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
|
|---|---|---|---|---|---|
Neurotoxicity in all prion disorders is believed to result from the accumulation of PrP-scrapie ( PrPSc ) , a β-sheet rich isoform of a normal cell-surface glycoprotein , the prion protein ( PrPC ) . Limited reports suggest imbalance of brain iron homeostasis as a significant associated cause of neurotoxicity in prion-infected cell and mouse models . However , systematic studies on the generality of this phenomenon and the underlying mechanism ( s ) leading to iron dyshomeostasis in diseased brains are lacking . In this report , we demonstrate that prion disease–affected human , hamster , and mouse brains show increased total and redox-active Fe ( II ) iron , and a paradoxical increase in major iron uptake proteins transferrin ( Tf ) and transferrin receptor ( TfR ) at the end stage of disease . Furthermore , examination of scrapie-inoculated hamster brains at different timepoints following infection shows increased levels of Tf with time , suggesting increasing iron deficiency with disease progression . Sporadic Creutzfeldt-Jakob disease ( sCJD ) –affected human brains show a similar increase in total iron and a direct correlation between PrP and Tf levels , implicating PrPSc as the underlying cause of iron deficiency . Increased binding of Tf to the cerebellar Purkinje cell neurons of sCJD brains further indicates upregulation of TfR and a phenotype of neuronal iron deficiency in diseased brains despite increased iron levels . The likely cause of this phenotype is sequestration of iron in brain ferritin that becomes detergent-insoluble in PrPSc-infected cell lines and sCJD brain homogenates . These results suggest that sequestration of iron in PrPSc–ferritin complexes induces a state of iron bio-insufficiency in prion disease–affected brains , resulting in increased uptake and a state of iron dyshomeostasis . An additional unexpected observation is the resistance of Tf to digestion by proteinase-K , providing a reliable marker for iron levels in postmortem human brains . These data implicate redox-iron in prion disease–associated neurotoxicity , a novel observation with significant implications for prion disease pathogenesis .
Imbalance of brain iron homeostasis is considered an important contributing factor of neurotoxicity in several neurodegenerative disorders including Parkinson's disease , Alzheimer's disease , and Huntington's disease [1]–[4] . Recent evidence suggests a similar alteration of iron homeostasis in prion disorders , a group of neurodegenerative conditions affecting humans and animals [5]–[10] . Since prion disorders are believed to result from a change in the conformation of cellular prion protein ( PrPC ) from an α-helical to a β-sheet rich PrP-scrapie form ( PrPSc ) , disturbance of brain iron homeostasis is an unexpected outcome [11] , [12] . Direct demonstration of increased total iron including Fe2+ and Fe3+ ions in the cerebral cortex , striatum , and brain stem of scrapie-infected mice , and an increase in markers of oxidative stress such as free malondialdehyde in diseased brains supports a role for redox-iron in prion disease pathogenesis , though the generality of this phenomenon and the underlying cause remain unidentified [5] , [6] , [10] , [13] . Bearing in mind the pathophysiology of prion disorders , it is difficult to explain disruption of iron metabolism by the conversion of PrPC to PrPSc , the principal event in all prion disorders . Under normal conditions , cellular iron homeostasis is maintained by a set of iron regulatory proteins that respond to intracellular iron levels by regulating their expression . In iron deficient conditions , iron uptake proteins transferrin ( Tf ) and transferrin receptor ( TfR ) are up regulated , and the iron storage protein ferritin is down regulated . The net result is an increase in the bio-available labile iron pool . The opposite scenario takes effect when iron is in excess [14] . A tight regulation of cellular iron is important since iron is not only an essential component of enzymes and proteins required for optimal neuronal growth and function , but is also toxic due to its ability to exist in two oxidation states , ferric ( Fe3+ ) and ferrous ( Fe2+ ) [14]–[16] . Under conditions where ferritin is unable to detoxify iron , Fe2+ iron can participate in one-electron transfer reactions such as Fenton or Haber-Weiss' reactions , resulting in cell death . Thus , alteration of brain iron homeostasis in diseased brains is likely to induce significant neurotoxicity , an observation that has received little attention [16] . Although the origin of iron imbalance in diseased brains is unclear , in vitro studies conducted previously provide some clues to the underlying mechanism ( s ) . It is likely that the affinity of PrPC for iron and hemin and the influence of this interaction on its conformational change to PrPSc renders iron inaccessible from PrPSc aggregates [17] , [18] . This would increase total cellular iron while decreasing the bio-available iron pool , inducing a state of cellular iron deficiency in the presence of excess iron . Alternately , since PrP is involved in cellular iron uptake and transport , a disruption of this function by aggregation of PrP to the PrPSc form may alter cellular iron homeostasis , creating a state of iron imbalance [19] . Though plausible , this phenomenon has not been evaluated in prion disease affected brains . To fill this gap , we examined a large sample size of sporadic Creutzfeldt-Jakob disease ( sCJD ) affected human brains , scrapie infected mouse brains inoculated with the 139A strain [20] , and hamster adapted Transmissible Mink Encephalopathy ( TME ) or 263K scrapie inoculated hamster brains to investigate the presence of iron imbalance during disease progression and at end-stage of disease [21] . In addition to measuring brain iron levels , particular attention was directed to Tf , the iron uptake protein that responds to minor changes in cellular iron levels , is unaltered by disease associated astrogliosis [2] , [14]–[16] , and resists harsh protease digestion ( see below ) , making it possible to assess iron homeostasis in post-mortem human brains . Snap frozen mouse and hamster brains served as additional controls to rule out any effects of post-mortem interval in the human tissue . In addition , cell models that replicate prions in vitro were used to understand the underlying cause of altered iron homeostasis due to prion infection [22] , [23] . We report that diseased human , hamster , and mouse brains show signs of iron imbalance , confirming the generality of this abnormality regardless of the prion strain or the affected species . The alteration of brain iron status appears early in the disease process , concomitant with the appearance of PrPSc , and is not an outcome of end-stage disease . Furthermore , we demonstrate the sequestration of iron in detergent insoluble PrPSc-ferritin complexes in PrPSc infected cell models and in sCJD brain homogenates , explaining the cause of iron imbalance in prion disease affected brains .
To evaluate if prion infection induces imbalance in brain iron homeostasis , brain tissue from control ( NH ) and end-stage diseased hamsters inoculated with the hyper strain ( HY ) of hamster adapted TME ( HaHY ) [17] was assessed for PrPSc levels , total iron , redox-active ferrous iron , and levels of major iron management proteins Tf and ferritin . Accordingly , NH and HaHY brain homogenates were digested with proteinase-K ( PK ) and analyzed by SDS-PAGE and immunoblotting . Probing with anti-PrP antibody 3F4 revealed the expected glycoforms of PrPC comprising of unglycosylated and two glycosylated forms migrating between 27 and 37 kDa in the NH sample ( Figure 1A , lane 1 ) . HaHY showed significantly more reactivity and additional bands migrating between 19 and 37 kDa , typical of PrPSc ( Figure 1A , lane 4 ) . Treatment with 50 and 100 µg/ml of PK digested all PrPC in NH ( Figure 1A , lanes 2 and 3 ) , whereas HaHY showed typical PK-resistant PrPSc forms migrating between 19 and 30 kDa ( Figure 1A , lanes 5 and 6 ) . Probing for Tf revealed relatively higher expression in HaHY compared to NH ( Figure 1A , lanes 1 and 4 ) , and surprisingly , partially cleaved PK-resistant Tf forms that migrated ∼10 kDa faster in both NH and HaHY samples ( Figure 1A , lanes 2 , 3 , 5 , 6 ) . Similar resistance to PK was noted in commercially available human Tf and normal and sCJD affected human brain tissue ( see below ) . Immunoblotting for ferritin showed a significant increase in HaHY infected samples as observed for Tf ( data not shown ) . Quantitative estimation of PrP , Tf , and ferritin after normalization with β-actin showed an increase of 4 . 3 , 2 . 8 , and 2 . 9 fold respectively in HaHY compared to matched NH samples ( Figure 1B ) . ( β-actin is cleaved by PK ( Figure 1A , lanes 2 , 3 , 5 , 6 ) . Estimation of ferrous ( Fe ( II ) ) and ferric ( Fe ( III ) ) iron showed significantly more reactivity in HaHY samples compared to matched NH samples as reported earlier for diseased human and mouse brains ( Figure 1C ) [16] . Total iron was increased by 4 . 9 fold in the HaHY sample relative to matched NH samples ( Figure 1D ) . A similar analysis of normal ( NH ) and scrapie infected end stage mouse ( MoSc ) brain tissue showed increased accumulation of PrP and up regulation of ferritin and Tf in MoSc samples compared to matched controls ( Figure 2A , lanes 1–3 and 4–7 ) . Quantitative estimation of PrP , ferritin , and Tf showed an increase of 4 . 9 , 3 . 9 , and 1 . 4 fold respectively in MoSc compared to matched controls ( Figure 2B ) . In addition , MoSc samples revealed increased reactivity for both Fe ( II ) and Fe ( III ) iron compared to matched controls ( Figure 2C ) as noted for hamster samples above and reported previously [18] . To evaluate if diseased human brains show a similar phenotype of iron imbalance , brain homogenates from 20 cases of sCJD ( CJD+ ) and matched controls ( CJD− ) were fractionated by SDS-PAGE and immunoblotted as above . Probing for PrP showed the expected glycoforms migrating between 27 and 37 kDa in CJD− samples ( Figure 3 , odd lanes ) . CJD+ samples showed relatively more reactivity for PrP and an additional 19–20 kDa band typical of PrPSc ( Figure 3 , even lanes ) . Re-probing for ferritin and glial fibrillary acidic protein ( GFAP ) , a marker for glial cells , showed an inconsistent difference between CJD− and CJD+ samples ( Figure 3 , odd and even lanes ) . However , Tf and TfR revealed a consistent increase in CJD+ samples compared to age-matched controls ( Figure 3 , odd and even lanes ) . Since autopsy samples represent end-stage disease and carry the additional risk of artifacts due to post-mortem interval , a similar evaluation was carried out on human biopsy samples . Due to limited amount of biopsy tissue , PVDF membranes used for diagnostic purposes were utilized . Membranes from two representative CJD positive cases containing control ( CJD− ) and CJD positive ( CJD+ ) samples either left untreated or subjected to PK treatment were reprobed for TfR , ferritin , Tf , and β-actin ( Figure 4 , odd and even lanes , respectively ) . As noted for autopsy samples , CJD+ biopsy samples demonstrated an increase in TfR and Tf levels , and an insignificant difference in ferritin levels compared to age-matched CJD− controls after normalization with β-actin ( Figure 4 , lanes 5–8 , 13–16 , 9–12 ) . These results provide confidence that the differences between CJD− and CJD+ samples are not influenced by post-mortem interval . As observed for diseased hamsters , Tf showed resistance to PK digestion in CJD− and CJD+ samples ( Figure 4 , lanes 14 and 16 ) , a novel finding that could prove useful in assessing the iron status of healthy and diseased post-mortem brains . This observation was therefore further validated by treating 10 and 5 ng of purified Tf isolated from human plasma with 50 µg/ml of PK for 30 minutes at 37°C in the presence of non-ionic detergents ( Figure 5A , lanes 1–4 ) . An aliquot ( 5 µl ) of CJD− brain homogenate was processed in parallel ( Figure 5A , lanes 5–8 ) . Purified Tf migrated as a single band ( Figure 5A , lanes 1 and 3 , black arrow ) and was cleaved to faster migrating forms when exposed to PK ( Figure 5A , lanes 2 and 4 , open arrowhead ) . Brain Tf migrates as a doublet , probably representing glycoforms of Tf ( Figure 5A , lanes 5 and 7 , black arrow ) , and is cleaved to faster migrating forms by PK treatment that co-migrate with similarly treated fragments of purified Tf ( Figure 5A , lanes 6 and 8 , open arrowhead ) . A similar resistance to PK has been reported for human brain ferritin [18] . To determine if the increase in Tf levels in CJD+ samples is accompanied by an increase in its iron content , CJD− and CJD+ samples were separated on a non-denaturing gel in duplicate . Proteins were either reacted with Ferene-S , a dye that forms a blue reaction product with iron ( Figure 5B , lanes 1 and 2 ) [24] , or transblotted to a PVDF membrane and probed for ferritin and Tf ( Figure 5B , lanes 3 and 4 ) , followed by silver staining of the membrane ( Figure 5B , lanes 5 and 6 ) . An increase in the level of ferritin and Tf expression and reactivity for iron was detected in the CJD+ sample ( Figure 5B , lanes 1–4 , black arrowheads ) . The slower migrating Tf reactive band probably represents a complex or glycosylated form of Tf that does not react with ferene-S for reasons that are unclear from our data ( Figure 5B , lanes 2 and 4 , white asterisk ) . Separation of purified 59Fe labeled Tf on a similar gel system reveals three labeled bands , though it is difficult to conclude which of the Tf bands from brain homogenates correspond to the isoforms of 59Fe-Tf [19] . Silver staining of the membrane shows two prominent unidentified bands in CJD− and CJD+ samples , and equal loading of protein in both samples ( Figure 5B , lanes 5 and 6 , black asterisk ) . Quantitative analysis of 20 CJD+ and an equal number of age-matched CJD− autopsy samples showed a significant increase in PrP , Tf , and TfR levels in CJD+ samples by 1 . 5 , 1 . 6 , and 1 . 5 fold respectively . The difference in ferritin levels between CJD− and CJD+ samples is statistically insignificant ( Figure 5C ) . Variability in the levels of different proteins within CJD− and CJD+ samples is shown in Figure 6A . The correlation between PrP and Tf was 0 . 05 in CJD− samples , and significantly higher at 0 . 444 in CJD+ samples . A similar assessment of PrP and iron revealed a correlation of 0 . 161 in CJD− samples , and 0 . 274 in CJD+ samples . Correlation between iron and Tf was 0 . 042 in CJD− samples as expected , and 0 . 319 in CJD+ samples ( Figure 6A ) . Although the correlation between PrP and ferritin was also higher at 0 . 602 in CJD+ samples relative to 0 . 327 in CJD− samples , these observations could reflect the effect of astrogliosis with disease progression . Correlative values for TfR were not significantly different between the two groups . It is notable that the values for total iron are surprisingly similar in CJD+ samples , suggesting similar iron levels at end stage disease despite obvious differences in the background and disease course of each case ( Figure 6A ) ( see Materials and Methods ) . Comparison of total brain iron in 55–66 year old CJD− patients carrying no definitive diagnosis with age-matched CJD+ cases shows a 1 . 4 fold increase of total iron in the latter ( Figure 6B ) . Other CJD− samples , including cases of Alzheimer's disease , vascular disorders , dementia , and inflammation showed high levels of iron , each varying with the specific disease ( Figure 6B ) . However , Tf and TfR levels did not show a corresponding decrease perhaps due to precipitation of iron in the extracellular space with little or no effect on cellular iron levels ( Figures 3 , 5C , and 6A ) . Linear regression analysis of the above data showed a positive correlation between PrP and Tf levels in CJD+ samples ( R = 0 . 444 ) , not in CJD− samples ( R = 0 . 05 ) ( Figure 6C ) . Immunostaining of brain sections revealed increased reactivity for Tf in cerebellar Purkinje cell neurons of CJD+ cases ( Figure 7 ) , indicating up-regulation of TfR expression in these cells . No visible deposits of iron were detected in any brain region of CJD+ cases , ruling out precipitation or co-aggregation of iron with PrPSc deposits ( data not shown ) . Increased levels of Tf , an iron uptake protein , in the presence of increased brain iron content observed above indicates imbalance of iron homeostasis in prion disease affected brains . A positive correlation between PrP and Tf levels further suggests that iron deficiency may occur as a result of PrPSc accumulation . Since a similar phenotype of iron deficiency is detected in human biopsy samples , alteration of brain iron homeostasis appears to occur before end-stage disease . It is notable that Tf levels are unlikely to be affected by secondary factors such as astrogliosis and neuronal loss , prominent features of end-stage prion disease [11] , [12] , and reflect neuronal iron status . In addition , our data indicate that Tf is a reliable marker of iron status in post-mortem tissue since it resists digestion by proteinase K and is likely to survive mild autolytic conditions . To evaluate the relationship between PrPSc accumulation and iron deficiency and to determine if imbalance of brain iron homeostasis appears earlier during the incubation period , hamster brain tissue harvested at 6 , 9 , and 12 weeks after inoculation with 263K strain of scrapie was analyzed . Immunoblot analysis showed a significant increase in PrP levels with disease progression as expected ( Figure 8A , lanes 1–9 ) . PK resistant PrPSc appeared 9 weeks after inoculation , and could be detected till 12 weeks , the end stage of disease ( Figure 8B , lanes 1–18 ) . Tf and TfR levels showed a significant increase at 9 weeks , and continued to increase till end stage of disease at 12 weeks post-inoculation ( Figure 8A , lanes 4–9 ) . Levels of GFAP , a marker of glial cells , showed a significant increase at 9 weeks and little change there-after ( Figure 8A , lanes 4–9 ) . Quantitative estimation of PrP , Tf , TfR , and GFAP levels in 9 and 12 week samples relative to 6 week post-inoculation samples revealed the following changes: PrP was increased by 3 . 4 and 4 . 4 fold , Tf was increased by 3 . 8 and 4 . 0 fold , TfR was increased by 5 . 2 and 5 . 7 fold , and GFAP was increased by 7 . 0 and 7 . 8 fold ( Figure 9 ) . Since Tf is a reliable indicator of brain iron status as demonstrated above , PrP and Tf levels at 6 , 9 , and 12 weeks post-infection were plotted to quantify the strength of association between the two proteins ( Figure 10 ) . Although limited data points suggest caution in interpretation , our results indicate highly significant and robust association indicating an exponential increase in Tf levels with increase in PrP expression ( R2 = 0 . 91 ) ( Figure 10 ) . These results suggest that the rate of neuronal iron deficiency increases with PrPSc accumulation , with substantially higher increase between 9 to 12 weeks relative to that evidenced between 6 and 9 weeks . Since these hamsters are relatively young and lack other associated diseases that could contribute to these observations , our results suggest increasing iron deficiency with increase in PrPSc accumulation . It is plausible that a state of iron deficiency despite elevated iron in diseased brains arises from sequestration of iron in biologically unavailable PrPSc-ferritin aggregates as suggested in a previous report [18] . To evaluate this possibility , N2a and prion infected ScN2a cells were exposed to 0 . 1 mM ferrous ammonium citrate ( FAC ) overnight to up-regulate ferritin expression . Treated cells were radiolabeled with 59FeCl3-citrate complex for 4 hours and chased in normal medium to accumulate 59Fe within ferritin . Cells were lysed in native lysis buffer and analyzed as such , or supplemented with 1% SDS for 10 minutes without boiling ( Figure 11A , lanes 1–4 ) or with boiling ( Figure 11A , lanes 5 and 6 ) . Subsequently , the samples were separated by native gel electrophoresis followed by autoradiography . A prominent iron labeled band was detected in both N2a and ScN2a lysates ( Figure 11A , lanes 1 and 3 ) . ( This band represents iron loaded ferritin as determined by re-fractionation of eluted proteins from the labeled band on SDS-PAGE followed by immunoblotting with anti-ferritin antibody ( Figure 11A , lanes 1–4 , lower panel ) ) . Exposure to SDS dissociated a significant amount of iron from ferritin in N2a lysates ( Figure 11A , lanes 1 and 2 ) , but had little effect on the iron content of ferritin in ScN2a lysates ( Figure 11A , lanes 3 and 4 ) . Boiling in the presence of SDS released all iron from ferritin in both N2a and ScN2a lysates ( Figure 11A , lanes 5 and 6 ) . Silver staining of lanes 2 and 4 showed insignificant difference in ferritin protein levels in SDS treated N2a and ScN2a cells ( Figure 11B ) , indicating that the difference arose from the iron content of ferritin . These results suggest that iron is sequestered in ‘SDS-insoluble’ ferritin aggregates in ScN2a cells , most likely in association with PrPSc . Similar results were obtained with scrapie infected SMB cells , another prion infected cell line ( data not shown ) [23] . It is notable that incorporation of iron in ferritin from N2a cells is significantly more compared to ScN2a cells ( Figure 11A , lanes 1 and 3 ) , probably due to inaccessibility of iron to aggregated ferritin in the latter . Since PrPSc from CJD+ brain homogenates is several-fold more resistant to PK digestion than PrPSc from ScN2a and SMB cells , a similar evaluation was carried out on CJD− and CJD+ brain homogenates . Thus , commercially available purified ferritin , CJD− and CJD+ brain homogenates were suspended in native lysis buffer or supplemented with 2% SDS and boiled for 10 minutes before separating on a non-denaturing gel . In-gel staining of separated proteins with the iron binding dye Ferene-S revealed a sharp blue band of iron in purified ferritin as expected , which was lost on boiling due to release of iron from denatured ferritin ( Figure 12 , lanes 1 and 2 , respectively ) . A similar evaluation of CJD− and CJD+ samples showed more iron in CJD+ samples relative to CJD− controls under both conditions ( Figure 12 , lanes 3–6 ) . A parallel set of samples was transferred to a PVDF membrane under non-denaturing conditions and probed for ferritin . Treatment of CJD− samples with boiling SDS denatured ferritin , some of which migrated with the dye front ( Figure 12 , lanes 7 and 8 ) . A similar treatment of CJD+ samples had minimal effect on ferritin that migrated as oligomers without boiling , and as a relatively compact band after boiling ( Figure 12 , lanes 9 and 10 , respectively ) . Re-probing for Tf revealed relatively higher levels in the CJD+ sample , though the bands did not react with Ferene-S , perhaps due to denaturation by SDS ( Figure 12 , lanes 11–14 ) . Silver staining of the same membrane revealed equal loading of protein in all samples ( Figure 12 , lanes 15–18 ) . To evaluate the cellular site where PrPSc and ferritin associate , ScN2a and SMB cells [22] , [23] analyzed in Figure 11 above were immunostained for PrP and ferritin and examined ( Figure 13 ) . In contrast to control ScN2a cells that displayed minimal reaction for PrP and cytoplasmic staining for ferritin ( Figure 13 , panels 1–3 ) , FAC exposed ScN2a cells revealed prominent intracellular reaction for PrP and ferritin , and significant co-localization of the two proteins ( Figure 13 , panels 4–6 ) . Scrapie infected SMB cells showed significant intracellular reaction for PrP and minimal reaction for ferritin at steady state ( Figure 13 , panels 7–9 ) . However , exposure to FAC revealed prominent intracellular vesicles that reacted for both PrP and ferritin ( Figure 13 , panels 10–12 ) . Together , the above results indicate that PrPSc and ferritin form detergent insoluble aggregates that sequester iron , creating a state of relative iron deficiency in prion infected cells .
In this report we demonstrate that prion infection induces a state of iron imbalance in diseased human , hamster , and mouse brains as revealed by a phenotype of iron deficiency in the presence of increased total iron . The phenotype of iron deficiency appears early in the disease process , probably with the initiation of PrPSc formation , and is not an outcome of end-stage disease . An important underlying cause of iron imbalance is sequestration of cellular iron in detergent insoluble PrPSc-ferritin complexes , thus rendering it bio-unavailable and creating a phenotype of apparent iron deficiency . Consequent up-regulation of iron uptake proteins Tf and TfR result in increased cellular iron uptake , creating a state of cellular iron imbalance . Since iron is potentially toxic due to its redox-active nature , these observations have significant implications for prion disease associated neurotoxicity . In evaluating iron homeostasis in healthy and diseased brains , we have focused primarily on total brain iron and levels of Tf and TfR . Although these parameters do not represent changes in iron levels in individual cell types that comprise the complex milieu of the brain , these are reliable markers of neuronal iron levels , the principal cell type affected in prion disorders [14] . The mechanisms underlying brain iron homeostasis are complex , and details of iron import and export from various cells within the brain are still emerging . Normally , iron enters the brain through the Tf/TfR pathway across capillary endothelial cells , and is delivered to the brain interstitium using mechanisms that are still debated . Once in the brain , iron is bound by citrate , ascorbate , and Tf present in the interstitial fluid . Citrate and ascorbate are mainly released by astrocytes , and brain Tf is derived from cells of the choroid plexus and oligodendrocytes and is saturated with iron under normal conditions . Neurons take up most of their iron from Tf through TfR mediated uptake . Under conditions of iron deficiency , neurons up-regulate TfR levels to internalize increased amounts of iron saturated Tf to make up for the deficiency . The opposite scenario holds true when iron is in excess . In contrast to other cell types such as astrocytes , oligodendrocytes , and microglia , neurons in most brain regions lack the iron storage protein ferritin , necessitating efflux of excess iron through the iron export protein ferroportin . Astrocytes , microglia , and oligodendrocytes lack TfR , and take up most of their iron through other mechanisms [2] , [14]–[16] . Thus , TfR levels in the brain reflect the iron status of neurons , the most vulnerable cell population in prion disorders . However , experimental manipulation of brain tissue often results in the degradation of TfR , providing inconsistent results from post-mortem brains and animal brains unless the tissue is handled with extreme care . Tf , on the other hand , is as resilient to degradation as PrPSc , and reflects TfR levels as demonstrated by our immunohistochemistry and immunoblot data from animal and human brains . It is for these reasons that we have focused on Tf and iron levels , and where possible , TfR levels to evaluate brain iron status in prion disease affected brains . Using the above criteria , our data demonstrate the presence of iron imbalance in prion disease affected brains from three different species , human , hamster , and mice . Although alteration of brain iron has been described in other neurodegenerative diseases such as AD and PD , it is notable that the changes in iron homeostasis in prion disease affected brains are distinct [1]–[4] . A prominent difference is the apparent phenotype of iron deficiency in prion disease affected brains despite increased brain iron content . A direct correlation between PrP and Tf levels in both sCJD and infected hamster samples suggests that iron deficiency arises as a direct consequence of PrPSc accumulation , perhaps by forming PrPSc-ferritin complexes as noted in scrapie infected cells and reported earlier [18] , [25] . An increase in Tf and TfR levels is also noted in human biopsy samples , indicating that brain iron levels decrease much before the onset of end-stage pathology . The presentation of AD and PD brains , however , is strikingly different . In the pool of CJD− cases examined in this report , several samples contained high levels of iron , including cases of AD , PD , dementia of unknown origin , and vascular disorders . As opposed to prion disease affected brains , these cases demonstrated a consistent down-regulation of Tf and TfR as expected , suggesting that the increased iron content was probably derived from insoluble deposits in the extra-cellular space . Brains from sCJD cases , on the other hand , show a modest and a surprisingly similar increase in iron content across twenty samples and a significant increase in Tf levels , suggesting a common pathway of iron accumulation and sequestration by scrapie infection . The data on scrapie infected hamsters support the above conclusions . In these animals , iron deficiency appears almost at the same time as PK-resistant PrPSc , and worsens with disease progression as indicated by increasing levels of Tf and TfR till end stage disease . Since this analysis was performed on experimental brains that were snap frozen after harvesting , the increase in TfR levels with disease progression provides a direct measure of neuronal iron deficiency , further supported by an increase in Tf that is likely to reflect TfR levels as observed in the Purkinje cell neurons of CJD+ samples . Since most neurons lack ferritin iron stores [14] , such a situation is likely to result in compromised neuronal health and toxicity . It is unlikely that advanced age contributes to the iron imbalance observed in prion disease affected brains since the hamster brains used for this evaluation range in age from ten to sixteen weeks , a relatively young age considering the life-span of hamsters . It is interesting to note that the extent of iron accumulation varies with the strain of prions used for infecting the experimental animal . Hamsters infected with TME accumulate much higher levels of total and redox-active iron than those infected with the 263K strain of scrapie ( unpublished observations ) . However , the increase in Tf levels is similar in both cases , suggesting that iron in some cases precipitates in the extra-cellular space and does not alter the intracellular iron status . The most likely cause of iron deficiency in prion disease affected brains is the sequestration of iron in PrPSc-ferritin aggregates as noted in scrapie infected cells and sCJD brain tissue . In a previous report we demonstrated that PrPSc from sCJD and scrapie infected mouse tissue forms a PK-resistant , detergent insoluble complex with ferritin [18] , [25] . Here we show that ferritin from sCJD affected brains sequesters iron , and is not denatured by boiling in 2% SDS for ten minutes . Ferritin from normal human brains is denatured by this treatment and releases bound iron . A similar observation is noted in scrapie infected cell lines , though cellular ferritin is less resistant to SDS . These observations suggest that ferritin co-aggregates with PrPSc to form a detergent insoluble complex and sequesters the associated iron , creating a state of iron deficiency . Such an occurrence would explain the development of iron deficiency with the appearance of PrPSc , increased levels of total iron in diseased brains , a direct correlation between PrP and Tf levels , and an overall state of iron imbalance in diseased brains . Although ferritin is mainly cytosolic , it is known to undergo degradation within the lysosomes , the cellular compartment where PrPSc and ferritin are observed in scrapie infected cells , and are hence likely to form the complex [18] . To summarize , this report demonstrates that brain iron imbalance is a common feature of prion disease affected brains , and is likely to contribute to prion disease associated neurotoxicity due to the redox-active nature of iron . The underlying cause of iron imbalance is the formation of PrPSc-ferritin complexes , a pathogenic process that may lend itself to therapeutic manipulation , thus providing a means to reduce disease associated neurotoxicity .
Mouse neuroblastoma cells ( N2a ) and ScN2a cells were obtained from Dr . Byron Caughey ( Rocky Mountain laboratories ) , and scrapie infected SMB cells ( mouse brain cells of mesenchymal origin were obtained from the TSE Resource Center , Institute of Animal Health ( Edinburgh , Scotland ) . PrP-specific monoclonal antibodies 3F4 and 8H4 were obtained from Signet ( Dedham , MA , USA ) and Drs . Man-Sun Sy and Pierluigi Gambetti ( National Prion Surveillance Center , Case Western Reserve University ) . Anti-ferritin antibody was from Sigma-Aldrich ( St . Louis , MO ) , anti-Tf from GeneTex ( San Antonio , TX ) and anti-TfR antibody from Zymed Laboratories Inc ( Carlsbad , CA ) . Horseradish peroxidase ( HRP ) -labeled secondary antibodies were from GE Healthcare ( Little Chalfont , Buckinghamshire , United Kingdom ) . All other chemicals were purchased from Sigma . Frozen human brain tissue from the frontal cortex of CJD+ and matched controls ( CJD− ) was obtained from the National Prion Surveillance Center , Case Western Reserve University . Additional samples of control tissue ( CJD− ) were obtained from the Harvard Tissue Bank . Collectively , the samples ranged in age from 37–80 years , and the disease duration ranged from 1–24 months in CJD− , and 2 to 4 months in CJD+ cases . CJD− cases carried the diagnosis of extensive infarcts , possible granulomatous angiitis , cerebral vascular disease , and necrotizing lymphoplasmacytic meningoencephalitis ( 1 case each ) , Alzheimer's disease ( 5 cases ) , and no definitive diagnosis ( 11 cases ) . CJD+ cases were classified as MM1 ( 11 cases ) , VV1 ( 2 cases ) , VV2 ( 4 cases ) , and MV2 ( 3 cases ) . Male and female representation was almost equal among the CJD− and CJD+ groups . Given the heterogeneity in the two groups , samples were matched only for age when loading on SDS-PAGE . For SDS-PAGE , 10% homogenate was prepared in lysis buffer ( 10 mM Tris , pH 7 . 5 , containing 150 mM NaCl , 0 . 5% sodium deoxycholate , 0 . 5% nonident P-40 containing PMSF and Protease inhibitor cocktail ) . For determination of iron , the sample was prepared in PBS and analyzed using the Teco diagnostic kit as described below . Brain biopsy samples were obtained from the Prion Disease Surveillance Center at Case Western Reserve University , and represent individuals ranging in from 46 and 80 years . CJD− cases presented with positive neurological symptoms for 1–12 months , and CJD+ cases for 2–4 months . Due to the limited amount of tissue available from biopsies , PVDF membranes used for diagnostic purposes were re-probed to evaluate the expression of major iron regulatory proteins ferritin , Tf and TfR . Control and TSE infected end stage hamster brain tissue was obtained from Jason Bartz ( Department of Medical Microbiology and Immunology , Creighton University , Omaha , Nebraska ) [21] was analyzed as above for protein analysis , and homogenized in PBS for estimating total , Fe ( II ) , and ( Fe ( III ) iron [26] . To evaluate the point of initiation of iron dyshomeostasis during disease progression , hamster brain tissue collected at 6 , 9 , and 12 weeks after inoculation with 263K strain of scrapie was obtained from Qingzhong Kong ( Case Western Reserve University ) . Half brain from each sample was homogenized in lysis buffer analyzed for PK resistant PrPSc as well as expression of other proteins . Control and scrapie infected end stage mouse brain tissue ( strains 137A and 22L ) was generated in our facility and obtained from Jason Bartz and analyzed as above . Normal ( NH ) and diseased brains homogenized in lysis buffer ( 10 mM Tris , pH 7 . 5 , containing 150 mM NaCl , 0 . 5% sodium deoxycholate , 0 . 5% nonident P-40 ) were incubated with 50 and 100 µg/ml of PK at 37°C for h . Digestion was stopped by addition of 1 mM PMSF and protease inhibitor cocktail . Digested and undigested samples were processed for SDS-PAGE and Western blotting as described previously [27] , [28] . Membranes containing transferred proteins were probed with anti-PrP antibodies 3F4 ( 1∶5000 ) or 8H4 ( 1∶3000 ) , anti-ferritin ( 1∶1000 ) , anti-Tf ( 1∶6000 ) , anti-TfR ( 1∶3000 ) , anti-GFAP ( 1∶2000 ) , and anti-β-actin ( 1∶7500 ) followed by appropriate secondary antibody conjugated with horseradish peroxidase ( 1∶6000 ) . Immunoreactive bands were visualized by ECL detection system ( Amersham Biosciences Inc . ) . A colorimetric method ( Teco Diagnostics , Anaheim , CA ) was used to estimate total iron in brain homogenates according to the manufacturer's instructions . In short , 20 µl of 10% homogenate in PBS was dissolved in 1 ml of acetate buffer containing 220 mM Hydroxylamine hydrochloride , pH 4 . 5 and 50 µl of 16 . 6 mM Ferrozine was added to develop the color . The change in color was recorded at 560 nm after 10 minutes of incubation at 37°C using 500 mg/dl FeCl2 solution as standard . Negative controls were recorded without the addition of ferrozine and background was subtracted from test samples . Mouse neuroblastoma cells ( N2a ) and scrapie infected ScN2a and SMB cells were cultured in the absence or presence of 0 . 1 mM ferrous ammonium citrate and labeled with 59FeCl3-citrate complex ( 1 mM sodium citrate and 2 µCi of 59FeCl3 in serum free Opti-MEM ) overnight at 37°C in 5% CO2 in a humidified atmosphere . Washed cells were lysed in native lysis buffer ( 0 . 14 M NaCl , 0 . 1 M HEPES , pH 7 . 4 , 1 . 5% Triton X-100 and 1 mM PMSF ) , and half of each sample was supplemented with 10 µl of 10% SDS for 10 minutes at room temperature . Both parts were mixed with glycerol , traces of bromophenol blue , and resolved on 3–9% native gradient gel as described by Petrak and Vyoral [29] , [30] with modifications [19] . Gels were vacuum dried and exposed to X-ray film to visualize labeled bands . A duplicate set of samples was analyzed on native gel , and the radioactive band was eluted with boiling SDS and re-fractionated on a SDS-PAGE gel followed by immunoblotting with anti-ferritin antibody . Brain homogenates prepared in 5 mM HEPES pH 7 . 4 , 0 . 25 M sucrose and protease inhibitors were resolved on native gradient gels ( 3–9% ) without Triton X-100 and stained for iron with freshly prepared Ferene-S ( 0 . 75 mM 3-[2-pyridyl]-5 , 6-bis ( 2-[-furyl sulfonic acid]-2 , 4-triazine , 2% ( v/v ) acetic acid , 0 . 1% thioglycolic acid ) [24] for 30 min at 37°C . A dark blue stain indicates the presence of iron or iron-bound protein . The identity of iron stained bands was confirmed by immunoblotting with specific antibodies . Prussian blue reaction was used to detect redox-active ferrous and ferric iron in brain homogenates as described in [26] . Paraffin embedded cerebellar sections of CJD+ cases and age matched controls were deparafinized with methanol followed by exposure to 8% H2O2 and water . Cleared sections were incubated with 10% NGS followed by anti-Tf antibody ( 1∶20 in 1% NGS ) overnight at 4°C . After washing with TBS the sections were incubated with peroxidase-conjugated secondary antibody for 60 min and developed with DAB . N2a and SMB cells exposed to FAC were processed for immunostaining as described in a previous report [18] . Human brain tissue from CJD+ and matched controls ( CJD− ) homogenized in PBS was used for the analyses . Brain homogenate was mixed with native lysis buffer ( 0 . 14 M NaCl , 0 . 1 M Hepes , 1 . 5% Triton X-100 , 1 mM PMSF , PH 7 . 4 ) , or supplemented with 2% SDS . Samples were either incubated at 37°C or boiled for 10 min and resolved by 3–9% native gradient gel . Iron levels were detected by Ferene S staining as described above . To confirm the identity of ferritin and Tf , proteins were immunoblotted under native conditions as described previously [19] and membranes were probed for ferritin and Tf . Equal loading was confirmed by silver staining of membranes . Data are presented as the mean SEM values . Statistical evaluation of the data was performed by using Students t-test ( unpaired ) . Regression analysis was performed by standard statistical methods and the Stata ( 2005 ) statistical package . | Prion disorders are neurodegenerative conditions of humans and animals that are invariably fatal . The main agent responsible for neurotoxicity in all prion disorders is PrP-scrapie ( PrPSc ) , a β-sheet rich isoform of a normal cell-surface glycoprotein , the prion protein ( PrPC ) . Deposits of PrPSc in the brain parenchyma are believed to induce neurotoxicity , though the underlying mechanisms are not entirely clear . Emerging evidence from prion-infected cell and mouse models implicates redox-iron in prion disease–associated neurotoxicity . However , a systematic evaluation of iron homeostasis in prion disease–affected brains and the underlying mechanism of iron dyshomeostasis are lacking . In this report , we demonstrate that prion disease–affected human , mouse , and hamster brains exhibit a state of iron deficiency in the presence of excess total brain iron , resulting in a state of iron imbalance . The underlying cause of this phenotype is likely sequestration of iron in cellular ferritin that becomes detergent-insoluble , possibly due to association with PrPSc . This results in a state of iron bio-insufficiency , leading to increased iron uptake by the cells and worsening of the state of iron imbalance . Since iron is highly toxic if mismanaged , these results implicate iron imbalance as a significant contributing factor in prion disease–associated neurotoxicity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"neurological",
"disorders/prion",
"diseases"
]
| 2009 | Abnormal Brain Iron Homeostasis in Human and Animal Prion Disorders |
Chromosomal location has a significant effect on the evolutionary dynamics of genes involved in sexual dimorphism , impacting both the pattern of sex-specific gene expression and the rate of duplication and protein evolution for these genes . For nearly all non-model organisms , however , knowledge of chromosomal gene content is minimal and difficult to obtain on a genomic scale . In this study , we utilized Comparative Genomic Hybridization ( CGH ) , using probes designed from EST sequence , to identify genes located on the X chromosome of four species in the stalk-eyed fly genus Teleopsis . Analysis of log2 ratio values of female-to-male hybridization intensities from the CGH microarrays for over 3 , 400 genes reveals a strongly bimodal distribution that clearly differentiates autosomal from X-linked genes for all four species . Genotyping of 33 and linkage mapping of 28 of these genes in Teleopsis dalmanni indicate the CGH results correctly identified chromosomal location in all cases . Syntenic comparison with Drosophila indicates that 90% of the X-linked genes in Teleopsis are homologous to genes located on chromosome 2L in Drosophila melanogaster , suggesting the formation of a nearly complete neo-X chromosome from Muller element B in the dipteran lineage leading to Teleopsis . Analysis of gene movement both relative to Drosophila and within Teleopsis indicates that gene movement is significantly associated with 1 ) rates of protein evolution , 2 ) the pattern of gene duplication , and 3 ) the evolution of eyespan sexual dimorphism . Overall , this study reveals that diopsids are a critical group for understanding the evolution of sex chromosomes within Diptera . In addition , we demonstrate that CGH is a useful technique for identifying chromosomal sex-linkage and should be applicable to other organisms with EST or partial genomic information .
The origin and evolution of sex chromosomes have long been of interest to geneticists and evolutionary biologists [1]–[3] . Besides their essential role in sex-determination , sex chromosomes are distinct from autosomes in their gene content , pattern of gene expression , rate of gene duplication and rate of protein evolution [4] . These differences are critical in shaping both the evolution of sexual dimorphism and patterns of speciation . Dramatic variation exists among animals in sex determination systems [3] and genomic organization of sex chromosomes [5] . Despite this variation , sex chromosome gene content is unknown for most species and only in a handful of taxonomic groups ( e . g . mammals , Drosophila ) are there sufficient comparative data to map the dynamics of sex chromosome evolution [6]–[12] . In this paper , we explore the utility of Comparative Genomic Hybridization ( CGH ) microarray experiments to differentiate X-linked genes from autosomal genes in several species of stalk-eyed flies from the genus Teleopsis . This analysis reveals the formation of a neo-X chromosome in Teleopsis relative to other Diptera , and identifies substantial gene movement between autosomes and sex chromosomes . Considerable theory [e . g . 13]–[17] has been developed under the assumption that sex chromosome identity is determined by the presence of a male-determining factor that initially appeared on an autosome . Once genic sex determination appears , sexually antagonistic alleles , i . e . those with beneficial effects in males but harmful effects in females , should increase in frequency if they are linked to the sex-determining factor [15] . As soon as multiple loci are required for male functionality ( or if there are alleles that cause sterility in the opposite sex ) , then selection will favor reduced recombination between the sex chromosomes . Lack of recombination , together with the joint effects of mutation , natural selection and genetic drift , are then expected to result in degeneration of the Y chromosome ( or W in female heterogametic species ) . Eventually , according to theory , extreme divergence in gene content between sex chromosomes will evolve with only a small number of genes essential for male function expected to survive on the Y . Once sex chromosomes appear , dosage compensation mechanisms will often evolve to equalize the gene expression differences that result from one sex being heterogametic [18] , [19] . In addition , selection for how genes are expressed on sex chromosomes relative to autosomes is expected to differ because X chromosomes are exposed to selection in males half as often as in females , assuming an equal sex ratio . Conflict will arise when alleles that favour one sex are harmful to the other . How such sexually antagonistic conflict is resolved depends on the dominance of the alleles [15] , [20] . A number of recent studies have attempted to evaluate these predictions and so far , largely support predictions based on dominant , rather than recessive , allelic effects . In Drosophila , a large fraction of the genome exhibits sex-biased expression [21] , [22] and X chromosomes show evidence of feminization , i . e . there are fewer male-biased genes and more female-biased genes on the X chromosome than on the autosomes and much of the bias in gene expression is associated with gametogenesis [23]–[25] . An alternative , but not mutually exclusive , explanation for why genes with male-biased expression are under-represented on the X chromosome is that this chromosome is transcriptionally silenced during spermatogenesis in flies [26]–[28] and mammals [29] . Meiotic sex chromosome inactivation ( MSCI ) has been used to explain why gene movements caused by retrotransposition ( detectable by the absence of introns in the derived copy ) more commonly involve movement from X chromosomes to autosomes than the reverse and typically exhibit expression in D . melanogaster testes [30] . Similar patterns have been found for several other Drosophila species in which genes have moved by both RNA and DNA-based mechanisms [9] , [31] . Movement of mammalian genes from the X to autosomes has also been linked to sex chromosome silencing [10] , [32] . Furthermore , the demasculinization of a neo-X chromosome in D . pseudoobscura appears to be driven not by shifts in sex-biased gene expression but rather by differential gene gain , loss and movement [25] . Male-biased genes on the neo-X of D . pseudoobscura have not evolved female-biased expression . Instead , they have been either lost from the X and then newly created on an autosome or moved from the X to an autosome . Stalk-eyed flies in the family Diopsidae are an excellent group in which to explore the dynamics of sex chromosome evolution . They have become an iconic system for studying sexual selection and the sex chromosomes harbor much of the genetic variation affecting numerous aspects of their reproductive biology . Many species exhibit dramatic sexual dimorphism in head shape in which the eyes are placed on the ends of stalks [33] . In some males , the outer most distance between the eyes ( eyespan ) is twice their body length . Sexual dimorphism in eyespan has evolved multiple times within the family [34] . Evidence suggests that long eyespan males in sexually dimorphic species succeed in male-male contests [35] and are also preferred by females [36]–[38] . Much of the genetic variation associated with exaggerated male eyespan in at least one dimorphic species , Teleopsis dalmanni , is located on the X chromosome [39] , [40] . In addition , X chromosome effects have also been found for sperm length [41] and parts of the female sperm storage organs [42] . An additional noteworthy feature of diopsids from southeast Asia is that several species in the genus Teleopsis [recently synonymized with Cyrtodiopsis , 43] exhibit a classic sex ratio polymorphism in which male carriers produce predominantly female offspring [44] , [45] . This meiotic drive system is genetically linked to eyespan such that male eyespan serves as an indicator of genetic quality due to an association between short eyespan and segregation distortion [46] . In T . dalmanni , the loci influencing drive is X-linked and appears to be associated with an inversion complex that restricts recombination on this chromosome [40] . Therefore , knowing which genes reside on the X chromosome in T . dalmanni would be extremely valuable for understanding the genetic consequences of sexual selection in the diopsids and provide another invertebrate system to compare with Drosophila . CGH allows the identification of sex-linkage on a substantially larger scale than has been possible to date by traditional linkage mapping . CGH microarray analysis is primarily used for fine scale identification of gene copy-number variation and is most commonly used in medical applications [47]–[49] . Because the technique involves the relatively simple approach of hybridizing DNA from two different tissues that may differ in gene copy-number at one or several regions of the genome , we expected that it might accurately distinguish the copy-number difference in X-linked genes between male and female DNA samples . Female stalk-eyed flies are XX and , therefore , should produce a hybridization intensity for X-linked genes that is twice that of males , which are XY . Alternatively , autosomal genes should exhibit no difference between the sexes , while Y-linked genes will produce a hybridization signal only for males . Therefore , oligonucleotide arrays were generated from an annotated EST library for T . dalmanni and genomic hybridization comparing male and female DNA were performed for this species and three congeneric species .
The T . dalmanni log2 ratio values of female-to-male signal intensities for 3444 genes , averaged across four hybridizations , exhibit a clear bimodal distribution with one peak at 0 and the second peak at 0 . 925 ( Figure 1; Table S1 ) . The average correlation in gene log2 ratio values across different hybridizations was 0 . 938 . Based on the histogram , the interval with the fewest entries ( 5 genes ) fell between 0 . 45 and 0 . 55 so a cut-off of 0 . 5 was used to distinguish autosomal from X-linked genes . With this criterion , 2891 genes were scored as autosomal , 533 as X-linked and 16 as unknown . Therefore , based on this sample of genes , the X chromosome represents approximately 15% of the T . dalmanni genome , a measure that is similar to an estimate of the relative size of the X chromosome ( 12% ) generated from mitotic chromosome lengths [Figure 2 , 39] . K-means clustering did not provide any chromosomal assignments that contradicted the first method but scored most of the unknown genes from the first method as X-linked . Overall , K-means scored 2895 genes as autosomal and 545 genes as X-linked . We validated chromosome classifications for 33 loci , four of which were putatively on the X and one was on the Y chromosome . After PCR and genotyping of at least 35 males and 35 females , all loci were assigned correctly with 28 putative autosomal loci containing male heterozygotes , and 4 putative X-linked loci lacking male heterozygotes ( Table S2 ) . One gene , ORF-126 , had an extremely low log2 ratio of −6 . 732 and the next lowest value for any gene was −0 . 569 . Therefore , we suspected this gene was Y-linked and designed primers from the EST sequence to test this possibility . We conducted PCRs each on 47 individual males and 47 individual females and found that all of the male PCRs produced a band while none of the female PCRs produced a band , confirming the location of ORF-126 on the Y chromosome . Inspection of the gene annotation for the D . melanogaster genes that are homologous to the T . dalmanni genes on the array reveals that 453 ( 90% ) of the X-linked genes in T . dalmanni are found on chromosome 2L in D . melanogaster ( Figure 2 ) . The remaining 46 X-linked genes that have a homologous gene in D . melanogaster are found on each of the other major chromosome arms in roughly equal numbers ( 15 from 2R , 11 from 3L , 11 from 3R and 9 from the X ) . Conversely , the T . dalmanni autosomal genes have relatively few homologs on 2L ( 86 genes ) compared to the other major Drosophila arms: 2R ( 601 ) , 3L ( 646 ) , 3R ( 805 ) , and X ( 569 ) ( Figure 2 ) . This pattern indicates a strong syntenic relationship between the X chromosome in T . dalmanni and chromosome 2L in D . melanogaster . Figure 1 depicts the log ratio values for the genes that violate this syntenic relationship . If the putative chromosomal movement indicated for these genes is an artifact of the methodology we would expect their log2 ratios to be concentrated more in the valley between the peaks but this is not the case . Using GeneMerge , for the small number of genes whose homologs are not located on 2L in D . melanogaster but are X-linked in T . dalmanni we tested whether their location on the non-2L chromosome in D . melanogaster is clustered within a specific chromosomal region on that chromosome . The program divides each chromosome into regions corresponding to the cytogenetic map designations and assesses whether a subset of genes in a sample are disproportionately represented in one of these regions . A clustered distribution might suggest the existence of a small primitive X chromosome that was subsequently fused with the 2L homolog , but the non-2L X-linked genes were randomly distributed across each of their respective homologous chromosome in D . melanogaster . In order to investigate the synteny between the non-2L chromosomes in D . melanogaster and the two autosomes in T . dalmanni we mapped 28 genes by genotyping flies from two previous F2 intercross experiments involving lines in which males were selected for increased or decreased relative eyestalk length [40] . Each of the genes contained a single amino acid repeat region that varied in length between the lines and , therefore , segregated as a length polymorphism . Inspection of the resulting linkage map ( Figure 3 , Table S2 ) reveals that one diopsid autosome contains multiple genes from the D . melanogaster X and 3L chromosome arms ( Muller elements A and D ) while the other autosome contains multiple genes from chromosome arms 2R and 3R ( Muller elements C and E ) . Two genes ( grainy head and CG32133 ) appear to have moved between Teleopsis autosomes based on their location in D . melanogaster . In addition to the T . dalmanni slides , we performed CGH experiments on three congeneric species—T . whitei , T . thaii and T . quinqueguttata . In all cases , cDNA from these species was hybridized to the microarray slide that contained probes generated from the T . dalmanni EST sequences . All three species exhibited a similar bimodal distribution of log2 ratio values ( Figure 4; Table S1 ) but with slightly reduced separation between the peaks . As a result , more genes were scored as ‘unknown’ for each of these species than for T . dalmanni . T . whitei , the species that is most closely related to T . dalmanni , had the fewest number of unknown genes ( 56 ) followed by T . quinqueguttata ( 144 ) and T . thaii ( 225 ) . The average correlation in log2 ratio values across different hybridizations ( 0 . 917 for T . whitei , 0 . 737 for T . thaii and 0 . 811 for T . quinqueguttata ) was also lower for these species than in T . dalmanni . The vast majority of genes ( 93 . 7% ) that are autosomal in T . dalmanni are also autosomal for the other Teleopsis species , indicating strong syntenic conservation within the genus . As with the T . dalmanni analysis , the K-means clustering of the data for the other Teleopsis species produced chromosome designations similar to the method based on the confidence intervals of individual probes . The K-means analysis scored the majority of the unknown genes as X-linked and in a few cases ( 5 genes in T . whitei , 4 genes in T . thaii , and 6 genes in T . quinqueguttata ) scored a gene that was autosomal in the first analysis as X-linked . As with ORF-126 in T . dalmanni , there were two genes , RNA polymerase II 215kD subunit ( RpII215 ) and a paralogous copy of Cuticular protein 35B ( Cpr35B ) which has duplicated several times within diopsids [50] , in T . quinqueguttata that had extremely low log2 ratio values , −5 . 107 and −4 . 378 respectively . PCR of 48 males and 48 females in T . quinqueguttata confirms a Y chromosome location for RpII215 but we have been unable to verify the location of the Cpr35B paralog through PCR presumably because we have not yet been able to design copy-specific PCR primers . Assuming a syntenic relationship between D . melanogaster 2L and the Teleopsis X chromosome ( and between all other D . melanogaster chromosomes and the Teleopsis autosomes ) and using the D . melanogaster and Anopheles gambiae chromosomal designations to polarize the reconstruction within Teleopsis , we reconstructed the pattern of gene movement on and off the X chromosome ( Figure 5; Table S3 ) . A total of 193 genes exhibit movement either between Drosophila and Teleopsis or within Teleopsis . Twenty-nine genes were not assigned a specific pattern of movement either due to a lack of homology with D . melanogaster and A . gambiae or because they produced ambiguous reconstruction within Teleopsis . For 65 of the genes that violated the syntenic relationship between D . melanogaster 2L and the Teleopsis X , it was more parsimonious , based on the syntenic relationship with A . gambiae , to ascribe the gene movement as occurring within the lineage leading to Drosophila rather than to Teleopsis . Of the remaining genes , there are a similar number of genes that have moved on to an autosome ( 27 ) as on to the X chromosome ( 20 ) in the lineage leading to Teleopsis . Within Teleopsis , movement of genes from the X chromosome to an autosome is concentrated ( 23 out of 29 gene moves ) on the branch leading to the monomorphic taxa , T . quinqueguttata , while all of the movement of genes off of the autosomes and onto the X chromosome occurs on the branches associated with the two most sexually dimorphic taxa , T . thaii and T . dalmanni ( Figure 5 ) . Analysis of the gene annotation associated with the pattern of gene movement indicates that the set of genes that have moved onto the X chromosome are overrepresented ( P = 0 . 049 ) for genes that function in the ‘structural constituent of ribosome’ ( GO:0003735 ) . This pattern is generated primarily by T . thaii as 5 of the 12 genes that have moved onto the X chromosome on the branch leading to this species are involved in this molecular function . Genes that had moved on to an autosome were not overrepresented for any functional category . Several studies in Drosophila have demonstrated that chromosomal location and the pattern of gene movement across chromosomes are associated with differential rates of protein evolution [51] , [52] , duplication [9] , [31] and divergent patterns of gene expression [22] , [24] , [25] . Estimates of each process were obtained for Teleopsis in an analysis of the eye-antennal imaginal disc transcriptome [50] . Therefore , we examined the relationship between these variables and gene location within the genus . The rate of protein evolution was measured as the percentage of amino acid change in the lineage leading to T . dalmanni relative to the amount of change for that gene in the Drosophila and Anopheles lineages . Using this index , there is little difference between genes that reside exclusively on the autosomes or the X chromosome in Teleopsis , but genes that have moved between these chromosomes are evolving significantly faster than those that have not moved ( P<0 . 0001; Kruskal-Wallis test; Figure 6 ) . This pattern is also supported if we look just at the genes that were determined , by a relative rate test , to be evolving significantly faster at the protein level in T . dalmanni than in D . melanogaster , D . pseudoobscura and D . virilis [50] . These genes are overrepresented in the set of genes that have moved chromosomes , particularly those that have moved onto the autosomes ( χ2 = 27 . 74 , P<0 . 0001; Chi-squared test ) . 2 . 62% of all genes were classified as evolving significantly faster in T . dalmanni , but 19 . 51% of the genes that have moved onto an autosome fell into this category , compared with 8 . 33% for the genes that have moved onto the X chromosome , 4 . 11% for exclusively X-linked genes and 1 . 97% for genes that are exclusively autosomal . The analysis of the eye-antennal imaginal disc EST database revealed 20 putative duplication events in the Teleopsis lineage [50] . An additional duplication that was not detected in that study , involving the sex determination gene transformer 2 ( tra2 ) , was revealed by the CGH results . Hybridization to probes generated from different consensus sequences ( conseqs ) that are both homologous to tra2 revealed that one sequence resides on an autosome while the second sequence is X-linked . Subsequent phylogenetic analysis of the divergence in protein sequence between the conseqs and Drosophila homologs confirms the CGH results . Of these 21 putative duplicates , 11 of their homologs in D . melanogaster reside on 2L while the remaining homologs are distributed in similar numbers among 2R ( 2 ) , 3L ( 2 ) , 3R ( 5 ) , and X ( 1 ) . This chromosome distribution is significantly overrepresented for genes on 2L ( χ2 = 14 . 143 , P = 0 . 0002; Chi-squared test ) . Given the syntenic relationship between 2L in D . melanogaster and the X chromosome in Teleopsis , there is also a significant relationship between gene duplication and residence on the X chromosome in T . dalmanni ( χ2 = 6 . 430 , P = 0 . 011; Chi-squared test ) , but the relationship is weaker because three of the duplicates have moved off the X chromosome in T . dalmanni . Overall , genes that belong to a duplicate set are significantly more likely to move between chromosomes than are other genes . There are 10 occurrences ( 27 . 7% ) of genes in the duplicate sets moving chromosomes within Teleopsis compared to 88 ( 2 . 6% ) chromosomal movements for the remaining genes ( χ2 = 31 . 031 , P<0 . 0001; Chi-squared test ) . Of the 10 duplicate gene movements , 7 are off of the X chromosome onto an autosome , two are off an autosome onto the X , and one is onto the Y chromosome . The EST study also compared the gene expression levels in male T . dalmanni eye-antennal imaginal discs between lines of flies that had been selected for over fifty generations for increased and decreased eye span and found over 350 differentially expressed genes . These genes , however , were not preferentially associated with chromosomal location ( χ2 = 3 . 502 , P = 0 . 174; Chi-squared test ) or gene movement ( χ2 = 2 . 322 , P = 0 . 508; Chi-squared test ) .
Diptera provide one of the most comprehensive model systems for studying the differentiation of sex chromosomes . The regulatory network controlling sex determination in Drosophila is well characterized at the genetic and molecular level [59] . While some aspects of this system appear unique to this clade , there are other features that are highly conserved across Diptera and provide a valuable comparative framework for examining the genetics controlling sex determination in other fly species [60]–[64] . There is also substantial variation in both sex chromosome composition and sex determination systems at various taxonomic levels within the order [59] , [64]–[66] . For instance , Tephritidae include species with isomorphic chromosomes and female heterogamety [67] and , within Musca domestica , the chromosomal location of the male-determining factor varies between populations , occurring in some cases on the X chromosome and , in other populations , on one of several autosomes [68] , [69] . Despite the substantial variation in these systems , a comprehensive catalog of genes on the sex chromosomes exists for only Drosophila and the mosquito , A . gambiae [70] . Therefore the data on chromosomal location presented in this study provide valuable information for reconstructing syntenic relationships and mapping the evolution of chromosomal organization within Diptera . Furthermore , the discovery of a neo-X chromosome in Teleopsis means that these flies represent a critical group in a comparative analysis of sex chromosome organization and provide a complementary model system for understanding key aspects of sex chromosome evolution . Because of the abundant interspecific variation in sex chromosome organization and the paucity of information regarding the specific gene composition of these chromosomes within Diptera , it is difficult to reconstruct the steps that may have led to the formation of the neo-X chromosome in Teleopsis . There is substantial syntenic conservation between the X chromosomes of A . gambiae and D . melanogaster [70] , [71] , flies that share a common ancestor approximately 250 million years ago and represent a more ancestral split than Teleopsis and Drosophila . However , tephritids , which are thought to be a closer relative of diopsids than are Drosophila [72] , appear to have a greatly reduced X chromosome containing few genes and a sex determination system controlled by a male-determining factor [67] . This pattern , combined with the fact that the calyptrate flies examined ( e . g . Musca domestica and Lucilia cuprina ) have a similar condition , suggests the ancestor leading to diopsids may also have had a small X chromosome and male-determining factor . If this is the case , there are a few scenarios that might describe the formation of the neo-X in diopsids . First , the neo-X may represent the fusion of an autosome to small ancestral sex chromosomes . In this scenario , we would predict that some homology would be maintained between the ancestral portion of the sex chromosome of diopsids ( i . e . the region not homologous to 2L ) and chromosomal regions in other flies . There was no syntenic relationship between the non-2L X linked genes in Teleopsis and any chromosomal region in Drosophila , but there may be synteny between these genes and the X chromosomes of tephritids or calyptrates . Second , the male determining factor ( M ) may have moved from a sex chromosome to an autosome ( i . e . the homolog of 2L ) and the sex chromosome creation process started anew . Similarly , M may have existed in a polymorphic state with respect to chromosomal location ( as in Musca ) and become fixed on an autosome at some point . In either case , we would predict a lack of homology between the sex chromosome of diopsids and tephritids or calyptrates and M should exhibit different micro-syntenic relationships within diopsids than tephritids or calyptrates . Finally , a new gene that is located on 2L may have become the primary sex determination signal supplementing M . The evolution of sex-determination pathways is generally characterized by modification of components higher up in the genetic hierarchy [73] , and Shearman [66] has outlined several processes that would result in the acquisition of a new gene in the sex determination pathway . If this scenario applies to stalk-eyed flies , there will be no homology between the primary sex determination gene in diopsids and the M factor of tephritids and calyptrates . Obviously , it is essential for future research to elucidate the sex determination system in Teleopsis and to determine the gene content of sex chromosomes in other dipteran systems such as tephritids and calyptrates . Regardless of the evolutionary transitions in sex chromosome composition leading to Teleopsis , the formation of a neo-X chromosome in this lineage provides an opportunity to examine several important components of sex chromosome evolution , such as the degradation of the Y chromosome and the evolution of dosage compensation . In a few Drosophila lineages , the presence of neo-X chromosomes has provided invaluable snapshots into these processes [53] , [74]–[76] . The neo-X chromosome in diopsids is likely to be more ancient than those in D . miranda [76] or D . americana [77] , but may be similar in age to the neo-X in D . pseudoobscura , which has been estimated at 8–10 million years old [78] . Therefore , Teleopsis offers an alternative system to Drosophila for examining sex chromosome evolution , but the neo-X chromosome in this species is also distinct from those in Drosophila in that it appears to represent a wholesale reconstitution of the X rather than a fusion of an autosomal arm to a substantial preexisting X . Only a single Y-linked gene was identified for T . dalmanni by the CGH microarrays , so it is premature to speculate on the evolutionary history of the Y chromosome in Teleopsis . However , results from this study , along with images of mitotic chromosome lengths [39] , clearly indicate the sex chromosomes in Teleopsis are heteromorphic in males . Therefore , a dosage compensation mechanism is expected to exist in this species . The microarray experiment conducted in Baker et al . [50] compared gene expression between males from different lines , and provides an opportunity to test for the presence of dosage compensation . If the X chromosome in males is hyper-transcribed there should be no difference in the level of expression between autosomal genes and X-linked genes , and , in fact , we find no difference in the average signal intensity between these groups ( P = 0 . 782; Wilcoxon test ) . Intriguingly , none of the male specific lethal ( MSL ) complex genes that control dosage compensation in Drosophila [79] were found in the EST survey of T . dalmanni . While their absence may result from random sampling of the transcript population , it is possible that diopsids , like Sciara [80] , utilize a mechanism of dosage compensation that differs from Drosophila . Transposition in Drosophila and mammals is characterized by an excess of movement off of the X chromosome [8] , [9] , [30] , [31] , [81] , [82] . The derived autosomal copies generally exhibit testes expression [8] , [9] , [30] , [31] , [81] , [82] , and therefore , selection to avoid meiotic sex chromosome inactivation ( MSCI ) has been hypothesized as a primary mechanism driving this pattern [8] , [9] , [30] , [31] , [81] , [82] . It is unknown whether silencing of X-linked genes during late spermatogenesis occurs in stalk-eyed flies . Overall , the reconstruction of gene movement within Teleopsis indicates similar amounts of movement on and off of the X chromosome ( Figure 4 ) . However , there was a striking relationship in the pattern of movement with respect to sexual dimorphism in eyespan . Movement of genes onto the X chromosome are concentrated on branches leading to the most dimorphic taxa while movement off the X is associated primarily with the monomorphic species , T . quinqueguttata ( Figure 4 ) . One possible explanation for this pattern is that sexual selection is operating more strongly on testes function within T . quinqueguttata than the dimorphic species , such that the genes moving off the X chromosome in this species are undergoing selection for increased testes expression . Regardless of the selection pressures affecting T . quinqueguttata , it is noteworthy that the branches leading to the most dimorphic species are characterized primarily by gene movement onto the X chromosome . If these genes play a role in the evolution of eyestalk dimorphism , then they are unlikely to be affected by MSCI . However , sexual antagonism has also been postulated as a factor driving gene movement off of the X chromosome [24] , [25] Therefore , we might expect genes affecting male eyespan to be preferentially located on an autosome but gene movement in the dimorphic species appear directed away from not onto the autosomes . As in Anopheles [83] , gene expression in Teleopsis may not be characterized by a demasculinization of the X chromosome . Consistent with this scenario , the largest QTL influencing variation in male eyespan has been mapped to the X chromosome [40] . Much of the ‘off-of-the-X’ gene movement in Drosophila is driven by duplication events in which the derived copy , often through retrotransposition , moves off the X and subsequently acquires testes expression [9] , [31] . The majority of gene movements reconstructed in Teleopsis cannot be traced to duplication events as complete genomic data is not available for this species . However , for the small number of duplicated genes identified in the EST analysis there was a clear relationship between the duplication process and chromosomal organization . Duplicated genes were concentrated on the 2L/X ( D . melanogaster/T . dalmanni ) chromosome and over 10 times more likely to move between chromosomes than genes that had not been duplicated . Furthermore , the pattern of gene movement for these duplicated genes is consistent with the ‘out-of-the-X’ hypothesis ( 7 movements off of the X versus 2 onto the X ) but the sample size is too limited to draw strong conclusions . Genes that have moved chromosomes also exhibit a faster rate of protein evolution within stalk-eyed flies than genes that have not moved . Theoretical models predict that patterns of divergence of X-linked protein coding genes will be distinct from autosomal genes [4] , and several empirical studies have identified differences in substitution patterns between X-linked and autosomal genes [51] , [84] , [85] , [but see 86]–[88] . Using a measure of protein evolution that calculates divergence in Teleopsis relative to Drosophila and Anopheles [50] , we found no difference in relative divergence between genes on the X chromosome and genes on an autosome . Genes that had moved between chromosomes , however , have higher levels of divergence than those that have not . Despite the extensive research on gene movement in Drosophila , to our knowledge , no study has examined , at the genomic level , the association between gene movement and protein evolution within the genus . There is , however , a positive relationship between male-biased gene expression and rates of protein evolution [23] , [51] . Therefore , if gene movement in Teleopsis is related to the evolution of sexual dimorphism then these genes may be under more intense selection pressures , and evolving faster , than other genes . Ultimately , it will be necessary to measure levels of sex-specific gene expression and quantify rates of evolutionary change before and after translocation in order to fully understand the factors driving this pattern . One methodological issue that may influence the pattern of gene movement within Teleopsis is the effect of sequence divergence on hybridization intensity . Chromosomal location for the three Teleopsis species not including T . dalmanni were inferred from microarray slides with probes designed from T . dalmanni ESTs and the effect of sequence divergence on relative intensity for these species is unknown . Some have suggested that assessment of gene expression levels from cross-species hybridization can be problematic [89] , although the resolution needed for CGH arrays is less than that for detecting differential gene expression . We attempted to mitigate some of these concerns by limiting probe design to protein coding sequence and using the median hybridization value for up to ten probes from a single EST contig . However , the number of ambiguous chromosomal designations and the between-slide correlations clearly indicate that the error associated with measuring hybridization intensity increases as the species becomes more distantly related to T . dalmanni . Therefore , we have taken a conservative approach by only assigning chromosomal location to genes that had consistent signal across all four hybridizations for each species . With respect to the pattern of gene movement in the genus , if probe sequence divergence is influencing hybridization intensities , and thus the reconstruction of gene movement , we would not expect the pattern found in this study where , for two divergent species ( i . e . T . thaii and T . quinqueguttata ) , all the movement is concentrated in one direction for one species and in the other direction for the second species . More likely , gene movement would be randomly distributed among these branches . In addition , for all of the putative gene movement in Teleopsis , the mean rank of relative sequence divergence is higher for the genes where the movement occurs on a branch leading to T . dalmanni than branches exclusive to the other three species , suggesting that the relationship between movement and sequence divergence is not an artifact created by hybridization effects . In the future , it will be essential for new hybridizations examining chromosomal location in other diopsid species to use probes designed from sequence data from that particular species , as well as to develop resources for direct genotyping in these species .
Diopsids are one of several families within a group of higher flies known as Acalyptrata , a paraphyletic group that also contains the fly families Drosophilidae and Tephritidae . Relationships among acalyptrate fly families have proven difficult to resolve and are actively under investigation , but current evidence indicates that the Tephritoideae is the most likely sister group to the Diopsoideae , and that these superfamilies likely had a common ancestor with Drosophila no more than 76 MYA [72] . Four diopsid species in the genus Teleopsis , T . dalmanni , T . whitei , T . thaii and T . quinqueguttata were used in this study . The first three species are all sexually dimorphic with respect to eyespan while T . quinqueguttata is monomorphic . In T . dalmanni , eye span exhibits X-linked inheritance [39] and eye span QTL are tightly linked to a sex-ratio factor that results in female-biased brood sex ratios [44] , presumably as a result of one or more inversions on the X chromosome [40] . As a consequence , males carrying a sex-ratio X chromosome have shorter eye span than other males [40] . They also have shorter sperm [41] . Sex chromosome meiotic drive has been detected in multiple populations of T . dalmanni and T . whitei [44] , [45] in southeast Asia , all of which exhibit some degree of postcopulatory reproductive isolation [90] . Phylogenetic relationships among the four Teleopsis species are well supported and sexual dimorphism in eyespan has been mapped on the phylogeny [91] , [92] . QTL mapping experiments [40] , [93] and visualization of mitotic chromosomes in spermatocytes [39] for T . dalmanni indicate that this species has three major chromosomes , two autosomal chromosomes and one set of sex chromosomes characterized by an acrocentric X and smaller Y chromosome . 60-mer Agilent oligonucleotide probes were designed based on contig sequences from an EST library made from the eye-antennal imaginal discs of T . dalmanni [50] . This study generated over 33 , 000 ESTs that assembled into over 3400 contigs with significant homology to a gene in Drosophila . Probes were designed for nearly all of these genes along with 168 contigs with open reading frames ( ORF ) greater than 500 basepairs . Because we were also conducting hybridization on Teleopsis species other than T . dalmanni , probes were limited to protein coding regions in order to minimize the amount of nucleotide divergence between T . dalmanni and the other species , which tends to be higher in the UTR regions of the transcripts . A maximum of 10 probes were designed for each gene . One rationale for selecting this number of probes is that the CGH results provided valuable information for a subsequent study comparing sex-biased gene expression between T . dalmanni and T . quinqueguttata . For the gene expression study , we wanted probes with the lowest amount of sequence divergence between species and the CGH results provided a means for selecting these probes . The microarray experiment consisted of 4 replicate hybridizations for each species ( the data is available from NCBI via accession number GSE20315 ) . Each hybridization sample consisted of 4 or 5 male or female adult flies that were taken from population cages maintained at the University of Maryland , College Park . After removing wings and heads DNA was extracted from macerated fly bodies with Qiagen DNeasy kits using the insect sample protocol . DNA concentration and quality was then estimated with a Nanodrop ND-1000 spectrophotometer and 3 µg of DNA was used in each sample . Each DNA sample was fractionated by restriction digestion with AluI and RsaI for 2 h at 37°C . Each sample was then labeled with either Cy-3 or Cy-5 and processed according to the Agilent array-based CGH protocol . After hybridization for 24 h at 65°C , arrays were scanned using an Agilent G2539A microarray scanner . Hybridization intensity was measured from array images scanned at each dye wavelength using Agilent's Feature Extraction Software . Features were excluded from further analysis if the majority of pixels were saturated or the median pixel intensity was less than two times the background . Intensity scores were normalized using the linear normalization methodology in the Feature Extraction Software . For each hybridization , we calculated the median log2 ratio ( all log2 ratios were defined as female/male intensity ) of all the probes for a given gene or ORF . Then , the log2 ratios for a given hybridization were centralized by adding or subtracting a constant value to the median log2 ratios so that the peak of the lower distribution was centered over zero . Standard CGH software and analysis techniques are not appropriate for this study because a chromosomal map of the genome does not exist for T . dalmanni . Therefore , after signal extraction and centralization , we conducted two analyses to separate the gene log2 ratios into two categories ( i . e . autosomal or X-linked ) . In the first method , we generated a histogram of the gene log2 ratios averaged across the four hybridizations for each species . Based on this histogram , intervals of size 0 . 1 were searched every 0 . 0125 across the log2 distribution to determine the interval with the fewest entries . The value in the center of this interval was then used as a cut-off for distinguishing chromosomal categories . Using the variation in log2 ratios across the four hybridizations , we then calculated a 95% confidence interval for each gene . If the confidence interval of a given gene did not contain the cut-off value , then that gene was assigned as autosomal if its average log2 ratio value was less than the cut-off and X-linked if its average log2 ratio value was greater than the cut-off . If a gene's confidence interval contained the cut-off value then the gene's location was designated as unknown . The second method we used to separate the probes into groups was K-means clustering [94] . This method performs well when specifying an exact number of clusters and , using the program tmev [95] we conducted 50 iterations of average linkage clustering based on the Euclidian distance to classify the genes into 2 clusters . Overall , the K-means analysis produced chromosomal designations that required about 3 times as much gene movement between the autosomes and the X chromosome within Teleopsis as the first method . Therefore in order to assess the pattern of gene movement relative to Drosophila and within the genus we used the more conservative designations provided by the confidence interval method . Reconstruction of gene movement within Teleopsis was conducted in MacClade ( V4 . 06 ) using a simple parsimony approach . Chromosomal locations in D . melanogaster were used to root the reconstruction of gene movement within Teleopsis assuming a syntenic relationship between chromosome 2L in D . melanogaster and the X chromosome in Teleopsis ( see Results ) . For a few genes that have undergone chromosomal translocation within Drosophila , the basal state for the genus , as reconstructed by Vibranovski et al . [31] , was used . Chromosomal locations in the mosquito , Anopheles gambiae , were used to polarize gene movement on the basal Teleopsis branch that forms a split with D . melanogaster . We assumed a syntenic relationship between chromosome 3R in A . gambiae and the 2L/X chromosome in D . melanogaster/Teleopsis [70] . ‘Unknown’ character states for Teleopsis chromosomal location were assigned the state that minimized the amount of chromosomal gene movement and all genes that produced an ambiguous reconstruction were discarded . A measure of the rate of protein evolution for each gene was taken from the analysis of the T . dalmanni EST database [50] and was based on maximum likelihood trees constructed from amino acid data ( using a JTT substitution model with no invariant sites ) for A . gambiae , three Drosophila species—D . melanogaster , D . pseudoobscura and D . virilis—and T . dalmanni . The index measures the percentage of the entire tree length comprised by the branch leading to T . dalmanni , and was calculated by dividing the length of the branch leading to T . dalmanni by the length of the entire tree . We also used , from the Baker et al . [50] study , data on the number of putative gene duplications with the genus as well as data from a microarray experiment that compared the level of gene expression between males from two artificial lines selected for increased and decreased eye span . Over-representation of gene ontology terms was evaluated with GeneMerge [96] , which uses a hypergeometric distribution and Bonferroni correction for multiple testing , to provide statistical rank scores for numerous functional categories . Statistical analyses were conducted with JMP [97] . Syntenic relationships for identified genes were established after constructing a linkage map based on anonymous microsatellites [98] and annotated EST loci [50] containing variable length tracts of repeated glutamines [99] . We conducted two F2 intercross experiments between different pairs of replicate lines of T . dalmanni selected for either long ( high line ) or short ( low line ) male relative eye span [cf . 40 , 99] . The first intercross was conducted after 32 generations of selection by mating a high line male with a low line female . This cross produced 490 flies , including 231 females and 259 males , in a single F2 family . For this analysis we genotyped 80 females and 95 males at 17 microsatellite and 26 amino acid repeat loci , of which 11 were X-linked and exhibited recombination . The second intercross was conducted after 45 generations of selection by mating a low line male with a high line female . The male used in this cross carried a drive X chromosome , which exhibits very little recombination with a standard X chromosome [40] . We used a single F2 family containing 464 flies including 271 females and 193 males . These flies were genotyped at 20 microsatellite and 18 amino acid repeat loci , of which four were X-linked . For each family we used Joinmap v . 4 . 0 [100] to assign microsatellite and amino acid repeat loci to linkage groups using Kosambi map distances . We used the CP population type for autosomal loci and X-linked loci in females because this option allows for heterogeneous segregation types , resulting from a mixture of homozygous and heterozygous parents , and coded each locus to fit segregation expectations for the corresponding parental genotypes . We used haploid segregation expectations for X-linked loci in males . Because some microsatellite or annotated loci were informative in only one of the crosses , we used Joinmap to produce a joint linkage map , which combines the linkage maps for males and females in both families . A total of 17 microsatellite and 6 annotated loci were genotyped for all flies in both families and provide a common framework for estimating linkage relationships . | The distribution and organization of genes on chromosomes vary widely among animals . Chromosomes can change in number and size , as well as gene composition , over short evolutionary time scales . Furthermore , chromosome location can influence how genes are expressed in various tissues and how they evolve . The sex chromosomes , in particular , have a dynamic impact on gene movement , expression , and evolution . Uncovering the chromosomal location of genes has traditionally been difficult for non-model organism species . In this study , we assess sex chromosome linkage using a new method that hybridizes DNA from males and females to probes representing over 3 , 400 genes in stalk-eyed flies . This technique identifies 533 genes ( 15% ) that are located on the X chromosome with the remaining genes located on two autosomes . Comparison of these genes with their location in Drosophila indicates that the X chromosome in stalk-eyed flies is nearly completely homologous to the autosome 2L in D . melanogaster . This result reveals the formation of a neo-X chromosome in the lineage leading to stalk-eyed flies and indicates that stalk-eyed flies provide a valuable new system to study the origin and evolution of sex chromosomes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"genetics",
"and",
"genomics/comparative",
"genomics",
"genetics",
"and",
"genomics/chromosome",
"biology",
"evolutionary",
"biology/genomics"
]
| 2010 | Comparative Genomic Hybridization (CGH) Reveals a Neo-X Chromosome and Biased Gene Movement in Stalk-Eyed Flies (Genus Teleopsis) |
The granulomatous lesion resulting from infection with the fungus Paracoccidioides brasiliensis is characterized by a compact aggregate of mature cells , surrounded by a fibroblast- and collagen-rich content . Granuloma formation requires signaling elicited by inflammatory molecules such as members of the interleukin-1 family . Two members of this family have been thoroughly studied , namely IL-1α and IL-1β . In this study , we addressed the mechanisms underlying IL-1α secretion and its functional role on the host resistance to fungal infection . We found that , the expression of caspase-11 triggered by P . brasiliensis infection of macrophages depends on IFN-β production , because its inhibition reduced procaspase-11 levels . Curiously , caspase-11 deficiency did not impair IL-1β production , however caspase-11 was required for a rapid pore-mediated cell lysis . The plasma membrane rupture facilitated the release of IL-1α , which was necessary to induce NO production and restrict fungal replication . Furthermore , P . brasiliensis-infected macrophages required IL-1α to produce optimal levels of IL-6 , a major component of Th17 lymphocyte differentiation . Indeed , IL-1α deficiency accounted for a significant reduction of Th17 lymphocytes in lungs of infected mice , correlating with diminished neutrophil infiltration in the lungs . Strikingly , we identified that IL-1α directly reprograms the transcriptional profile of Th17-committed lymphocytes , increasing cellular proliferation , as for boosting IL-17 production by these cells . Beyond neutrophil chemotaxis in vivo , IL-17 also amplified IL-1α production by infected macrophages in vitro , endorsing a critical amplification loop of the inflammatory response . Therefore , our data suggest that the IFN-β/caspase-11/IL-1α pathway shapes a protective antifungal Th17 immunity , revealing a molecular mechanism underlying the cross-talk between innate and adaptive immunity .
During pulmonary Paracoccidioides brasiliensis infection , the granulomatous inflammation is a crucial process to control dissemination and prevent systemic chronic paracoccidioidomycosis ( PCM ) . Concerted efforts of both innate and adaptive immune cells are necessary for fungal recognition and elimination by the host . However , the same mechanisms that destroy the pathogen may also damage the host and exacerbate the disease [1] . Deregulated immunity and tissue remodeling arising from a persistent fungal stimulus are major pathological features of this infection [2] . Resistance to this fungus is primarily mediated by Th1 immunity , while susceptibility is associated with an imbalance towards Th2 response . Nonetheless , cells expressing interleukin-17 ( IL-17 ) , such as Th17 lymphocytes , have been detected within and around the granulomas in the skin and oral mucosa lesions from PCM patients [3] . Indeed , IL-17 exerts important roles during P . brasiliensis infection [4–6] . Macrophages produce diverse pro-inflammatory mediators that initiate and maintain granulomas . Among them , IL-1 signaling has a well-determined function in regulating the recruitment and activation of cells in inflamed tissues [7 , 8] , but the exact contribution of different members of the IL-1 superfamily to this process still needs to be elucidated . The IL-1 family is now comprised by 11 members , which exhibit complimentary or distinct biological functions [9 , 10] . The most well-studied cytokines from this family , IL-1α and IL-1β , bind to the type I IL-1 receptor ( IL-1RI ) [11] , but differ in their maturation processes . In contrast to IL-1β , IL-1α does not require proteolytic processing by caspase-1 inflammasome to become biologically active [12] , whereby immature IL-1α also binds to IL-1RI [13] . Nevertheless , because pro-IL-1α lacks a signaling peptide that mediates its secretion from the cell , inflammasome activation can promote IL-1α release under certain circumstances , mostly as a result of cell death [14 , 15] . The inflammasome-mediated cell death , termed pyroptosis , is characterized by an immediate pore formation in the cell membrane that increases its permeability [16 , 17] . Integrity loss of plasma membrane , osmotic potential imbalance , cell swelling , and eventual bursting is followed by the passive leakage of cytosolic pro-inflammatory molecules from dying cells [18] . Although pyroptosis was initially identified as a caspase-1 inflammasome-dependent process , caspase-11 activation also elicits membrane rupture and cell death [19 , 20] . Extracellular release of cytoplasmic components fuels local inflammation as an alert of tissue damage . In fact , bystander cells respond to IL-1α by releasing chemokines and inducing leukocyte infiltration to clear pyroptotic macrophages and avoid excessive inflammation [21 , 22] . IL-1α has been shown to play significant roles during fungal diseases [23 , 24] . However , the molecular mechanisms that control IL-1α secretion and resistance to pathogenic fungi remain poorly understood . Here , we demonstrate that caspase-11 operates downstream of IFN-β to promote pyroptotic cell death and IL-1α release by bone marrow derived macrophages ( BMDMs ) . IL-1α mediates effective production of nitric oxide ( NO ) and IL-6 , enhancing control of fungal replication by BMDMs . Moreover , this cytokine reprograms the transcriptional profile of Th17 committed lymphocytes , increasing cellular proliferation and IL-17 production . Importantly , IL-1α boosts the Th17 response in vivo , promoting adequate neutrophil recruitment into the lung , which reduces fungal burden and confers resistance to the infection . Interestingly , IL-17 stimulation of P . brasiliensis-infected BMDMs significantly increased IL-1α production , indicating an inflammatory loop between IL-1α and the Th17 response . Collectively , the positive effects of IL-1α over the Th17 response and resistance to P . brasiliensis infection reveal unrecognized molecular mechanisms at the interface of the innate and adaptive immunity to fungi .
Caspase-11-containing inflammasome has been described across a broad range of infections , especially with gram-negative bacteria [25 , 26] . To address the molecular mechanisms governing caspase-11 function during fungal diseases , we investigated whether P . brasiliensis triggers its activation in macrophages . We observed that infected WT BMDMs express both procaspase-11 and cleaved caspase-11 , in which the p30 subunit ( cleaved caspase-11 ) was evident in cells infected with a MOI of 5 ( Fig 1A ) . Next , we assessed the internalization of the membrane-impermeant dye EtBr to evaluate the requirement of caspase-11 for inducing pores in host cell membranes ( pyroptosis ) . The fungal infection clearly induces EtBr internalization by WT BMDMs , demonstrating that host cells display pores in their membranes after interaction with fungi cells ( Fig 1B ) . However , Casp11-/- BMDMs were significantly less permissive to EtBr , confirming a requirement for caspase-11 on P . brasiliensis-induced pore formation ( Fig 1B ) . Using propidium iodide internalization assay and flow cytometry , we also found that the membrane integrity was maintained in BMDMs derived from Casp11-/- mice but lost in WT cells ( Fig 1C ) . To corroborate that host cell pore-formation and reduced membrane integrity indeed reflect pyroptosis , we quantified LDH levels in cell culture supernatants after fungal infection . Compared to WT BMDMs , LDH levels were significantly reduced in Casp11-/- BMDMs culture supernatants ( Fig 1D ) . Interestingly , the caspase-11-mediated lysis requires P . brasiliensis to be alive , once heat-killed yeasts did not stimulate significant LDH release ( Fig 1E ) . These results indicate that caspase-11 activation by P . brasiliensis is essential for pore-mediated cell lysis and pyroptosis . Next , we investigated whether the fungus-induced caspase-11 activation and host cell pyroptosis are followed by passive IL-1α release . To achieve this , we used BMDMs from mice deficient for caspase-11 and gasdermin D , a 53 kDa protein that generates an N-terminal pore-forming fragment after inflammasome-dependent activation of caspase-1 or -11 [19 , 20 , 27–29] . We found that IL-1α levels were significantly reduced in cell culture supernatants from both infected Casp11-/- and Gsdmd-/- BMDMs ( Fig 2A ) , suggesting that caspase-11 and pyroptosis are vital to ensure optimal IL-1α release . Caspase-1 and caspase-8 are required for IL-1β processing and secretion during this fungal infection [30 , 31] . Thus , we tested whether caspase-11 is also involved in IL-1β release in this context . WT and Casp11-/- BMDMs incubated with P . brasiliensis synthesized and secreted similar amounts of IL-1β , demonstrating that caspase-11 activation is only necessary for IL-1α secretion ( Fig 2B ) . Detailed time course analysis of IL-1β secretion by infected WT and Casp11-/- cells revealed that caspase-11-deficient BMDMs behave similarly to WT cells . Even pondering different timepoints , IL-1β release occurred in a caspase-11-independent manner ( S1A Fig ) , excluding that the IL-1β might peak at a time other than 48 hours . Importantly , levels of IL-1β were similar in the lungs of WT and Casp11-/- mice at 15 and 30 dpi , validating the in vitro data ( S1B Fig ) . P . brasiliensis infection of macrophages increased levels of IFN-β over time ( Fig 2C ) , whereas IFN-β induces caspase-11 expression and activation [32] . To understand the upstream events of caspase-11 activation , we explored the role of IFN-β during P . brasiliensis-induced IL-1α release . Incubation of WT BMDMs with IFN-β-neutralizing antibody at the same time of the infection markedly reduced procaspase-11 expression ( Fig 2D ) . As expected , IL-1α release was also impaired in these BMDMs ( Fig 2E ) . Moreover , caspase-11 expression ( Fig 2F ) and IL-1α release ( Fig 2G ) were strongly enhanced by addition of rIFN-β to the culture 2 hours prior P . brasiliensis infection . Collectively , the data point to a model where the fungus induces IFN-β production , whose signaling promotes procaspase-11 expression , cleavage and activation of caspase-11 , resulting in pyroptosis and IL-1α release to the extracellular environment . To determine the role of caspase-11 during host responses in vivo , we infected WT and Casp11-/- mice with P . brasiliensis . Confirming the findings obtained in cell culture experiments , caspase-11-deficient animals exhibited reduced pulmonary levels of IL-1α after 15 and 30 dpi compared to WT mice ( Fig 3A ) . Casp11-/- mice also harbored increased fungal burden in the lung at the same time of infection ( Fig 3B ) . However , survival of infected Casp11-/- mice remained similar to those of WT mice ( Fig 3C ) . In accordance , in situ Grocott staining of fungi cells showed that infected Casp11-/- mice were unable to limit the fungal growth in the lung tissue at 30 dpi ( Fig 3D ) . We presumed that elevated fungal burden would be associated with an intense pulmonary inflammation . However , despite the increase in the inflammatory infiltrate , H&E-stained histological sections of infected Casp11-/- mice revealed well-organized granulomas with a few diffuse inflammatory reactions in the pulmonary parenchyma ( Fig 3E ) . These data demonstrate that , although dispensable for granuloma formation , caspase-11 plays a significant role in the control of P . brasiliensis infection . Reduced levels of IL-1α in the lungs of infected Casp11-/- mice along with the lower expression of inducible nitric oxide synthase ( iNOS ) ( Fig 3F ) , suggest that effector mechanisms downstream caspase-11 activation could be mediated by IL-1α and nitric oxide ( NO ) , an important antifungal compound . Thus , we tested the fungal killing capacity of Casp11-/- BMDMs and the effects of exogenous rIL-1α . Compared to WT BMDMs , infected Casp11-/- cells exhibited increased fungal loads ( Fig 3G ) . However , rIL-1α rescued the fungicide activity of Casp11-/- cells ( Fig 3G ) , correlating with the increased production of NO ( Fig 3H ) . In addition , the administration of recombinant IL-1α to Casp11-/- infected mice reversed their pulmonary fungal load , which was similar to that of the WT phenotype ( Fig 3I ) . These data indicate that caspase-11 regulates IL-1α release to limit fungal replication by macrophages . To better characterize the biological role of IL-1α , we first analyzed the kinetics of IL-1α production over the time of fungal infection . In vitro , WT BMDMs initiated IL-1α secretion after 12 h , and increased until 72 h ( Fig 4A ) . In vivo , pulmonary levels of IL-1α increased in the first 7 days of infection but declined to baseline levels at 15 dpi ( Fig 4B ) . Next , we evaluated the biological function of IL-1α in PCM by following survival of P . brasiliensis-infected WT and Il1a-/- mice . Strikingly , 100% of Il1a-/- animals succumbed to infection , in contrast to 30% of WT mice ( Fig 4C ) . We also recovered more fungal colonies in the lungs , but not livers of Il1a-/- animals after 15 days of infection ( Fig 4D ) . However , Il1a-/- mice showed higher fungal counts in the lungs and livers at 30 dpi ( Fig 4D ) . Gomori staining of lung tissue sections confirmed that IL-1α deficiency suppresses pulmonary fungal clearance ( Fig 4E ) , providing further evidence for a protective role of IL-1α in the control of P . brasiliensis infection . The granulomatous inflammation is a crucial process to control P . brasiliensis dissemination and prevent systemic disease . IL-1 signaling has been shown to promote efficient granuloma formation [7] , which would explain higher susceptibility of Il1a-/- mice to the fungal infection . However , caspase-11 does not seem to be important for this process ( Fig 3E ) , suggesting that the caspase-11/IL-1α axis is not critical for the structural formation of granulomas in this context . Indeed , we observed that the elevated number of cells , presented in IL-1α-deficient mice after 15 and 30 dpi , was arranged in organized and well-formed granulomas ( S2A Fig ) . Furthermore , Il1a-/- animals presented a large number of reticulin fibers delimiting the peripheral area and compartmentalizing the granulomatous structures ( S2B Fig ) . These data indicate that the caspase-11/IL-1α pathway promotes resistance to P . brasiliensis by mechanisms other than enhancing the granulomatous response . Exogenous rIL-1α enhanced NO production and fungal clearance by Casp11-/- cells ( Fig 3G and 3H ) , showing that IL-1α drives efficient NO production by macrophages in response to P . brasiliensis . Moreover , BMDMs from Il1a-/- mice incubated with P . brasiliensis produced an average of 19 μM of nitrite , contrasting with the 31 μM produced by WT cells ( S3A Fig ) . This data correlated with reduced fungicide capacity of Il1a-/- cells ( S3B Fig ) . Even after activation with IFN-γ , IL-1α deficiency impaired NO production and yeast clearance ( S3A and S3B Fig ) . Importantly , the addition of rIL1-α at 50 ng/mL to Il1a-/- cell culture increased NO synthesis by 210% ( Fig 4F ) and reduced fungal burden by approximately 70% ( Fig 4G ) . Accordingly , the importance of IL-1α in driving NO production in response to P . brasiliensis was also evident in vivo . Compared to WT animals , there was a down-regulation of Nos2 gene expression in the lungs of Il1a-/- mice ( S3C Fig ) . These results show that IL-1α regulates NO production by macrophages , coordinating an important fungicide mechanism . Several studies have described that Th1 and Th17 responses are implicated in the host resistance to P . brasiliensis [4 , 33] . Interestingly , a fungal infection of Il1a-/- BMDMs resulted in reduced IL-6 production , whereas exogenous rIL-1α rescued IL-6 levels ( Fig 4H ) . Considering the pivotal role of IL-6 for Th17 lymphocyte differentiation [34 , 35] , we hypothesized that IL-1α deficiency would particularly compromise the Th17 cell-mediated immunity . In fact , infected Il1a-/- mice exhibited reduced transcriptional and protein levels of IL-17 ( Fig 5A and 5B , respectively ) but not of IFN-γ ( Fig 5C and 5D ) in lungs . Corroborating this data , Il1a-/- animals exhibited reduced frequency and number of IL-17-expressing CD4+ T cells in the lung at 30 days post P . brasiliensis infection ( Fig 5E ) . Importantly , infection of Casp11-/- mice with P . brasiliensis recapitulated all aspects of the Th17 response observed in the lung of Il1a-/- animals at 30dpi ( S4A–S4C Fig ) . Treatment of Casp11-/- mice with rIL-1α not only reduced CFU counts ( Fig 3I ) , but also increased IL-17 levels in the lung at 30 dpi , supporting a role for this cytokine in boosting the Th17 response ( S4D Fig ) . Together , these data strongly support a model in which IL-1α is released due caspase-11 activation and pyroptosis to promote efficient antifungal Th17 immunity . A key question that arises from these results is whether IL-1α-deficiency impairs Th17 cell migration from mediastinal lymph nodes to the lung . However , we observed that IL17+ T cells did not accumulate in the lymph nodes collected from infected Il1a-/- mice ( S5A Fig ) . In addition , a disruption of the chemoattractant axis CCL20/CCR6 in Il1a-/- mice could explain the reduced Th17 response after 30 days of infection . Despite increased transcriptional and protein levels of CCL20 in the lung of Il1a-/- animals ( S5B and S5C Fig ) , CCR6 expression in Th17 cells was similar between infected Il1a-/- and WT mice ( S5D Fig ) . Finally , Th17 responses promote a strong neutrophil influx into sites of inflammation [36–38] , suggesting that IL-1α deficiency impacts the neutrophilic response against the fungus . Accordingly , we found that decreased IL-17A production correlated with a significant reduction in neutrophil recruitment to the lungs of infected mice ( Fig 5F ) . We confirmed this result by showing that Il1a-/- animals infected for 30 days express lower amounts of Ly6G ( neutrophil surface marker ) in the lung tissue compared to WT mice ( Fig 5G ) . Collectively , these data demonstrate that caspase-11 and IL-1α are critical for the development of a productive Th17 response to P . brasiliensis infection . In view of the significant impact of IL-1α deficiency over the Th17 lymphocyte compartment , we isolated CD4+ T cells from WT and Il1a-/- mice and cultured them under Th17-polarizing conditions to rule out the possibility that IL-1α directly influences Th17 cell differentiation . In both groups , CD4+ T cells differentiated equally well for the Th17 profile ( S6A and S6B Fig ) . Hence , we conclude that the decline of IL-17+ T cell population in Il1a-/- mice was not due an impaired intrinsic capacity to differentiate into Th17 lymphocytes . To understand the immune-modulatory effects of IL-1α over the Th17 response , isolated CD4+ T cells were polarized to the Th17 profile , whereby we added rIL-1α to these cells on the 3rd day of differentiation . As a readout for the effectiveness of the Th17 cell differentiation , we confirmed that these cells were unable to produce IFN-γ ( S6C and S6D Fig ) . Surprisingly , the addition of rIL-1α elevated the percentage of IL-17 producing cells ( Fig 6A ) , also visualized by increased IL-17 mean fluorescence intensity ( MFI ) ( Fig 6B ) . Furthermore , addition of rIL-1α also amplified mRNA ( Fig 6C ) and protein levels of IL-17 ( Fig 6D ) , showing that IL-1α boosts IL-17 production by CD4+ T cells that are already committed to the Th17 subset . Because IL-1 signaling promotes T cell proliferation [39] we analyzed the expansion capacity of IL-1α-treated Th17 cell by evaluating the cellular proliferation ratio of CFSE-labeled cells after 5 days of anti-CD3 plus anti-CD28 stimulation . Th17 committed CD4+ T cells incubated with IL-1α displayed greater cellular division compared to conventional Th17 cells ( Fig 6E ) . This suggests that the reduced frequency and number of CD4+ IL-17+ T cells in the lung of infected Il1a-/- animals are affected by an impaired proliferative capacity . Beyond increasing cellular proliferation and IL-17 production , IL-1α could induce a molecular reprogramming of Th17 cells to enhance their effector functions . To address this question , we profiled the expression of Th17-related genes in naïve CD4+ T cells ( Th0 ) , conventional Th17 cells and conventional Th17 cells + IL-1α . Principal component analysis ( PCA ) , including the whole array of genes , demonstrates that each of the conditions acquired distinct transcriptional profiles ( Fig 6F ) . Furthermore , the direct comparison between the three conditions revealed the differential expression of several genes . By grouping 20 of these genes based on their biological functions , we observed that genes associated with Th17 differentiation such as Il1r1 , Il23r , Il6ra , Il23r , and Rorc were equally expressed by conventional Th17 cells and IL-1α-treated Th17 cells ( Fig 6G ) . However , IL-1α upregulated Ccr6 , Ifng and Tgfb1 expression ( Fig 6H ) , while downregulating the expression of genes coding for IL-17 inhibitors , such as Socs1 and Il27 ( Fig 6I ) . Notably , IL-1α downregulated the expression Stat4 , Il7r , Il10 , and Icos , while upregulating Il9 ( Fig 6H ) . Taken together , these results demonstrate that IL-1α promotes transcriptional and functional adaptations of Th17 cells . Of note , we observed that IL-17 enhances IL-1α production by BMDMs infected with P . brasiliensis ( Fig 6J ) . To validate that the decreased Th17 response is responsible for the increased susceptibility of IL-1α deficient mice , we treated P . brasiliensis-infected Il1a-/- mice intranasally with 500 ng of rIL-17A . We found that the protective immunity was restored by IL-17A treatment , which significantly reduced the pulmonary fungal load when compared to untreated Il1a-/- mice at 30dpi ( Fig 6K ) . This indicates that IL-α reprograms Th17 lymphocytes to produce high levels of IL-17 , which promote an inflammatory loop by sustaining IL-α production and enhancing effector antifungal mechanisms by innate cells . Taken together , these results demonstrate that IL-1α is released via IFN-β/caspase-11 axis to shape Th17 immunity and resistance against P . brasiliensis infection .
Murine caspase-11 was discovered in 1993 by Yuan et al . [40] owing on its similarity to caspase-1 . The 5'UTR promoter region , which regulates the expression of caspase-11 gene , has multiple NF-κB-binding sites but only a single binding site for STAT [41] . It has been reported that caspase-11 is activated , independently of TLR4 signaling [42 , 43] , mainly in response to the lipid A contained in the LPS from gram-negative bacteria [32 , 42–45] . Compelling findings with Candida albicans [46] and Aspergillus fumigatus [47] demonstrate that the activation of the non-canonical caspase-11 inflammasome pathway is not limited to gram-negative bacteria . In a model of candidiasis , the host cell recognizes , engulfs and phagocytizes the fungus . While the phagosome matures , the fungus senses and responds to entry into the host cell initiating filamentation and cell wall biosynthesis [48 , 49] . The remodeling of its surface exposes moieties , mainly ergosterol , capable of inducing macrophage lysis via inflammasome activation [50 , 51] . Our study highlights the importance of caspase-11 in a non-bacterial infection , integrating the IFN-β/caspase-11 axis with IL-1α release and resistance to P . brasiliensis challenge . We have demonstrated that P . brasilensis-induced IFN-β activates caspase-11 to release IL-1α through a gasdermin D-dependent cell death . Nonetheless , we cannot discard that ergosterol , one of the most prevalent lipids found in P . brasiliensis vesicles [52–54] , is also involved in the macrophage toxicity . Caspase-11 induces cell death by promoting the proteolysis of the 53 kDa protein gasdermin D . Cleavage of human and mouse gasdermin D at Asp275 residue generates a 31 kDa N-terminal fragment that oligomerizes and forms ~20nm pores in the plasma cell membrane [19 , 20 , 27 , 28] , which become permeable to low molecular weight dyes such as EtBr and PI [55 , 56] . Caspase-11 was required for EtBr and PI incorporation and LDH release by P . brasiliensis-infected macrophages , indicating pyroptotic cell death . The death of P . brasiliensis-infected cells allows the leakage of IL-1α in a caspase-11 and gasdermin D-dependent manner . The lysis of macrophages is a process that has also been widely observed for the opportunistic pathogenic fungus C . albicans [57 , 58] . Caspase-11-mediated pyroptosis seems to be actively induced by P . brasiliensis , because dead fungal cells were unable to stimulate this process in macrophages . This could also mean that dead yeasts lack molecules that are recognized by innate immune receptors and necessary to trigger this pathway . Likewise , it has been reported that the lysis of macrophages also requires live C . albicans that forms hyphae , which stress phagosomal membranes [59] , disrupt the macrophages with their associated toxin [57] and compete with the host for glucose [60] . In this context , a pertinent question that comes out of these findings is whether the induced pyroptosis benefits the host , the fungus or both . For instance , activation of pyroptosis in response to Candida might serve to augment proinflammatory responses , but C . albicans hijacks host pathways of programmed cell death for macrophage killing and evasion of the immune response [61] . However , considering that impaired pyroptosis in Casp11-/- macrophages correlates with elevated fungal burden in vitro and in vivo , our study provides evidence for caspase-11-mediated pyroptosis as a critical process for resistance against the infection . Endogenous danger signals released due to caspase-11-mediated pyroptotic death could activate the NLRP3 inflammasome and IL-1β/IL-18 production , a platform that plays a protective role during experimental PCM [30] . Because pore formation usually happens earlier than IL-1β production and processing , pyroptosis has been proposed as a mechanism by which IL-1β is secreted [18] . Nevertheless , our data do not support this hypothesis . Even though infected Casp11-/- macrophages displayed reduced pore formation in cell membranes , IL-1β levels were still detected in vitro and in vivo . This suggests that pore formation is not an essential prerequisite for IL-1β secretion . Corroborating this concept , intact IL-1β processing and release in caspase-11-deficient mice was also reported in aspergillosis [47] . However , it is still possible that pores formed in response to caspase-1 activation in Casp11-/- cells are enough for IL-1β secretion . Greater susceptibility of Casp11-/- cells to P . brasiliensis infection was associated with reduced IL-1α production , because the treatment of Casp11-/- mice with recombinant IL-1α protected the mice from fungal replication when compared to non-treated knockout mice . Nevertheless , diminished levels of pulmonary IL-1α in Casp11-/- mice did not correlate with reduced survival , once these animals exhibited mortality rates comparable to that of WT mice after P . brasiliensis challenge . Because the production of IL-1α was not completely abrogated in Casp11-/- mice , we speculate that alternative mechanisms [62] could mediate the release of this cytokine during P . brasiliensis infection . Modest IL-1α levels in Casp11-/- mice might attenuate mortality , but they are not sufficient to control fungal growth in the lung . Moreover , studies with A . fumigatus have suggested that caspase-11 induces actin-mediated phagosomal killing to control fungal dissemination in vivo [63 , 64] . Therefore , caspase-11 might also regulate IL-1α-independent pathways and interfere with fungal colonization and growth in other organs . Further studies are needed to confirm this hypothesis . In contrast to IL-1β , unprocessed IL-1α is also active , but pro-IL-1α can be cleaved by calpains [65 , 66] , independently of caspase-1 activity , in a C-terminal fragment with a calculated mass of 18kDa [12] . Our study demonstrates that caspase-11 coordinates IL-1α release in response to P . brasiliensis , whereas IL-1β maturation and secretion depends on different molecular pathways [30 , 31] . However , we have not addressed whether IL-1α was cleaved in the supernatant or the cell lysates after P . brasiliensis infection . Because many activators of IL-1α secretion , described by Tschopp group [12] , are inducers of Ca2+ influx and patients with disseminated PCM present hypercalcemia [67] , we suppose that calpain-like proteases can also influence IL-1α cleavage after P . brasiliensis stimulation . Given that IL-1α and IL-1β bind and activate the same receptor [13] , they should in principle have identical biological functions . However , both IL-1 cytokines exert non-redundant roles on the host immune response against different pulmonary pathogens such as A . fumigatus , Histoplasma capsulatum and Cryptococcus neoformans [23 , 68 , 69] . In systemic candidiasis , for example , IL-1α increases leukocyte antifungal activity , while IL-1β recruits neutrophils [24] . Herein , we showed that IL-1α-deficiency led to exacerbated P . brasiliensis replication and a non-resolving granulomatous inflammation . Accordingly , our previous work revealed that , P . brasiliensis-infected Il1r1-/- animals harbor more yeasts inside organized and compact granulomas [30] . These similarities with IL-1α-knockout mice indicate a prominent role for IL-1α in activating IL-1RI signaling . Nonetheless , we do not discard that these events could also be mediated by IL-1β . Future research using Il1b-/- and Il1a/Il1b-/- mice will provide deeper understanding about what the role of IL-1β in PCM . During respiratory tract infections , the early expression of IL-1α by CCR2+ monocytes precedes the recruitment of neutrophils to the lung [21 , 23] . In addition , cytosolic extract of necrotic WT cells , but not from Il1a-/- cells , induced recruitment of neutrophils via CXCL1/CXCR2 [70] . We found reduced neutrophil chemotaxis to the lung of infected Il1a-/- mice . Equivalent to intratracheal LPS injection model [71] , one possibility is that IL-1α absence prevents tissue extravasation of the circulating neutrophils as a result of the greater cellular adhesion to the vascular endothelium [72] . However , our data point to a major role for IL-1α in shaping antifungal Th17 immunity , which boosts IL-17 production to regulate neutrophil recruitment into sites of infection . We have found that the treatment of Casp11-/- mice with rIL-1α not only reduced CFU counts , but also increased IL-17 levels in the lung . Furthermore , macrophages increase the expression of IL-17A receptor in response to an inflammatory stimulus [73] , while we showed that IL-17 amplifies the production of IL-1α by P . brasiliensis-infected macrophages and reverses the susceptibility of Il1a-/- mice to P . brasiliensis infection . Dendritic cells ( DCs ) infected with Leishmania amazonensis preferentially drive CD4+ T cells to an IFN-γlowIL-17high profile , because they secrete more IL-1α than IL-1β , besides being weak IL-12p40 producers [74] . Following this logic , M-CSF-differentiated macrophages expressing membrane-bound IL-1α converts IL-17- CD4+ T memory cells into conventional Th17 cells [75] . Nonetheless , IL-1α-deficiency did not affect the intrinsic capacity of naïve CD4+ T cells to differentiate into Th17 cells . Instead , we observed that IL-1α favors Th17 cell expansion , correlating with reduced amounts of Th17 cells in the lungs of infected Il1a-/- mice . Previous study showed that long-term Th17 cells are maintained by IL-1 [76] , while these cells are completely absent in IL-1R knockout mice [77] . Thus , IL-1α promotes Th17 cell survival during infection , which is essential for the host immune defense against fungi [78] . Importantly , IL-1α also induces IL-6 expression and its suppression reduces IL-6 synthesis [79 , 80] . In our murine model of P . brasiliensis infection , impaired IL-6 production seen in infected IL-1α-knockout cells could have a major effect on Th17 differentiation . However , it is important to note that the cytokine requirement for Th17 cells polarization differs substantially between mouse and human . The conditions that regulate the development of mouse Th17 cells in vitro and in vivo are well defined [81] . It is widely accepted that Th17 generation in mice is commonly induced by synergistic treatment with TGF-β and IL-6 [82–84] . While TGF-β is required for initiation , IL-6 promotes STAT3-mediated activation of RORγt as well as alleviates Foxp3-mediated inhibition of the Th17 program [82 , 85] . However , IL-6 does not seem to be absolutely necessary for human Th17 cultures [86] . Alternatively , IL-21 , activating STAT3 , can act with TGF-β to generate human Th17 cells [87] . The cytokines IL-23 and IL-1β are most consistently found to play crucial roles in both mouse and human Th17 development [88–91] . The IL-1β dependency for Th17 differentiation has been already reported in Candida infection [91 , 92] . Here , we demonstrate that IL-1α induces a transcriptional reprograming of cells already committed to the Th17 phenotype . In synergy with IL-23 , IL-1β would control RORγt and IL-23R expression to sustain IL-17 production by effector Th17 cells [77 , 88] . However , IL-1α seems to exert its effects independently of changes in RORγt and IL-23R abundance . Indeed , IL-1α suppresses the expression of inhibitory molecules , such as Socs1 and IL-27 , and enhances the levels of TGF-β1 transcripts , a cytokine that ( in mice ) collaborates with IL-6 to promote T helper cell differentiation into Th17 profile [93] . Remarkably , Th17 cells stimulated with IL-1α simultaneously co-express IL-17A and IFN-γ , but not IL-10 . Because A . fumigatus-exposed human PBMCs stop producing IL-10 to upregulate IL-17 [94] , IL-1α possibly disrupts IL-10 /IL-17 balance , enabling IL-17 synthesis . Also , the flexibility of Th17 cells to acquire effector functions that are normally associated with Th1 responses such as IFN-γ production has been reported in C . albicans and A . fumigatus -primed cultures [91 , 95] . Their shift to a Th1 cell-like phenotype has been linked to the pathogenicity of Th17 cells in colitis [96] , arthritis [97] , diabetes [98] , and experimental autoimmune encephalomyelitis ( EAE ) [99] . According to Wang et al . , 2014 [100] , the development of IL-17/IFN-γ double producing cells is dependent on T-bet , Runx1 , and Runx3 transcription factors . Nevertheless , the shift in cytokine production under IL-1α stimulation was not accompanied by an increase in Tbx21 mRNA levels . We speculate that the expression of IFN-γ by Th17 cells collaborates for effective antifungal immunity because IFN-γ is important for host resistance against fungal cells [101] . IFN-γ induces P . brasiliensis-infected macrophages to secrete TNF-α , required for the development and persistence of well-formed granulomas [102] , and NO , which plays a well-documented role in fungal clearance [103] . Overall , this study discloses a pathway in which IL-1α is released due to caspase-11-dependent pyroptosis to enhance effector functions of innate cells and T lymphocytes . Further investigations shell reveal whether this molecular mechanism translates to human cell responses to P . brasiliensis and how it can be explored to design novel immunotherapeutic interventions .
For experimental infection and isolation of bone marrow macrophages ( BMDMs ) , six to seven-week-old male Casp11-/- ( kindly provided by Vishva Dixit , Genentech ) and Il1α-/- ( kindly provided by Bernhard Ryffel , University of Orleans , Orleans , France ) and strain-matched wild-type ( C57BL/6 ) mice were obtained from the Isogenic Breeding Unit , Ribeirão Preto Medical School , University of São Paulo , Ribeirão Preto , Brazil . Mice were bred and maintained under specific pathogen-free conditions and provided with clean food and water ad libitum in the animal housing facility of the Department of Biochemistry and Immunology . Mice experiments were conducted in compliance with the institutional guidelines on ethics for animal care approved by the Ethical Commission in Animal Research ( approved protocol no . 218/2017 ) , which follows the Brazilian National Guidelines recommended by the Brazilian Society of Science in Animals Laboratory . Yeast cells from a P . brasiliensis-virulent strain ( Pb18 ) were cultured for 7 days at 37°C in brain heart infusion ( BHI ) -agar medium ( Sigma ) supplemented with gentamicin ( 96 μg/mL ) , and 5% fetal bovine serum ( FBS ) ( Gibco ) . Pb18 cells were harvested and incubated overnight under agitation in F12 Coon's Modification medium ( Sigma ) at 37°C . Cell viability was determined by the fluorescein diacetate-ethidium bromide method , as previously described [104] . Only fungal suspensions containing more than 90% viable cells were used . For in vivo infection , six to seven-week-old male mice were anesthetized with ketamina ( 90mg/kg ) and xylazine ( 5mg/kg ) by intraperitoneal ( i . p . ) administration , followed by intravenous ( i . v . ) inoculation of 1x106 yeast cells . The numbers of colony-forming units ( CFU ) in the organs were calculated at 15 and 30 days post-infection ( dpi ) by normalizing the count per gram of tissue . The survival of the Pb18-infected wild-type and knockout mice ( 10–12 mice of each group ) was verified daily for 120 days . Lungs from Pb18-infected mice , were removed , weighed , triturated in sterile PBS-containing protease inhibitor ( Complete , Roche ) , and centrifuged . Supernatants were collected and stored at −20°C . Levels of IL-1α , IL-17A , IFN-γ and , IL-6 were measured by enzyme-linked immunosorbent assay ( ELISA ) according to manufacturers’ recommendations ( BD Pharmingen ) . The reading was held in eMax ELISA reader ( Molecular Devices ) at 450 nm . Animals selected at random from each group were euthanized at 15 and 30 days after infection . The lungs were excised , fixed with 10% formalin for 24 hours and embedded in paraffin . Tissue sections ( 5 μm ) were stained with hematoxylin and eosin ( H&E ) for analysis of the lesions or impregnated with silver for fungal cell labeling ( Gomori method ) and reticulum fiber staining . The nuclei percentage and the integrated density of P . brasiliensis were estimated using Image J software . Lungs from each mouse were excised , minced with scissors , and enzyme-digested at 37°C for 30–35 minutes in 1 mL of digestion buffer ( 2 mg/ml collagenase IV ( Sigma ) , and 1 mg/ml DNase ( Sigma ) ) . Tissue fragments were further dispersed by repeated aspiration , crushed through a 50-μm pore size nylon filter ( BD Biosciences ) and then centrifuged ( 400 g , 10 min , 4°C ) . Erythrocytes in the cell pellets were lysed , and the remaining cells were resuspended in 5% RPMI . For pulmonary cell activation , 1x106 cells/well were cultured for 4 h with PMA ( 50 ng/mL ) , ionomycin ( 500 ng/mL ) , and brefeldin A ( 5 mg/mL ) , followed by staining for extracellular and intracellular markers and analysis by flow cytometry . CD4+ T cells from spleens and lymph nodes of 5-7-week old C57BL/6 mice were purified by magnetic separation using anti-APC microbeads ( Miltenyi Biotec ) . For Th17 differentiation , the cells were activated for 5 days by plate- bound anti-CD3 ( 2 μg/ml ) and anti-CD28 ( 1μg/m ) in the presence of 5 ng/mL TGF-β , 20 ng/mL IL-6 , 50 ng/mL IL-23 , 10 ug/mL anti-IL-2 , 10 ug/mL anti-IFN-γ and 10 ug/mL anti-IL-4 . All recombinant cytokines were obtained from R&D and neutralizing antibodies from BioXCell . When indicated , 10 ng/mL of IL-1α was added to the cell culture after 3 days of differentiation . Differentiated Th17 cells were stained with CFSE and lymphocyte proliferation was monitored by flow cytometry . Cells were stained with fluorochrome-conjugated antibodies specific for the surface molecules CD3 , CD4 , MHCII , CD11c , CD11b , Ly6G , CCR6 and for the intracellular cytokines IL-17 and IFN-γ . For the intracellular cytokine staining , cells were previously permeabilized using PBS containing 1% FBS , 0 . 1% sodium azide and 0 . 2% saponin . Data acquisition performed using a FACSCanto II flow cytometer ( BD Biosciences ) were plotted and analyzed using FlowJo software . Lungs from C57BL/6 and Il1a-/- mice were immersed in OCT medium ( Sakura Finetek ) , snap-frozen in liquid nitrogen , and stored at −80°C until analysis . Frozen tissue sections ( 5 μm ) were fixed with acetone and endogenous peroxidase activity and non-specific sites were blocked with 3% hydrogen peroxide and with PBS plus 3% ( W/V ) non-fat milk , respectively . Following , the slides were incubated with rat IgG anti-mouse Ly6G ( Biolegend ) or IgG anti-mouse ( as control ) antibodies before the incubation with biotin-labeled antibody ( Vector Laboratories ) . Staining was developed with 3 , 3'-diaminobenzidine ( Vector Laboratories ) and the reaction was counterstained with Mayer’s hematoxylin . We measured immunostained areas using Image J software . Briefly , the range of positivity was defined using the IHC Tool box . Next , the images were converted to 8-bit , and the grayscale was converted to binary ( black and white ) . The threshold was adjusted , and the labeled areas became the black portions . Finally , the percentage of stained area was analyzed . Total RNA was extracted from mouse tissue using the TRIzol reagent ( Invitrogen ) and the SV Total RNA Isolation System Kit ( Promega ) according to the manufacturer’s instructions . Complementary DNA was synthesized through a High Capacity cDNA Reverse Transcription Kit ( Applied Biosystems ) . SYBR Green Mix–based real-time quantitative PCR assays were performed using the StepOnePlus Real-Time PCR System ( Applied Biosystems ) . The mean cycle threshold ( Ct ) values of triplicate measurements were used to calculate the expression of the target genes , which were normalized to the housekeeping gene β2m and analyzed with the 2–ΔΔCt method . Primers ( presented in S1 Table ) were designed using the Primer Express software package v2 . 0 ( Applied Biosystems ) , based on the nucleotide reference sequences available at GenBank database . The 96-well precoated plate ( PAMM-073Z; Qiagen ) together with the SYBR Green Quantitative PCR Master Mix ( SAbiosciences ) was used to examine by qPCR a set of 89 Th17 response related genes in samples of CD4+ T cells cultured under the following three conditions: Th0 , Th17 and Th17 plus IL-1α . The samples from four different wells were polled together to obtain enough RNA for the whole array . RNA integrity was analyzed by Agilent Bioanalyzer following the manufacturer’s instructions . The raw data were normalized to the cycling threshold ( Ct ) from the housekeeping gene Gusb ( β-glucuronidase ) and analyzed with the 2–ΔCt method . Principal component analysis ( PCA ) and heat maps were generated using the R Language and Environment for Statistical Computing v . 3 . 5 . 0 and the packages ggplot2 and gplots . Where indicated , Casp11-/- mice were treated intranasally ( i . n . ) on days 0 , 1 , 4 , 7 , 14 , 21 and 28 with 100 ng/mice of recombinant mouse IL-1α ( rIL-1α; R&D Systems ) and were killed on day 30 post P . brasiliensis infection for analysis of CFU and cytokines in the lung . Where indicated , every three days from days 15 to 27 after P . brasiliensis challenge , recombinant mouse IL-17A ( rIL-17A; Gibco ) was administrated i . n . at a dose of 0 . 5 ug per mouse in a total of 25 uL . The pulmonary fungal burden was determined at day 30 post infection . Bone marrow-derived macrophages ( BMDM ) were differentiated from femurs of 7-week-old WT , Il1a-/- , and Casp11-/- naïve mice , by culturing precursor cells in RPMI 1640 medium supplemented with 20% FBS and 30% L929 cell–conditioned media for 7 days at 37°C and 5% CO2 , as previously described [105] . After differentiation , cells were harvested , seeded and infected with P . brasiliensis at MOI ( multiplicity of infection ) of 5 . IL-1β and IL-1α levels in culture supernatants were measured 48 hours post-infection using standard sandwich ELISAs according to the manufacturer’s recommendations ( BD Biosciences ) . When indicated , BMDMs were treated with 5 or 10 ug/mL of neutralizing IFN-β antibody ( PBL Assay Science ) at the same time of the infection; or pre-treated with or without recombinant IFN-β ( 100 or 1000 U/mL ) 2 hours before P . brasiliensis infection; or incubated for 6 h with 50 or 100 ng/mL of recombinant IL-17A ( BioSource-Life Technologies ) before infection . BMDMs ( 1x106 cells/well ) were infected with viable fungi at a MOI of 5 in the indicated time points . Cell-free supernatants and cell lysates were collected and subjected to western blotting . Primary antibodies included a monoclonal rat anti-caspase-11 ( Sigma ) and a goat anti-mouse IL-1β ( Sigma ) . In some cases , proteins from the cell culture supernatants were precipitated with methanol-chloroform . In this assay , fungus was added at a MOI of 5 to 2x105 BMDMs plated on 13-mm glass coverslips in 24-well tissue culture dishes . After 24 hours of incubation at 37°C in 5% CO2 , the coverslips were washed and then inverted onto a 5uL of PBS containing 25 μg/mL EtBr and 5 μg/mL acridine orange . Images were acquired using a fluorescence microscope ( Olympus America ) and analyzed with the ImageJ software . Pore formation was calculated according to the percentage of permeable cells to EtBr that were photographed ( 10x objective ) from three different fields containing approximately 400 cells/field . BMDMs were seeded into 24-well plates ( 2x105 cells/well ) and cultured in RMPI 1640 media lacking phenol red with 15mM HEPES and 2g/L NaHCO3 supplemented with 10% FBS . Lactate dehydrogenase ( LDH ) released by 24-hour P . brasiliensis-infected macrophages was quantified in culture supernatants using the CytoTox 96 LDH-release kit ( Promega ) . Percentage of LDH release was calculated as follows: ( LDH infected—LDH uninfected/LDH total lysis—LDH uninfected x 100 ) . LDH total lysis was determined by lysing the cultures with Triton X-100 . BMDMs ( 106 cells/well ) were infected with P . brasiliensis at a MOI of 5 . After 24 h of incubation , the frequency of macrophages undergoing pyroptosis cell death was defined by the loss of membrane integrity evaluated by propidium iodide ( PI ) staining . Data were acquired on a FACSCanto II flow cytometer ( BD Biosciences ) and analyzed using FlowJo software ( Tree Star , Ashland , OR ) . BMDMs were plated on 96 wells culture plates at 2x105 cells/well , pre-treated for 6 hours with 10 or 50 ng/mL of recombinant IL-1α ( rIL-1α; R&D Systems ) or overnight with 50 ng/mL of recombinant IFN-γ ( rIFN-γ; BD Biosciences ) , infected with P . brasiliensis ( MOI 1:25 fungi per macrophages ) and incubated at 37°C , 5% of CO2 for 2 hours . After this period , extracellular yeast cells were removed by washing the wells , and cells were cultured for 48 hours . Finally , cells were lysed with saponin ( 0 . 05% ) for colony forming unit ( CFU ) counting . Nitrite production in the culture supernatants was estimated using the Griess reaction in BMDM cultures stimulated at a MOI of 5 , as described previously [106] . Data are expressed as means ± SD . For comparison between multiple groups , one-way ANOVA followed by the Tukey-Kramer post-test was applied . The log-rank test ( Mantel–Cox ) was used to compare survival curves and unpaired parametric student´s t-test was used to compare differences between two groups that assumed a normal distribution . Comparisons between two experimental groups not normally distributed were performed with non-parametric Mann-Whitney U statistical test . The normal distribution of data was analyzed by D’Agostino test . All analyzes were performed using the Prism 5 . 0 software ( GraphPad Software ) . P values < 0 . 05 were considered statistically significant . | After infectious insults , the release of intracellular molecules from dying cells , such as IL-1α , provokes a local inflammation as an alert of tissue damage . In this study , we investigated the steps necessary for IL-1α-mediated inflammation and its implications after fungal infection . We identified a molecular mechanism in which IL-1α plays a pivotal role on the control of Paracoccidioides brasiliensis infection . We observed that after fungal recognition , macrophages produce IFN-β , a cytokine that promotes the expression of procaspase-11 . This enzyme is then activated to trigger a rapid pore-mediated cell lysis , leading to the passive leakage of the cytosolic IL-1α . Once extracellularly , IL-1α functions via paracrine signaling on surrounding cells to enhance the inflammatory response against the pathogen . IL-1α coordinates nitric oxide and IL-6 production by macrophages upon infection , but it also acts directly on CD4+IL-17+ T lymphocytes by reprograming their transcriptional profile and potentiating IL-17 production . While NO has intrinsic fungicidal properties and IL-6 drives Th17 cell differentiation , IL-17 recruits neutrophils into lungs of infected mice . Furthermore , macrophages synthesize more IL-1α in response to an IL-17-rich milieu . Therefore , IL-1α initiates a sustained and self-perpetuating inflammatory loop that is required for host resistance to P . brasiliensis infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"blood",
"cells",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"yeast",
"infections",
"pathology",
"and",
"laboratory",
"medicine",
"immune",
"physiology",
"cytokines",
"pathogens",
"immunology",
"microbiology",
"cell",
"differentiation",
"developmental",
"biology",
"fungi",
"molecular",
"development",
"paracoccidioides",
"brasiliensis",
"fungal",
"diseases",
"fungal",
"pathogens",
"infectious",
"diseases",
"mycology",
"white",
"blood",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"t",
"cells",
"microbial",
"pathogens",
"immune",
"system",
"eukaryota",
"cell",
"biology",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"macrophages",
"organisms"
]
| 2019 | Caspase-11-dependent IL-1α release boosts Th17 immunity against Paracoccidioides brasiliensis |
Campylobacter jejuni is the major cause of bacterial food-borne illness in the USA and Europe . An important virulence attribute of this bacterial pathogen is its ability to enter and survive within host cells . Here we show through a quantitative proteomic analysis that upon entry into host cells , C . jejuni undergoes a significant metabolic downshift . Furthermore , our results indicate that intracellular C . jejuni reprograms its respiration , favoring the respiration of fumarate . These results explain the poor ability of C . jejuni obtained from infected cells to grow under standard laboratory conditions and provide the bases for the development of novel anti microbial strategies that would target relevant metabolic pathways .
Campylobacter jejuni is the leading cause of bacterial food-borne disease in the USA and Europe and a major cause of diarrheal disease in developing countries [1] , [2] . The typical route of C . jejuni transmission is through handling and ingesting contaminated food , milk , or water [3] . In many cases , poultry is a common source of food contamination , and the gastrointestinal tract of many species of birds often serves as the natural reservoir of C . jejuni [4] . Despite its important impact on public health , surprisingly little is known about its pathogenesis [5] . Biochemical analysis and comparative genomic studies of C . jejuni have revealed unique features in its metabolic pathways [6]–[8] . Unlike most enteric bacteria , C . jejuni lacks the key glycolytic enzyme 6-phosphofructokinase as well as alternative pathways for sugar catabolism . Consequently , C . jejuni utilizes amino acids and organic acids as the major carbon sources . In addition , C . jejuni possesses highly branched electron transport chains , which allow it to respire not only oxygen but also a variety of electron acceptors including fumarate , nitrate , nitrite , trimethylamine-N-oxide ( TMAO ) , and dimethyl sulfoxide ( DMSO ) [9]–[12] . Such remarkable diversity in its respiratory pathways may contribute to its ability to survive or grow in substantially different environments [13] . We previously reported that C . jejuni survives within cultured intestinal mammalian cells by avoiding its delivery to lysosomes [14] . We also found that 24 hours after infection , C . jejuni recovered from cultured intestinal mammalian cells cannot be efficiently cultured under microaerophilic conditions unless it is pre-incubated under oxygen-limiting conditions . We hypothesized that within host cells C . jejuni may alter its physiology and change its mode of respiration perhaps due to low level of oxygen within the C . jejuni-containing vacuole . Such metabolic change should most likely be reflected in alterations on the levels of different sets of metabolic enzymes . In an effort to understand the physiological changes in intracellular C . jejuni we surveyed its proteome at different time after infection of cultured mammalian cells and compared it to the proteome of bacteria grown in vitro . Our data indicate that within cultured mammalian cells , C . jejuni undergoes a severe metabolic downshift and an apparent change of its respiration mode . These data provide major insight into the in vivo metabolism and pathogenesis of this important pathogen .
To obtain a snapshot of C . jejuni protein expression during infection , we developed a protocol to isolate bacteria away from mammalian host cell proteins by exploiting C . jejuni's moderate resistance to detergent treatment . This protocol resulted in the presence of very low amounts ( <15% ) of host cell-derived proteins in the final preparation . COS-1 cells were infected with C . jejuni and 2 hs and 20 hs after infection , intracellular bacteria were isolated , and proteins from bacterial lysates were pre-fractionated by gel electrophoresis prior to LC-MS/MS analysis . We chose to compare these two time points because while at 2 hs after infection C . jejuni is readily culturable under microaerophilic conditions , at 20 hs after infection it is not [14] . We also reasoned that after 2 hs of infection , the C . jejuni proteome would closely resemble that of extracellular bacteria allowing the opportunity to examine the potential changes that may occur during the transition from an extracellular to an intracellular niche . More importantly , the comparison of proteomes of bacteria isolated by similar procedures minimized potential differences due to sample processing that could occur by comparing the proteomes of bacteria grown in culture with that of bacteria obtained from within mammalian cells . Samples generated with the outlined protocol ( see Figure 1A ) exhibited limited host-protein contamination when examined by both Coommassie blue staining after SDS-PAGE ( Figure 1B ) , and mass spectrometry analysis ( less than 15% of the identified proteins corresponded to the mammalian cell proteome ) . Through the assignments of ∼400 , 000 MS/MS spectra we were able to detect 1 , 428 C . jejuni proteins , which represent ∼86% of the entire C . jejuni proteome ( see Dataset S1 ) . We used spectral counting to assess the relative protein abundance in the different samples [15] . Most proteins in the dataset were assigned from multiple spectral counts , which combined with the numerous technical and biological replicates , provided strong confidence to the protein assignments . In addition we quantified a selected group of proteins by selective reaction monitoring ( SRM ) ( see below ) [16] . Our data showed that proteomic differences determined by spectral counts agreed well with those measured by SRM although in some cases ( i . e . proteins present in large amounts ) spectral counting slightly underestimated the extent of proteomic differences ( reported as the fold changes ) between samples ( see below ) . Such observation is consistent with the fact the dynamic range of spectral counting is not as high as that of SRM [16] . To our knowledge , this represents the most comprehensive C . jejuni proteomic dataset available . After applying the criteria described in Materials and Methods , we found 225 proteins whose levels at 20 hs after infection differed significantly from those at 2 hs after infection ( Dataset S2 ) . The vast majority of these proteins ( 211 ) showed significantly ( p<0 . 05 ) reduced levels at 20 hs after infection . Only a few proteins showed a very moderate increase in their levels at 20 hs post-infection ( Dataset S2 ) . As predicted , the proteome of C . jejuni grown in vitro closely resembled the proteome of C . jejuni isolated from within cultured mammalian cells 2 hs after infection ( Dataset S3 ) . We specifically examined the profile of proteins whose levels are known to increase as a consequence of general stress responses [17]–[19] and found that their levels were either unchanged or were even slightly decreased ( Table 1 ) . Furthermore , the levels of proteins that are thought or known to protect C . jejuni against oxidative stress [20] were also either decreased or unaltered ( Table 1 ) , although their levels were relatively high in all time points . These results suggest that C . jejuni may not be subject to environmental stress when located within cultured mammalian cells and have important implications for the understanding of C . jejuni intracellular stage during infection . Functional grouping of the proteins whose expression changed upon C . jejuni's transition from the extracellular to the intracellular environment indicated an over representation of proteins involved in different metabolic pathways or in the transport of nutrients or compounds across the bacterial envelope ( Figure 2 ) . To better evaluate the potential physiological impact of the reduction of expression of proteins involved in metabolic pathways or transport mechanisms , we generated a graphical metabolic network database , using the Pathway Tools v14 . 0 [21] . For this purpose , we input the C . jejuni 81–176 genome sequence [22] to the program , which can predict pathways by matching the annotated enzymes of the input genome with its own comprehensive pathway/reaction database Metacyc [23] . We then manually curated this computation-derived C . jejuni pathway database based on published literature , such as biochemical studies on metabolic enzymes , transporters or regulators specifically in Campylobacter species . C . jejuni has unique metabolic capabilities presumably to adapt to different environments , so we customized the C . jejuni database by adding new pathways/reactions that are not present in Metacyc , or by linking related pathways to form super-pathways to better visualize the metabolite flow . Overall this database could account for 155 metabolic pathways , 885 associated enzymatic reactions , 75 transporters and 49 transport reactions . In this graphic representation , C . jejuni proteins were mapped into individual pathways , and were differentially labeled based on whether they were detected in our analysis or whether their expression level was decreased or increased . Following this template , our proteomic analysis was able to monitor 151 metabolic pathways , 382 associated enzymes , 429 enzymatic reactions , and 41 transporters ( Figure 3 ) ( The entire database will be made available in the BioCyc Database collection web site [http://biocyc . org] ) . Evaluation of the proteins ( and associated pathways ) whose expression significantly decreased 20 hs after infection indicates that C . jejuni undergoes a significant metabolic downshift within host cells ( Dataset S2 ) . Several proteins associated with various anabolic pathways showed significantly reduced levels in samples obtained 20 hs after infection ( Dataset S2 and Figure 3 ) . This includes components of amino acid biosynthesis pathways such as those associated with histidine , lysine , valine , leucine , isoleucine , methionine , glycine , and alanine biosynthesis . Similarly , components of other important biosynthetic pathways including those associated with the synthesis of prostetic groups and co-factors , fatty acids and lipids , pentose phosphate , as well as purine and pyrimidine nucleotides were also significantly down-regulated 20 hs after infection . In addition , the levels of several proteins associated with various catabolic pathways were significantly reduced in C . jejuni isolated from cells 20 hs after infection . These include pathways involved in the degradation and utilization of amino acids ( e . g . proline , L-asparatate , lysine , serine and L-arginine ) , and C1 compounds . Pathways potentially involved in nutrient acquisition were also down-regulated , including several amino acid ABC-type transporters as well as transporters for phosphate , potassium , tungstate and molybdate . Overall , <50% of all C . jejuni proteins whose levels were decreased after 20 hs of infection were associated with metabolic pathways or transport reactions . Consistent with an overall metabolic downshift , the levels of several ribosomal protein subunits ( e . g . 30S ribosomal proteins S6 , S19 , S15 , and S20 , and 50S ribosomal protein L9 and L24 ) were significantly decreased . This metabolic downshift is consistent with the observation that C . jejuni does not replicate within cultured mammalian cells but , rather , it seems to go into a dormant , non-culturable state [14] . In addition , the observed metabolic downshift resembles the one observed during C . jejuni's transition to late stationary phase [24] . However , unlike what was observed in C . jejuni stationary phase transcriptional reprogramming , we did not observe increased expression of heat shock proteins ( e . g . GroEL , GroES , GrpE , ClpB ) or proteins associated with oxidative stress resistance ( e . g . AhpC , SodB , and Tpx ) . This suggests that the remodeling of the C . jejuni proteome that occurs within cultured mammalian cells is not simply the result of its transition to a different “growth phase” but most likely , the result of its specific adaptation to the intracellular environment . The observation that intracellularly localized C . jejuni becomes culturable after incubation under very low oxygen conditions suggests at least two possibilities to explain its non-culturability under standard microaerophilic conditions . One possibility is that the intracellular remodeling of its proteome renders C . jejuni oxygen sensitive . This hypothesis would be supported by the observation that some proteins potentially involved in conferring protection against oxidative stress showed reduced levels at 20 hs after infection ( Table 1 ) . However , even though slightly reduced , the levels of these enzymes remained relatively high . Furthermore , no significant changes in the levels of proteins thought to be most important in oxidative stress protection ( e . g . SodB and KatA ) [25]–[27] were observed suggesting that other factors must account for C . jejuni's inability to grow under microaerophilic conditions when directly obtained from cultured cells . Consistent with this hypothesis , examination of the proteome of intracellularly-localized C . jejuni suggests that it undergoes reprogramming of its respiration . Indeed , the levels of many proteins that are central to the main respiration pathways were significantly reduced in the 20 hs sample ( Table 2 ) , an observation confirmed by SRM experiments ( Table 3 ) . For example , key components of the aerobic respiration pathway [7] were greatly reduced . The C . jejuni genome only encodes two terminal oxidases , a cbb3-type cytochrome c oxidase and a CioAB-type ( cyanide-insensitive oxidase ) oxidase [6] . The expression levels of cbb3 type cytochrome c oxidase ( CcoP and CcoO ) decreased markedly 20 hs after infection ( Table 2 and Table 3 ) . The CioAB-type oxidase could be hardly detected before and after infection , which is consistent with the hypothesis that the cbb3-type cytochrome c oxidase plays a more important role in aerobic respiration [28] . This hypothesis is also supported by the observation that a C . jejuni ccoN mutant was unable to colonize chickens while a cydA mutant colonized as well as the wild type [29] . Reduced levels of the only enzyme that directly uses O2 as electron acceptor may make it difficult for intracellular C . jejuni to grow under microaerophilic conditions after its extraction from within mammalian cells . Some components of anaerobic respiration also show reduced levels in C . jejuni samples obtained 20 hs after infection . For example , the levels of NapA/NrfA , involved in nitrate/nitrite respiration , and TorA/TorC , which are central components of TMAO respiration , were also markedly decreased . In contrast , the levels of the fumarate reductase FrdA between the 2 hs and 20 hs time points remained constant and relatively high suggesting that fumarate may be an important electron acceptor for C . jejuni during its intracellular stage . Thus , our proteomics data suggest that 20 hs after infection , aerobic respiration is greatly reduced , while the anaerobic respiration pathway can still be maintained by using fumarate as alternative electron acceptors . Furthermore , these results may help to explain C . jejuni's inability to grow under standard microaerophilic conditions when plated immediately after its release from mammalian cells 20 hs after infection . Although the actual oxygen concentration within the C . jejuni-containing vacuole is unknown , during the infection process , C . jejuni are exposed to a higher oxygen tension ( 5% CO2 vs . 10% CO2 when grown microaerophilically ) at least until they reach the intracellular environment . To evaluate if this ( presumably brief ) change in oxygen concentration may itself result in changes in the C . jejuni proteome , we compared the proteome of C . jejuni incubated in the infection medium under 5% CO2 for 2 hs ( without host cells ) with the proteome of in-vitro grown bacteria . However , we found no significant differences in the C . jejuni proteome under these two conditions ( Dataset S4 ) indicating that exposure to the higher oxygen conditions during infection cannot account for the observed changes in the intracellular population of C . jejuni 20 hs after infection , which are most likely the direct consequences of its adaptation into the host intracellular environment . We have previously shown that intracellular C . jejuni at late time points of infection ( e . g . >18 hs ) can be rendered culturable under standard microaerophilic conditions if pre-incubated under oxygen-limiting conditions ( O2<0 . 2% , CO2 ∼6 . 5% ) [14] . Therefore , to gain further insight into C . jejuni's intracellular metabolic reprogramming , we carried out a proteomic analysis of C . jejuni obtained immediately after its incubation under oxygen-limiting conditions . We reasoned that by examining changes induced by this incubation step , which render it culturable , we might be able to gain further insight into the changes undergone by C . jejuni when located within the cells , which render it non-culturable . Bacteria obtained from infected cells 20 hs after infection and incubated for 48 hs under low oxygen conditions were scraped from plates and their proteome was examined as described above ( Dataset S5 ) . We observed significantly increased levels of enzymes involved in known C . jejuni anaerobic respiration pathways including FrdA , NapA , TorA , and NrfA [12] ( Dataset S6 ) . In addition , the levels of the anaerobic C4-dicarboxylate transporters DcuA and DcuB were also increased after incubation under low oxygen conditions . Strikingly , we also observed that the levels of several enzymes involved in aerobic respiration were also increased after incubation under oxygen-limiting conditions . For example , all three subunits of ubiquinol-cytochrome c reductase ( PetA , PetB and PetC ) were markedly increased under these conditions . These results may provide an explanation for the observation that pre-incubation of intracellular C . jejuni under oxygen-limiting condition renders it ready for growth under microaerophilic environment , in which oxygen acts as the main electron acceptor [14] . Furthermore , it suggests the possibility that the low oxygen environment encountered by C . jejuni in the gut may also prepare the bacterium for its phase outside the host , in which oxygen levels are likely much higher . Interestingly , the ability to prepare for a change in a future environment has been previously observed in V . cholera , in which genes required for the aquatic environment are induced within the intestinal track , where they are not needed [30] . To gain more insight into the potential respiratory reprogramming of C . jejuni within mammalian cells we specifically compared the relative abundance of key enzymes for different respiration pathways under the different conditions examined in this study . We specifically examined the levels of FdhA , PetC , CcoP and CcpA , which are involved in aerobic respiration ( Figure 4A ) , and FrdA , NapA , NrfA , and TorA , which are central components in the respiration of fumarate , nitrate , nitrite and TMAO/DMSO , respectively ( see Figure 4B ) [11] , [31] , [32] . For quantification we used both spectral counting ( Figure 4A and 4B ) as well as SRM ( Table 3 ) , which yielded equivalent results . We found that the levels of aerobic respiration enzymes were markedly decreased in intracellular C . jejuni 20 hours after infection . A similar pattern was observed in the case of NapA and TorA . Although the levels of NrfA remained unchanged during infection , the actual protein levels were very low suggesting that this respiration pathway may not be central for the metabolism of C . jejuni during its intracellular state . The levels of FrdA also remained unchanged in C . jejuni samples 2 and 20 hs after infection . However , in contrast to NrfA , the levels of FrdA were high , suggesting that fumarate may be an important electron acceptor during C . jejuni's intracellular stage . Consistent with this hypothesis , the levels of AspA , which is an aspartase that can provide fumarate for respiration [33] , showed a similar pattern to the levels of FrdA ( Figure 4C ) , which were characterized by unchanged although high levels 20 hs after infection . Furthermore , the levels of AspA , FrdA and MfrA , another fumarate reductase [11] , significantly increased after incubation under low oxygen conditions ( Figure 4C ) . To test the potential significance of fumarate respiration during C . jejuni intracellular survival , we constructed C . jejuni strains carrying mutations in aspA or the fumarate reductases frdA and/or mfrA and examined their viability 20 hs after infection . Since in the absence of fumarate the C . jejuni aspA , frdA , or mfrA mutants have been shown to be defective for invasion , these mutant strains were grown and allowed to infect cells in the presence of fumarate , which results in wild type levels of bacterial internalization [34] . The aspA , frdA and frdA/mfrA mutants showed a significant decrease in viability 20 hs after infection ( Figure 5 , bottom graph ) indicating that respiration of fumarate is central to the metabolism of intracellular C . jejuni . Complementation of the aspA mutant restored viability to wild-type levels . In contrast , the mfrA mutant showed intracellular survival close to wild type . This is consistent with the extremely low levels of MfrA ( barely detectable by our LC-MS approach ) observed in both extracellular and intracellular C . jejuni preparations . These data are also consistent with previous observations indicating that C . jejuni aspA or frdA mutant strains are impaired in chicken colonization while the mfrA mutant is not [30][28] . To further corroborate the hypothesis that fumarate respiration plays an important role in C . jejuni intracellular survival , we examined C . jejuni strains carrying mutations in napG , torA , cydA , and fdhB for their ability to survive within cultured mammalian cells . These genes are involved in aerobic and anaerobic respiration pathways [29] [32] . In contrast to mutants with impaired fumarate respiration , these mutant strains showed intracellular viability indistinguishable from that of the wild-type ( Figure 5 ) . Therefore , our proteomic and functional data indicate that fumarate respiration plays an important role in C . jejuni intracellular survival . We have shown here through a detailed proteomic study that C . jejuni undergoes a significant metabolic reprogramming upon internalization within mammalian cells . A salient feature of this metabolic reprogramming is a significant metabolic downshift . This metabolic downshift is qualitatively different from the one associated with C . jejuni's transition to stationary phase and therefore represents a unique metabolic state presumably associated with its adaptation to the intracellular environment . It is also consistent with the observation that although C . jejuni survives within mammalian cells , it does not replicate [14] . It is therefore possible that the intracellular environment , presumably shielded from many innate immune defense mechanisms , may serve as a reservoir for this pathogen but not necessarily as a replication site . The proteomic profile indicates that although non-replicating , C . jejuni is certainly not subject to environmental stress in this intracellular environment since none of the proteins usually associated with responses to different environmental stress showed increased levels . Another salient feature of the intracellular metabolic reprogramming is a change in respiration mode , which may be responsible for its decreased culturability under microaerophilic conditions in which oxygen is the main electron acceptor . Indeed , our data suggest that C . jejuni may favor fumarate respiration inside mammalian cells . Consistent with our data , C . jeuni aspA and frdA mutants , which are defective for fumarate respiration , were shown to be significantly impaired for chicken colonization [30][28] , while mutations in genes essential for other respiration pathways were not [Olson , Appl Environ Microbiol . 2008 Mar;74 ( 5 ) :1367–75 . ] . Therefore , this study highlights the importance of metabolic reprogramming in the biology of C . jejuni and may help the development of novel anti microbial strategies that would target relevant metabolic pathways .
The C . jejuni 81–176 wild-type strain , its aspA mutant , and the complemented aspA mutant derivative have been previously described [34] , [35] . All other mutant strains were constructed as previously described [22] . Bacteria were routinely grown on blood agar plates ( tryptic soy broth agar supplemented with 5% sheep blood ) or in brain heart infusion ( BHI ) broth at 37°C under 10% CO2 ( microaerophilic ) or low oxygen conditions ( GasPak Plus , BD-Diagnostic Systems , New Jersey ) . All C . jejuni strains were stored at −80°C in BHI broth containing 30% glycerol . COS-1 ( African green monkey kidney fibroblast-like cell line ) cells were obtained from the American Type Culture Collection ( Manassas , VA ) and grown in Dulbecco's modified Eagle medium ( DMEM ) supplemented with 10% bovine calf serum ( BCS ) . All cell lines were grown under an atmosphere of 5% CO2 . In order to obtain sufficient amounts of bacterial proteins for mass spectrometric analysis ( see strategy described below ) , we determined that ∼5×107 to 108 bacteria ( equivalent to ∼50–100 µg of protein yield ) would be needed for proteomic analysis . To obtain this number of bacteria C . jejuni infections were carried out in 15-cm plates as previously described [34] with a multiplicity of infection of 1000 . Infection was allowed to proceed for 2 hs in HBSS medium . Subsequently , cell monolayers were washed extensively with pre-warmed HBSS and incubated in gentamicin-containing ( 100 µg/mL ) DMEM media for another 2 hs to kill extracellular bacteria . After gentamicin treatment , cells were washed and lysed to release intracellular bacteria . Alternatively , the culture media was replaced with fresh DMEM containing lower concentration of gentamicin ( 10 µg/mL ) and cells were further incubated for 16 hs before releasing intracellular bacteria . A two-step differential centrifugation strategy was developed to separate intracellular bacteria from host cell debris . COS-1 cells were lysed in a buffer containing 0 . 5% Triton X-100 , 20 mM Tris-HCl ( pH 7 . 6 ) and 150 mM NaCl . The sample was first pelleted at 300× g for 5 min to remove host cell nuclei , and then the post-nuclear supernatant was centrifuged again at 3000× g for 20 min . These conditions resulted in the recovery of most of the bacteria in the second pellet as shown by CFU assays . The resulting bacterial pellets were washed extensively with RIPA buffer to minimize host-protein contamination . We carried out proteomic analysis of proteins extracted from different C . jejuni samples by LC-MS/MS . The experimental design is depicted in Figure 1 . Briefly , intracellular C . jejuni were harvested from mammalian cells at 2 hs and 20 hs of infection by differential centrifugation as described above . Bacterial cell lysates were separated by 10% SDS-PAGE to pre-fractionate the bacterial proteins . Electrophoresis was stopped when the dye front reached 1 . 5 cm below the stacking gel . Subsequently , the gel was divided equally into three slices , each slice was subjected to in-gel protein digestion ( see below ) , and extracted peptide samples were analyzed by LC-MS/MS for protein identification . Spectral counts associated with protein assignments were utilized to quantitatively assess their relative abundance in each sample . Triplicate LC-MS/MS runs were conducted for each sample , thereby allowing the assessment of the technical reproducibility of the measurements . Statistical analysis was performed on data obtained from 7 biological replicates . Each replicate was comprised of two samples corresponding to two bacterial populations ( from 2 hs and 20 hs of infection ) . Since each bacterial sample was divided into three gel fractions and each fraction was run in triplicate , in total 126 LC-MS/MS experiments were carried out to obtain the entire data set . Protein and peptide assignments from different gel fractions were pooled within the same bacterial sample . All mass spectrometric data have been deposited at the public data base PRIDE hosted by EMBL-EBI ( http://www . ebi . ac . uk/pride/ ) Gel slices were excised into small ( ∼1×1 mm ) cubes and transferred into sample tubes . Protein disulfide bonds were reduced with 10 mM DTT in 100 mM NH4HCO3 at 56°C for 30 min and subsequently alkylated with 55 mM iodoacetamide ( IAM ) in 100 mM NH4HCO3 at room temperature for 20 min ( in complete darkness ) . Upon alkylation , the gel samples were destained with 50% acetonitrile ( ACN ) in 50 mM NH4HCO3 and dehydrated by addition of neat ACN . After removal of the destaining buffer and ACN , the samples were subjected to in-gel digestion . The digestion buffer contained 10 ng/µL trypsin in 50 mM NH4HCO3 and the enzymatic reaction was allowed to proceed overnight at 37°C . The resulting tryptic peptides were extracted from the gel matrix by equilibrating the samples with 50% ACN and 5% formic acid ( FA ) . Finally the extracted peptides were vacuum dried prior to LC-MS/MS analysis . Nanoflow reverse-phase LC separation was carried out on a Proxeon EASY-nLC System ( Thermo Scientific ) . The capillary column ( 75 µm×150 mm , PICOFRIT , New Objective , Woburn , MA ) was packed in-house . Briefly , a methanol slurry containing 5 µm , 100 Å Magic C18AQ silica-based particles ( Microm BioResources Inc . , Auburn , CA ) was forced to pass through an empty capillary ( with a frit in the end ) using a pressurized device . Tryptic peptides were dissolved in HPLC-grade water and ∼200 ng of samples were loaded onto the analytical column in a single LC-MS/MS experiment . The mobile phase was comprised of solvent A ( 97% H2O , 3% ACN , and 0 . 1% FA ) ) and solvent B ( 100% ACN and 0 . 1% FA ) . The LC separation was carried out with the following gradient: solvent B was started at 7% for 3 min , and then raised to 35% in 40 min; subsequently , solvent B was rapidly increased to 90% in 2 min and maintained for 10 min before 100% solvent A was used for column equilibration . Peptides eluted from the capillary column were electrosprayed directly onto a linear ion trap mass spectrometer ( LTQ Velos , ThermoElectron , San Jose , CA ) for MS/MS analysis . A data-dependent mode was enabled for peptide fragmentation . One full MS scan was followed by fragmentation of the top 10 most intense ions by collision-induced dissociation ( CID ) . Dynamic exclusion ( with a duration of 6 seconds ) was enabled to preclude repeated analyses of the same precursor ion . Both pre-column and analytical column were washed intensively between different samples to remove any carryover from previous runs . Peptides and proteins were assigned by a database search method . Specifically , MS/MS scans were processed with commercial software ( Bioworks Browser , ThermoElectron , San Jose , CA ) to generate DTA files . Then those files were searched with MASCOT ( Matrix Science Ltd . , London , UK ) against a C . jejuni 81–176 protein database . The resulting peptide and protein assignments were filtered to achieve a peptide false discovery rate ( FDR ) of <1% and a protein FDR of <3% using the target-decoy method ( e . g . reverse database search ) . Protein relative abundance between different samples was assessed using a spectral counting method . Spectral counts represent the total number of repeated identification of peptides for a given protein during the entire analysis and provides a semiquantitative measurement of protein abundance . Raw spectral counts were normalized against constant protein signals ( between 2 h and 20 h samples ) . Eight highly abundant C . jejuni proteins ( GroEL , AcnB , flagellin , pyruvate ferredoxin , RpoB , RpoC , Tpx , methyl-accepting chemotaxis protein/CJJ81176_1205 ) having little change in expression level during the course of infection were chosen to calculate the final normalization factor . A protein fold change ( or ratio ) between paired samples was calculated by dividing the average spectral counts of triplicate measurements . Then the averaged value of ratios for 7 biological replicates was calculated for each detected protein . A ratio ( 20 h over 2 h samples ) greater or smaller than 1 indicates that levels of a given protein in 20 h samples were higher or lower , respectively , than those at 2 h after infection . Paired Student's t test was performed on data from 7 biological replicates . For the purpose of our analysis , differences between the average fold change of a given protein 2 h and 20 h after infection of >2 and a p-value<0 . 05 were considered significant . Among those proteins that met our criteria , a subset of metabolic enzymes of particular interest to this study were further examined by selected reaction monitoring ( SRM ) . C . jejuni infection of cultured mammalian cells was also carried out as previously described [14] with an MOI of 100 . In all cases , the C . jejuni aspA , frdA or frdA mfrA double mutant strains were grown in the presence of fumarate and fumarate was added during the initial infection of cultured mammalian cells . After 2 hs of infection , cells were washed extensively with HBSS followed by the addition of pre-warmed DMEM containing 100 µg/mL gentamicin for 2 hs to kill extracellular bacteria . Cells were lysed in a buffer containing 0 . 5% Triton X-100 , 20 mM Tris-HCl ( pH 7 . 6 ) , and 150 mM NaCl , and the number of intracellular bacteria was assessed by CFU assays . Alternatively , infected cells were washed extensively and incubated overnight with DMEM containing 10 µg/ml of gentamicin . Cells were then lysed as described above and the released intracellular bacteria were plated and incubated under low-oxygen conditions for 2 days before further incubation under microaerophilic conditions as previously described [14] . The ability of C . jejuni to survive intracellularly was assessed as a percentage of bacteria recovered after 20 hs incubation relative to those recovered 2 hours after infection . | Campylobacter jejuni is one of the most common causes of food-borne illness in the United States and a major cause of diarrheal diseases in developing countries . This pathogen can invade intestinal epithelial cells , which is very important for its ability to cause disease . Once it gains access to epithelial cells , C . jejuni becomes unable to grow under standard growth conditions , although it can grow if pre-incubated under oxygen limiting conditions . This study compares the protein compositions of C . jejuni grown under standard growth conditions and obtained from within epithelial cells . This analysis indicates that , within cells , C . jejuni undergoes a significant metabolic downshift and reprograms its respiration , favoring the respiration of fumarate . These results may provide the bases for the development of novel anti microbial strategies that would target relevant metabolic pathways . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
]
| [
"biology"
]
| 2012 | Quantitative Proteomics of Intracellular Campylobacter jejuni Reveals Metabolic Reprogramming |
Since its discovery in 1969 , enterovirus 71 ( EV71 ) has emerged as a serious worldwide health threat . This human pathogen of the picornavirus family causes hand , foot , and mouth disease , and also has the capacity to invade the central nervous system to cause severe disease and death . Upon binding to a host receptor on the cell surface , the virus begins a two-step uncoating process , first forming an expanded , altered “A-particle” , which is primed for genome release . In a second step after endocytosis , an unknown trigger leads to RNA expulsion , generating an intact , empty capsid . Cryo-electron microscopy reconstructions of these two capsid states provide insight into the mechanics of genome release . The EV71 A-particle capsid interacts with the genome near the icosahedral two-fold axis of symmetry , which opens to the external environment via a channel ∼10 Å in diameter that is lined with patches of negatively charged residues . After the EV71 genome has been released , the two-fold channel shrinks , though the overall capsid dimensions are conserved . These structural characteristics identify the two-fold channel as the site where a gateway forms and regulates the process of genome release .
Enterovirus 71 ( EV71 ) is an emerging pathogen with pandemic potential [1] . First isolated in 1969 , EV71 is classified as a member of the family Picornaviridae , genus Enterovirus [2] , [3] . This single-stranded positive-sense RNA virus is a causative agent of hand , foot , and mouth disease , a childhood illness that is usually mild and self-limiting . However , for some infected individuals , EV71 invades the central nervous system causing severe disease ranging from meningitis to fatal encephalitis [4] . Recently , outbreaks of EV71 have been common in the Asia-Pacific region , with high prevalence of severe disease and mortality being reported [5] . The link between EV71 and worldwide health risks has spurred increased research , as there is no effective treatment or vaccine . Extensive structural and biochemical studies have identified five distinct particles that represent the lifecycle of a human enterovirus: putative procapsid , provirion , mature infectious virus , the altered particle termed the “A-particle” , and empty capsid ( Fig . 1 ) [6] , [7] , [8] , [9] , [10] , [11] , [12] . Assembly begins in the cytoplasm where translation of the viral structural polyprotein and subsequent cleavage allows the formation of protomers consisting of VP0 , VP1 , and VP3 . Five protomers then join together to form the pentameric assembly subunit , twelve of which self-assemble into a naturally occurring empty capsid ( procapsid ) [13] , [14] , [15] . Two distinct assembly pathways have been proposed . In one pathway , this procapsid is a true intermediate , into which the genome would be packaged by some yet unknown mechanism to form provirion [16] , [17] , [18] . Alternatively , in a second proposed pathway , twelve pentamers associate around a viral genome , directly forming a provirion [19] . In this case , the procapsid would serve as a reservoir of capsid components , and represent an off-pathway byproduct of assembly . Regardless of their role in assembly , procapsids are found in all infected cells , and contain 60 copies each of the structural proteins VP1 , VP3 , and VP0 , but are devoid of genome [20] , [21] , [22] . After the provirion is formed , interactions between the structural proteins and the RNA genome cause the autocatalytic cleavage of VP0 into VP2 and VP4 [15] , [23] , [24] . Whereas portions of VP2 map to the capsid exterior , all of VP4 is internal [6] , [7] . The resultant mature virus particles are released from infected cells to continue the lifecycle . Mature viruses must interact with a receptor on the surface of a host cell to initiate the uncoating process and successfully infect the host . For many picornaviruses these interactions rely on the highly conserved architecture of the human enterovirus capsid and the use of a cellular receptor that is composed of an immunoglobulin-like ( Ig-like ) domain [25] , [26] . Surrounding each five-fold axis is a depression , forming a narrow channel referred to as the canyon [6] , [7] . A small opening is present at the bottom of the canyon , which leads to a hydrophobic “pocket” filled with a lipid moiety named the “pocket factor” . The presence of the pocket factor likely stabilizes the mature capsid [27] . One current model for picornavirus uncoating states that the Ig-like receptor binds into the canyon , dislodges the pocket factor , and allows the proteins to shift as rigid bodies . These events lead to the formation of the expanded A-particle . During the expansion process the N-terminus of VP1 is externalized , and VP4 is expelled from the capsid [28] , [29] . These alterations poise the particle for genome release . Subsequently , a second and unknown trigger causes the genome to egress , leaving behind an empty capsid shell [28] . The resulting productive infection will allow the capsid assembly and morphogenesis process to continue anew within the next host . Recent x-ray crystallography studies have revealed the structures of the EV71 procapsid and mature virus [30] , [31] . The procapsid is ∼4% larger than the native virus and has a large hole present at each two-fold axis of symmetry [30] , [31] . The canyon is fully formed , but is shallow compared to other known human enterovirus structures . Additionally , the pocket is collapsed , making it inaccessible to lipids . Relative to the procapsid , the structural protein shell of the mature virus is condensed , and each protomer is rotated 5 . 4° clockwise at each three-fold axis [30] , [31] . These protein movements open the pocket , allowing for a lipid to be inserted . The EV71 pocket factor was found to be longer compared to those of other picornaviruses , and is exposed at the tip [31] . Additionally , the holes at each two-fold axis are closed . These structural rearrangements place the mature virus into a conformation capable of proceeding through the uncoating process . Five cellular receptors have been identified for EV71 , P-selectin glycoprotein ligand-1 ( PSGL-1 ) , annexin II , heparin sulfate , sialyted glycoprotein , and scavenger receptor B2 ( SCARB2 ) [32] , [33] , [34] , [35] , [36] , [37] . The exact function of each of these proteins in virus attachment and entry has yet to be discerned . Surprisingly , none of the identified receptors have an Ig-like domain . However , the shallow nature of the EV71 canyon suggests that an immunoglobulin fold may not be necessary to penetrate the canyon , displace the pocket factor , and trigger the uncoating process [30] , [31] . Though the binding footprint for SCARB2 is not known , mutational evidence suggests that it may act in this manner to trigger A-particle formation and subsequent genome release [38] . Importantly , receptor use is not the only way to initiate picornavirus uncoating . Many studies have shown that heating the virus in vitro is an acceptable surrogate for receptor binding [39] . Wang et al . has reported that heated EV71 mature capsids crystallize isomorphously with procapsid , indicating that both types of empty particles are expanded and retain the same surface features [30] . To elucidate the structural rearrangements that occur during the EV71 uncoating process , we have reconstructed high-resolution cryo-electron microscopy ( cryoEM ) maps of the A-particle ( 6 . 3 Å ) and empty capsid ( 9 . 2 Å ) . Structurally similar to the procapsid [30] , [31] , these new cryoEM structures have an expanded form with large open channels at the icosahedral two-fold axes of symmetry that are lined with patches of negatively charged residues . The A-particle two-fold openings are larger in diameter and the internal capsid residues link the protein shell to the viral RNA . The location of the N-terminal portion of VP1 that is disordered in the procapsid crystal structure , maps to pentameric pillars of density that connect to the base of the canyon . The A-particle and empty capsid have remarkably similar structures , with small differences in protein composition . The cryoEM maps of the A-particle and empty capsid provide a detailed view of uncoating for EV71 . These structural characteristics specifically identify the channel at the two-fold axis as the site for the formation of a gateway , through which the process of viral genome release is regulated .
Purified EV71 mature virus was heated to 56°C for 12 minutes to produce A-particles ( see Materials and Methods ) . Electron microscopy revealed that the heated capsids were in one of two forms: 1 ) the A-particle , with genetic material remaining encapsidated ( ∼85% ) , or 2 ) the empty capsid containing no genetic material ( ∼15% ) ( Fig . S1 ) . As A-particles could be easily distinguished from empty capsids , the sample was vitrified and used for cryoEM data collection . These data were used to produce 3D reconstructions of the A-particle ( 6 . 3 Å ) and the empty capsid ( 9 . 2 Å ) ( Table 1 , Fig . 2A , C ) . Both capsid forms were radially expanded compared to the native virus and measured ∼30 nm in diameter [30] , [31] . The A-particle interior contained density corresponding to the RNA genome , whereas the empty capsid was completely devoid of internal density ( Fig . 2B , D ) . Both capsid conformations had open channels at the two-fold axis of symmetry , consistent with the expanded native empty capsid crystal structure [30] . The five-fold vertex opened at the surface of the capsid in both structures , though the resulting channel through the capsid shell was plugged in the interior . The canyon existed as a thin shallow depression around each five-fold vertex . Fitting of the procapsid crystal structure ( PDB accession code 3VBU ) into the EV71 A-particle cryoEM density map ( cc = 0 . 82 ) identified specific protein-RNA interactions [30] . The first ordered residue in the N-terminal portion of VP1 in the procapsid crystal structure is His73 . This residue is positioned directly within the density pillar that connects the capsid shell to the interior RNA ( Fig . 3 ) . Without further analysis , the unfilled cryoEM density in the EV71 A-particle map could be attributed to RNA , the N-terminus of VP1 , or a combination of the two molecules . Difference map analysis was used to identify the nature of the unfilled density remaining after fitting the procapsid crystal structure into the cryoEM map of the EV71 A-particle . In order to ensure this analysis was highly rigorous , equal quality maps were generated for the EV71 procapsid structure , and the EV71 A-particle map . Three difference density maps were calculated to make the analysis: A-particle minus empty capsid , A-particle minus procapsid , and empty capsid minus procapsid ( Fig . 4 ) . The density differences between the EV71 A-particle and empty capsid were limited to a region in the capsid interior that did not contact the protein shell ( Fig . 4A ) . The lack of interaction between the capsid interior and difference density unequivocally indicates that it corresponds to viral RNA . Subtracting the density of the procapsid from the A-particle left density in the center of the particle that protrudes outward in columns towards a region between the two-fold and five-fold axes ( Fig . 4B ) . This density does contact the interior of the capsid shell , indicating that it can be attributable to both protein and RNA . The density that remained when the procapsid was subtracted from the empty capsid corresponded to a region at the base of the canyon on the capsid exterior and as extra density at the interior shell directly below the canyon . There is an additional difference density on the interior of the empty capsid that is in the same position as the density pillars in the A-particles ( Fig . 4C ) . The procapsid and empty capsid are both void of genome , thus this density must be ascribed to protein . Therefore the portion of the pillar density present as difference density in Fig . 4C and the canyon-associated density must be attributed to protein that is present in the A-particle and empty capsid , but missing or disordered in the procapsid crystal structure . Neither the A-particle nor empty capsid contains VP4 , so the only protein that could fill these densities is the first 72 N-terminal residues of VP1 . To determine if the disordered residues of VP1 in the procapsid crystal structure could fill the density pillar and extend through the capsid shell , a density volume analysis was done . Icosahedral symmetry operators were applied to amino acid residues 1–72 of VP1 from the crystal structure of mature virus ( 3VBS ) [30] and surface contoured at 1 sigma . The volume of this structure was measured to be 116 . 8e3 Å3 ( data not shown ) . The volume of the empty capsid minus procapsid difference map when displayed at a contour level of 1 sigma was 97 . 81e3 Å3 ( data not shown ) . Thus , the disordered N-terminus of VP1 in the EV71 procapsid , is large enough to occupy a portion of the pillar density , extend through the capsid , and become externalized in the A-particle and empty capsid . To compare maps of similar resolution , the density map of the EV71 A-particle generated from ∼2000 particles was used to compare the diameter of the channel present at the two-fold axis with the same channel in the empty capsid . Fitting 3VBU into the A-particle revealed that two α-helices from adjacent VP2 proteins flank the opening of the two-fold channel , and a portion of the fitted structure ( cc = 0 . 82 ) extends outside of the A-particle cryoEM density ( Fig . 5A ) . Alternatively , the same crystal structure fitted into the empty capsid map ( cc = 0 . 93 ) showed the α-helices fully within the density ( Fig . 5B ) . Three residues of VP2 ( Lys 52 , Thr 54 , and Val 58 ) are out of density in the fitted A-particle , whereas only Val 58 protrudes from the density in the fitting of the empty capsid ( Fig . 5A , B ) . The two-fold opening in the A-particle has a measured diameter of 9 . 3 Å , whereas the empty capsid has an opening of only 6 . 4 Å , a 2 . 9 Å difference ( Fig . 5C , D ) . This size difference suggests that the diameter of the channel is linked to the presence or absence of RNA . The electrostatic potential of the fitted crystal structure ( 3VBU ) was determine by Coulomb's Law , and the resultant surface charges were modeled on the fitted crystal structure . The A-particle shows that the interior surface has large patches of negative charge at the two-fold axis of symmetry with minimal interspersed positive charges ( Fig . 6A ) . In contrast , the five-fold vertex is lined with positively and neutrally charged residues ( Fig . 6B ) . These charge distributions are similar in the empty capsid ( data not shown ) . The positive charges may attract the highly negative RNA towards the two-fold channel , whereas the large number of negatively charged amino acids would prevent the genome from becoming trapped in the channel . By this mechanism , the RNA may be positioned in the A-particle so that it can be released from a site between the two-fold and quasi three-fold axis ( see discussion ) .
The new cryoEM structures presented here provide structural details of the EV71 lifecycle . A radial expansion takes place during uncoating to form the expanded A-particle and empty capsid . A previous study reported that heated mature EV71 particles , which had released their genome , crystallized isomorphously with procapsids [30] . The high correlation of the crystal structure to the cryoEM maps presented here , the nearly identical surface features of the A-particle compared to the empty capsid , and the lack of discernible difference density confirm the synonymous topographies of all types of expanded EV71 capsid forms . However , the procapsid and empty capsid vary in protein composition . The 81 N-terminal residues of the uncleaved VP0 protein and the N-terminus of VP1 are located on the interior of the procapsid . In contrast , the downstream empty capsid that is formed after the cleavage of VP0 has extruded VP4 protein and externalized the N-terminus of VP1 . These differences are reflected in the density at the interior surface of the empty capsid . In the procapsid crystal structure , 72 residues of the N-terminus of VP1 are disordered [30] . The cryoEM data and calculated difference maps show these residues constitute a significant portion of the pillar of density extending from the interior capsid surface towards the center of the capsid in both the A-particle and empty capsid . The remaining pillar density could be attributable to viral RNA . Volume calculations show that the disordered residues of VP1 can account for the volume differences shown in the calculated difference maps ( Fig . 4C ) . Additionally , the procapsid has a flattened opening at the base of the canyon [30] . This opening was not observed in the A-particle or empty capsid map . The corresponding region was identified as difference density ( Fig . 4C ) suggesting that the N-terminus of VP1 exits the EV71 A-particle at the base of the canyon and remains externalized in the empty capsid . Aside from the presence of viral RNA the most notable difference between the EV71 A-particle structure and the empty capsid , is the diameter of the two-fold channel . These differences may account for the lower correlation coefficient obtained from fitting the procapsid crystal structure into the A-particle ( cc = 0 . 82 ) compared to the empty capsid fitting ( cc = 0 . 93 ) . The two-fold channel is formed by the separation of VP2 α-helices from adjacent protomers , during the transition of native virus to A-particle [30] , [31] . Over all , these two-fold channels are broader and more open in the A-particle ( Fig . 5 ) , likely to facilitate the release of ordered RNA , an important feature as the termini of the picornavirus genome retains large amounts of secondary structure [40] . A conserved feature of picornaviruses is the amphipathic helix at the extreme N-terminus of VP1 , which associates with host cell membranes as a critical event in RNA release [28] , [41] , [42] . In mature poliovirus , the N-terminus of VP1 is located near the five-fold axis [7] . However , in poliovirus A-particles the location of the N-terminus of VP1 shifts towards the tips of the three-fold propellers , and becomes externalized , as shown by antibody fragment labeling studies [43] , [44] . The difference maps calculated here reveal two unfilled densities: one that is localized to the interior as a protein pillar , and another that is associated with the canyon ( Fig . 4C ) . Volume calculations suggest that the N-terminus of VP1 is sufficient to fill these densities . The hole identified at the base of the canyon in the expanded EV71 procapsid is also present in the A-particle and empty capsid , and likely serves as the site for the externalization of the N-terminus of VP1 [30] . The location of the VP1 protein , its interaction with the viral genome , and the expansion of the two-fold channel would likely position the RNA to be released on a line between the two-fold axis and canyon . This site of genome egress has also been observed for poliovirus in a tomographic study [45] . The models presented here predict that concentrated patches of negatively charged amino acid residues , with interspersed regions of minimal positive charge , surround the two-fold channels . This local high concentration of negative charge has also been reported with human rhinovirus 2 [12] . These positive charges may serve to attract the viral genome , while the negative charges would prevent the RNA from sticking to the interior of the capsid , creating a “slippery” channel that begins at the two-fold axis and continues towards the base of the canyon for genome egress . Furthermore , the predicted positive “attractive” charges and the density plug below the five-fold vertex make the five-fold pore an unlikely site of genome release , as was once thought possible for other picornaviruses . The data presented here provide for a more complete model of picornavirus uncoating . In the mature EV71 capsid the two-fold channels are closed . The virus remains in a metastable state , transiently exposing the N-terminus of VP1 in a process deemed “breathing” , and is capable of withstanding the harsh extracellular environment [46] , [47] . When the capsid interacts with its receptor on a host cell membrane the pocket factor is expelled from the base of the canyon [26] . This release allows the capsid to expand and open the two-fold channels , leading to formation of the A-particle , in which a portion of the N-terminus of VP1 is tethered to the RNA genome . The VP4 proteins are expelled , and insert into the membrane to form a pore ( Fig . 7 ) . The extreme N-termini of VP1 extend from the opening at the canyon base , anchoring the amphipathic helices into the membrane . These events poise the RNA for release . It has been proposed that at one unique site on the A-particle the thin protein layer between the openings at the two-fold axis and canyon base breaks , creating a large pore for genome release [30] . The structures presented here indicate that the movement of the N-terminus of VP1 away from the stabilizing VP2 α-helices at the two-fold , and towards the base of the canyon regulates the formation of a gateway in the protein shell , through which the genome is released .
Confluent HeLa cell monolayers were infected with EV71 strain 1075/Shiga at an MOI of 0 . 1 in culture with DMEM supplemented with 2 . 5% fetal bovine serum . Cells and media were harvested 24 hours post infection and subjected to three freeze-thaw cycles . To remove cell debris , the lysate was centrifuged at 13K rpm in a SLA1500 rotor at 4°C for 15 minutes . After the addition of 8% PEG 8K and 0 . 5 M NaCl virus was precipitated overnight at 4°C and then centrifuged in a SLA1500 rotor ( 4°C , 13K rpm , 45 minutes ) . Pellets were resuspended in purification buffer ( 10 mM Tris-HCl , 200 mM NaCl , 50 mM MgCl , pH = 7 . 5 ) and 0 . 05 mg/ml DNase was added . The suspension was incubated at room temperature for 10 minutes with gentle rocking . After incubation 0 . 1 M EDTA pH = 8 . 0 was added ( 10% total volume ) , the pH was readjusted using ammonium hydroxide , and supernatant was cleared by low speed centrifugation . The supernatant was then pelleted through a 30% sucrose-buffer cushion ( 50 . 2ti rotor , 48K rpm , 4°C , 90 minutes ) . The pellet was resuspended in 2 mL purification buffer , centrifuged at 4000K to remove any remaining cellular debris , and applied to a 10–35% potassium tartrate step gradient for final purification by ultracentrifugation ( 36K rpm , 4°C , 2 hours , SW41 rotor ) . Two distinct bands of virus were collected by side puncture and buffer exchanged into purification buffer . As previously reported [22] , [30] the upper band consisted of procapsid , characterized by the presence of uncleaved VP0 and lack of genomic material . The lower band consisted of native virus comprised of VP2 and VP4 , with packaged genome . Each species was diluted to a concentration of 0 . 1 mg/ml for storage at 4°C in purification buffer . To produce A-particles and empty capsids , purified mature EV71 particles concentrated to 0 . 1 mg/mL in 10 mM Tris-HCl , 20 mM NaCl , 5 mM MgCl , pH = 7 . 5 buffer were heated to 56°C for 12 minutes . The presence of both A-particles and empty capsids was confirmed by negative stain electron microscopy using the JEOL 1400 transmission electron microscope in the microscopy imaging shared core facility at the Pennsylvania State University College of Medicine . Aliquots of sample were blotted onto glow discharged holey carbon Quantifoil electron microscopy grids and plunged into a mixture of liquid ethane and propane ( 50∶50 ) cooled by liquid nitrogen using a FEI Vitrobot mark III freezing robot ( FEI , Hillsboro , OR ) [48] . Data were collected in a FEI Tecnai TF-20 FEG electron microscope operating at 200 kV with magnification 50 , 000× , on Kodak SO-163 film ( Kodak , Rochester , NY ) . The films were scanned using a NIKON Super Coolscan 9000 with a sampling rate of 6 . 35 micron/pixel to generate a pixel size of 1 . 27 Å at the sample . Empty capsids were selected individually , whereas A-particles were selected using semi-automated boxing with the e2poxer . py module of EMAN2 [49] . The data were then linearized , normalized , and apodized using AUTO3DEM [50] . During the reconstruction process using AUTO3DEM , phases were flipped to correct for contrast transfer function ( CTF ) . The resolution of each map was assessed by comparing half-dataset maps using Fourier shell correlation ( FSC ) at a cutoff of 0 . 5 ( Fig . S1 ) . The cryo-EM maps of the A-particle and empty capsid have been deposited in the Electron Microscopy Data Bank ( accession codes EMD-5465 and EMD-5466 , respectively ) . A similar reconstruction procedure was used to calculate a second A-particle map from 2000 particles that was limited to 8 . 7 Å resolution . This map was used for comparisons with the empty capsid map ( Table 1 ) . The asymmetric unit of the expanded naturally occurring EV71 empty capsid ( 3VBU ) was obtained from the PDB and used to generate an icosahedral map in Chimera with a calculated pixel size = 1 . 3 Å [25] . The pixel size of the empty capsid cryoEM density map ( σ = 1 . 00 ) was altered by 0 . 01 increments while assessing the correlation coefficient ( cc ) compared to the calculated map with a finite pixel size = 1 . 3 Å . The value that corresponded to a cc closest to 1 provided the absolute pixel size of both the A-particle and empty capsid density maps ( pizel size = 1 . 27 Å , Table 2 ) . To obtain an accurate fit for the A-particle , a single asymmetric unit of the EV71 naturally occurring empty capsid ( 3VBU ) was preliminarily fitted into the A-particle density map ( σ = 1 . 00 ) using “Fit to Map” in Chimera and icosahedral symmetry operators were applied to generate the structure of the entire capsid . The resulting full map was then docked as a rigid body into the A-particle cryoEM density using Situs to obtain a refined fit [33] . This process was repeated for the density map of the empty capsid . The coordinates of the protomers fitted into the empty capsid and A-particle were deposited in the Protein Data Bank ( accession codes 3J23 and 3J22 , respectively ) . A difference map was calculated by subtracting the density of the 9 Å EV71 empty capsid ( pixel size = 1 . 27 Å ) from the density of the EV71 A-particle map of similar quality [51] . A calculated procapsid map was produced using Situs by mass-weighting the atoms and writing the map to a resolution of 6 . 5 Å . This map was subtracted from the density of the A-particle and the empty capsid to produce two separate difference maps in Situs [51] . Volume measurements of the disordered N-terminus of VP1 and the empty capsid minus procapsid difference density map were completed using the “Measure Volume and Area” tool in Chimera [52] . Chimera was used to predict the Coulombic charges of the fitted crystal structure . The capsid-RNA interactions in the A-particle , measurements , and images were completed using Chimera with maps displayed at a contour level of 1 sigma [52] . | In a picornavirus capsid structural integrity must not be compromised until a key mechanism triggers genome release into a permissive cell . It has long been established that the majority of members of the picornavirus family solve this dilemma with a two-step uncoating process initiated by receptor recognition . For human enteroviruses , binding of an entry receptor triggers a series of conformational changes , resulting in an “A-particle” that is primed for genome release . After endocytosis , an unknown trigger causes the A-particle to expel the viral genome , leaving behind an emptied capsid . This process can be mimicked in solution by heating mature virus . Though the capsid species for both of these steps have been isolated , the fine details of the uncoating process have yet to be elucidated . Cryo-electron microscopy reconstructions of the enterovirus 71 A-particle and empty capsid provide compelling structural evidence to suggest that the icosahedral two-fold axis opens a channel that acts as a gateway in the viral capsid , regulating the release of genomic material from the altered particle . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"virology",
"emerging",
"viral",
"diseases",
"biology",
"microbiology",
"viral",
"structure"
]
| 2013 | The Enterovirus 71 A-particle Forms a Gateway to Allow Genome Release: A CryoEM Study of Picornavirus Uncoating |
Although lactic acidosis is a prominent feature of solid tumors , we still have limited understanding of the mechanisms by which lactic acidosis influences metabolic phenotypes of cancer cells . We compared global transcriptional responses of breast cancer cells in response to three distinct tumor microenvironmental stresses: lactic acidosis , glucose deprivation , and hypoxia . We found that lactic acidosis and glucose deprivation trigger highly similar transcriptional responses , each inducing features of starvation response . In contrast to their comparable effects on gene expression , lactic acidosis and glucose deprivation have opposing effects on glucose uptake . This divergence of metabolic responses in the context of highly similar transcriptional responses allows the identification of a small subset of genes that are regulated in opposite directions by these two conditions . Among these selected genes , TXNIP and its paralogue ARRDC4 are both induced under lactic acidosis and repressed with glucose deprivation . This induction of TXNIP under lactic acidosis is caused by the activation of the glucose-sensing helix-loop-helix transcriptional complex MondoA:Mlx , which is usually triggered upon glucose exposure . Therefore , the upregulation of TXNIP significantly contributes to inhibition of tumor glycolytic phenotypes under lactic acidosis . Expression levels of TXNIP and ARRDC4 in human cancers are also highly correlated with predicted lactic acidosis pathway activities and associated with favorable clinical outcomes . Lactic acidosis triggers features of starvation response while activating the glucose-sensing MondoA-TXNIP pathways and contributing to the “anti-Warburg” metabolic effects and anti-tumor properties of cancer cells . These results stem from integrative analysis of transcriptome and metabolic response data under various tumor microenvironmental stresses and open new paths to explore how these stresses influence phenotypic and metabolic adaptations in human cancers .
Human cancers are extremely heterogeneous due to multiple mutations in oncogenes and tumor suppressor genes , a range of inherited germline variations and varying degrees of microenvironmental stresses . These tumor microenvironmental stresses include tumor hypoxia , accumulation of lactic acid ( lactic acidosis ) and depletion of glucose , glutamine and other nutrients [1] . These stresses are largely caused by a combination of poor tissue perfusion , abnormal tumor vasculature , uncontrolled proliferation and dysregulated energy metabolism of cancer cells during tumor development and progression . Importantly , these microenvironmental stresses also directly modulate physiological and metabolic phenotypes of cancer cells and ultimately affect the clinical outcomes of patients . With major variations known to exist among different tumors , advances in the pretreatment assessment of the influences of these stresses will aid in improved selection of appropriate therapeutic strategies for individual patients . These stresses and their downstream effects are also the targets of cancer therapeutics , including anti-angiogenesis and hyperthermia treatments . It is therefore important to fully understand the impact and mechanism of how these stresses affect various tumor and non-tumor cells in human cancers . It is well known that cells resort to glycolysis instead of oxidative phosphorylation to utilize glucose as energy source during hypoxia . In addition , cancer cells have a preferential use of glycolysis pathways for energy generation even in the presence of oxygen – so called “aerobic glycolysis” as first proposed by Dr . Otto Warburg [2] . These factors all likely contribute to high glucose flux and form the basis of using glucose analog 18F-FDG to detect tumor cells . Such dysregulated metabolisms in cancer cells also lead to the accumulation of the metabolic product of glycolysis – lactic acids in solid tumors . Many measurements have been performed to determine the level of tumor lactate and significant variations were found , with the medium range of 7–10 mM/g and up to 25 . 9 mM/g [3]–[5] . These studies show that high tumor lactate levels are typically associated with more aggressive tumors and resistance to treatment [3]–[5] . How lactic acidosis affects tumor and non-tumor cells in human cancers has been the focus of many elegant studies as summarized in several reviews [6]–[8] . The exposure of cultured cells to lactic acidosis in vitro has been shown to trigger calcium signaling [9] , induce angiogensis gene expression ( e . g . VEGF , IL8 ) [10]–[12] , HIF-1α stabilization [13] and cell death [14] . Recent genomic analyses identify the transcriptional responses of different cell types to acidosis and high lactate [15]–[17] . These in vitro studies have clearly shown the significant impact of lactic acidosis on the gene expression and phenotypes of cancer cells . However , it is often challenging to relate the influences of these microenvironmental stresses in vitro to the complex cancer phenotypes in vivo . We have previously overcome this challenge by defining “gene signatures” based on sets of genes whose expression levels are altered by lactic acidosis in vitro as quantitative “common phenotypes” , and projecting them to the in vivo microarray expression data of patients' tumors [18] . We have used this approach to successfully investigate the pathways of hypoxia [19] , vascular injury [20] and lactic acidosis [18] in human cancers . This reciprocal flow of information between the in vitro and in vivo systems demonstrate that lactic acidosis triggers significant metabolic reprogramming by forcing cells to rely more on oxidative phosphorylation as an energy source with the suppression of glycolytic phenotypes [18] . Therefore , cellular metabolism is critical in determining the impact of lactic acidosis to tumor phenotypes and clinical outcomes . In the use of microarrays to connect cultured cells with human cancers , primary epithelial cells are often used to provide the common and shared responses to the defined perturbations due to their intact genetic materials and signaling circuitry [18]–[22] . However , given our intention to relate gene signatures to human cancers , it is probably more relevant to assess signatures in cancer cells . Cancer cell lines are also often easier to be transfected , thus allowing genetic manipulations for mechanistic studies . In this study , we used breast cancer cell line , MCF-7 , to conduct a detailed temporal analysis of lactic acidosis response and compare the transcriptomic and metabolic responses of cancer cells to hypoxia , lactic acidosis and glucose deprivation in parallel to gain a further understanding of the underlying molecular mechanisms . One important goal in understanding the cellular responses to tumor microenvironmental stresses is to identify the key regulator ( s ) responsible for these observed gene expression response since such understanding will lead to important insights into the development and progression of human cancers . For example , the identification of hypoxia-inducible transcription factors ( HIFs ) as central regulators in the hypoxia response and its regulation at the level of protein stability has become crucial in our understanding of tumor hypoxia [23]–[26] . HIF-1α protein stabilization can also be seen in the development of multiple neoplasms in patients with von Hippel-Lindau disease [27] or mutations in several enzymes of the tricarboxylic acid ( TCA ) cycle , such as succinate dehydrogenase ( SDH ) and fumarate hydratase ( FH ) [28]–[30] . The glucose deprivation is another feature of tumor microenvironmental stress caused by the imbalance between supply and consumption [31] . Glucose deprivation triggers “starvation-like” signaling through the activation of AMPK and LKB-1 , which in turn activates TSC1/TSC2 and inhibits central energy sensor mTOR activities to inhibit ribosomal biogenesis , translation activities and proliferation [32] . The mutations in LKB-1 and TSCs in the glucose sensing pathway lead to cancer development in the Peutz-Jeghers Cancer Syndrome [33] and Tuberous Sclerosis [34] , [35] , respectively . In contrast , very little is known about the transcriptional regulation of the lactic acidosis gene response program . To uncover the molecular mechanisms by which tumor cells respond to their microenvironment , we conducted a detailed temporal transcriptional analysis of the MCF7 breast cancer cell line under lactic acidosis and compared this response to those elicited by glucose deprivation and hypoxia . We identified a novel growth suppressive pathway ( MondoA-TXNIP ) that portends better prognosis in breast cancer and contribute to the anti-Warburg effects of lactic acidosis .
We examined the temporal gene expression patterns of a breast cancer cell line MCF-7 exposed to lactic acidosis ( 25 mM lactic acid with pH 6 . 7 ) at various time points during the first 24 hours . Cells were first brought to replicative arrest by serum withdrawal for 24 hours before exposure to either control or lactic acidosis conditions composed of pH 6 . 7 and 25mM lactic acid in the presence of high levels of glucose ( 4 . 5 g/L ) in triplicate of different time points at 1 , 4 , 12 and 24 hrs to characterize the temporal changes of gene expression patterns . The RNA samples harvested from these MCF-7 cells were interrogated with Affymetrix GeneChip U133 plus 2 . 0 arrays ( ∼54 , 000 probe sets on ∼47 , 000 transcripts and variants ) with results deposited in Gene Expression Omnibus ( GSE19123 ) . Gene expression profiles of all MCF-7 cells were normalized by RMA , zero-transformed against the average expression levels of the same probe sets of the time-matched control samples as performed previously [19] , [20] . 1761 probes sets showing with at least two fold changes in at least two samples were selected and arranged by hierarchical clustering according to similarities in expression patterns ( Figure 1A ) . This analysis showed that lactic acidosis induced a dramatic change in the gene expression with significant temporal patterns ( Figure 1A and 1B ) . Among the genes induced in MCF-7 by lactic acidosis , we found PLAU , major histocompatility complex ( MHC ) type I and CD44 , and REDD1 ( Figure 1B ) , a p53 transcriptional target following DNA damage [36] . The expression of CD44 has been reported to be induced by lactosis in cancer cells [37] . Tumor necrosis factor ( TNF ) , Fas and connective tissue growth factors ( CTGF ) were only induced in earlier time points and returned to baseline in the later time points ( Figure 1B ) . Among clusters of genes repressed by lactic acidosis in the later time points is a large group whose expression is closely linked to cell proliferation ( cyclin D , PCNA , CCRK , E2F3 , and E2F6 ) ( Figure 1C ) . These lactic acidosis-repressed genes may reflect a physiological alteration that halts cell proliferation as the cells try to preserve energy consumption under metabolic stress [31] , [32] . To define the lactic acidosis response in MCF-7 at a pathway level , we performed Gene Set Enrichment Analysis ( GSEA ) [38] to compare the pathway composition in the gene expression of all control vs . lactic acidosis samples of MCF-7 cells . We found that samples exposed to lactic acidosis were enriched in gene sets representing nutrient deprivation [39] , the treatment of histone deacetylase inhibitor trichostatin A ( TSA ) in cancer cells , breast cancer good prognosis [40] and exposure of DNA damaging agent ( bleomycin ) ( Table S1 ) . The gene expression of MCF-7 which has been exposed to lactic acidosis was depleted in gene sets representing E2F1 target genes , DNA replication , breast cancer poor prognosis [40] , mitotic cycles and RNA processing ( Table S1 ) . We previously showed that the hypoxia and lactic acidosis response signatures elicited in cultured primary non-cancerous cells HMECs could provide a molecular gauge of hypoxia and lactic acidosis response for cancerous human tissues in vivo , and also predict clinical outcomes . To test robustness of this gene signature approach based now on cancerous cells , the lactic acidosis gene signatures generated in MCF-7 were projected to the Miller breast cancer data to evaluate a tumor-specific numerical score representing the predicted signature of lactic acidosis – i . e . , quantifying lactic acidosis pathway activity across tumors [41]; see Materials and Methods section below and the detailed statistical supplemental materials from a previous study [18] . Statistical survival analysis then indicates that patients with tumors showing higher levels of lactic acidosis pathway activity – defined by projected signatures of 12 and 24 hours exposure in MCF-7 – had significantly better clinical outcomes ( Figure 1D ) , consistent with our previous studies using HMEC [18] . The lactic acidosis gene signature at 12 hour time point had the most consistent prognostic significance across different tumor datasets , including three other breast cancer expression studies with different stages of diseases ( Figure S1 ) . These datasets include a study of 286 lymph node negative early breast cancers from NKI ( Wang ) [42] , and two studies of invasive breast carcinomas ( Sotiriou , Pawitan ) [43] , [44] . We also measured the reproducibility and consistency of the predicted lactic acidosis pathway activities in each tumor using gene signatures generated in both MCF-7 and HMECs . We found that the two variants of the lactic acidosis signature resulted in highly similar predicted lactic acidosis activity , evidenced by strongly correlated signature scores in Miller and other tumor datasets ( Figure 1E and Figure S2 ) . This result suggests that although there are differences between the lactic acidosis responses at the gene level , the pathway level predictions based on the gene signatures are similar and reproducible in both normal primary and breast cancer cells . In the GSEA analysis for the MCF-7 cells exposed to lactic acidosis , we found enrichment in the starvation pathways caused by the deprivation of glucose and glutamine obtained in an independent study [39] . Such association between lactic acidosis and nutrient deprivation was consistent with the reduced ATP production under lactic acidosis [18] . To test for such association , we directly compared the gene expression of MCF-7 cells of lactic acidosis with cellular starvation stress caused by glucose deprivation . We performed parallel global transcriptional analysis on the MCF7 cells in quintuplicate which had been exposed to lactic acidosis and 1% oxygen ( hypoxia ) in the media with high glucose ( 4 . 5g/L ) and glucose deprivation ( using media 0g/L glucose ) for 4 hours . Similar parallel analysis was performed for subsets of genes involved in metastasis and invasive capacity [45] , [46] . RNAs from these MCF-7 cells were extracted and hybridized to Affymetrix GeneChip U133 plus 2 . 0 arrays , normalized by RMA and zero-transformed by deducting the mean expression values for samples under control condition . 3903 probe sets were selected by the criteria of at least two observations with at least two fold changes and arranged by hierarchically clustering ( Figure 2A and 2B ) . Unexpectedly , we noted a highly similar transcriptional response to lactic acidosis and glucose deprivation which is distinct from the hypoxia response ( Figure 2A ) . The similarity of the lactic acidosis and glucose deprivation response was observed in both induction and repression of a large number of common sets of genes , including major histocompatility complex ( MHC ) type I & II , DNA repair gene , alkB ( alkyation repair homolog 7 ) , and CTGF , Jun and TNF ( Figure 2B ) , which were also seen in our previous time course experiment ( Figure 1B ) . In contrast , hypoxia elicited a very distinct transcriptional response with the upregulation of many known hypoxia-inducible genes such as EGLN3 , CA9 , stanniocalcin1 ( STC1 ) , BNIP3 and genes involved in glycolysis , including pyruvate kinase ( PKM2 ) ( Figure 2B ) [19] . In a previous microarray study of the stress response in yeast Saccharomyces cerevisiae , the induction of a share set of “common stress genes” was a prominent feature [47] . In contrast , there was only a small cluster of genes which were induced by lactic acidosis , glucose deprivation and hypoxia , which may represent “common stress genes” in MCF-7 cells ( Figure 2B ) . These genes include HIG-2 ( hypoxia-inducible gene ) and REDD , both genes reported to be induced by hypoxia ( Figure 2B ) . Since glucose deprivation is known to trigger cellular starvation response , this similarity in the transcriptional responses suggests that lactic acidosis may also trigger a “starvation” response seen for nutrient deprivation . To test this possibility , we measured the effects of lactic acidosis on several biochemical markers of starvation response in MCF-7 ( Figure 2C ) . AMP-activated protein kinase ( AMPK ) is a highly conserved energy-sensing heterotrimeric complex that plays a key role in the regulation of energy homeostasis; it becomes phosphorylated at Thr172 of AMPK α by an elevated AMP/ATP ratio due to various stress indicating energy stress and starvation . We find that lactic acidosis significantly increased Thr172 phosphorylation of AMPK , even in the presence of high levels of glucose and amino acids in the media ( Figure 2C ) . The Mammalian Target of Rapamcyin ( mTOR ) is another crucial cellular sensor for energy status and a crucial downstream target of AMPK [48] , [49] . We examined how lactic acidosis affected mTORC1 activities through the phosphorylation of its downstream target S6 kinase ( S6K ) and found it dramatically reduced the S6K phophorylation ( Figure 2C ) . Taken together , these data supported the notion that lactic acidosis triggers a cellular starvation response similar to that of glucose starvation , with AMPK activation and mTOR inhibition . These biochemical changes are likely to contribute to reduced cell growth and proliferation under lactic acidosis as evidenced by gene expression ( Figure 1C ) . Knowing that lactic acidosis can inhibit tumor glycolysis [18] , we evaluated how lactic acidosis and glucose deprivation affect glucose uptake using 2-deoxy-D-[2 , 6-3H]-glucose . While glucose deprivation increased the uptake of the glucose by 84% , lactic acidosis significantly reduced the glucose uptake by 67% ( Figure 3A ) . The inhibition of glucose uptake by lactic acidosis was consistent with the reduced glucose consumption and decreased lactate production we have reported previously . However , this result was not consistent with the expected increased uptake of glucose to increase energy generation associated with AMPK activation and cellular starvation , as seen in glucose deprivation [50] . Therefore , in spite of the shared AMPK activation , mTORC1 inhibition and similar transcriptional responses , lactic acidosis and glucose starvation triggered opposite effects on glucose uptake . To identify transcriptionally regulated genes that may contribute to these opposite metabolic responses , we used the discordance to select genes with significant but opposite expression changes under lactic acidosis and glucose deprivation . This analysis identified 115 probe sets ( Table S2 ) ; 49 were induced by lactic acidosis and repressed by glucose deprivation while 66 were repressed by lactic acidosis and induced by glucose deprivation ( Figure 3B ) . The three probe sets with the largest induction levels under lactic acidosis were all associated with thioredoxin interacting protein ( TXNIP or Vitamin D3-upregulated protein 1 , VDUP1 ) ( Figure 3B ) . TXNIP was strongly induced by lactic acidosis , and suppressed by glucose deprivation . While hypoxia was reported to trigger TXNIP expression [51] , we only noted a modest increase ( ∼40% ) in our experiment . We tested for the prognostic significance of this differentially expressed signature by projecting the 115 gene signature into the Miller breast tumor data set . Due to different arrays used in that data set , this results in 109 of the 115 signature probes present in the Miller data set . We find that higher levels of this signature also indicate favorable outcomes ( Figure S3A ) . The three TXNIP probe sets alone in fact have prognostic significance comparable with that of the signature ( Figure S3B ) ; on re-evaluating the signature after removal of these three TXNIP probe sets , we find that the reduced signature maintains prognostic significance albeit at a slightly reduced level ( Figure S3C ) . When we examined expression of MCF-7 and HMEC under lactic acidosis at different time points , we also noted significant induction of TXNIP and its parralogue alpha-arrestin domain containing 4 ( ARRDC4 ) ( Figure S4 ) . Real-time RT-PCR further confirmed the induction of TXNIP ( up to 4–5 fold ) and ARRDC4 transcripts ( up to 3 . 4 fold ) in MCF-7 ( Figure 3C and Figure S5 ) . In addition , lactic acidosis also induced TXNIP expression in other cancer cell lines , including WiDr ( colon cancer cell ) and SiHa ( cervical cancer cell ) ( Figure S6 ) . TXNIP expression is known to be induced by glucose exposure [52] , [53] . Since there were high levels of glucose ( 4 . 5g/dL ) in both control and lactic acidosis media , we further clarified the role of glucose vs . lactic acidosis in the TXNIP induction using real-time PCR to determine the level of TXNIP in MCF-7 cultured under different glucose and lactic acidosis conditions ( Figure 3D ) . Glucose ( 4 . 5g/dL ) or lactic acidosis alone increased TXNIP levels by approximately 4–5 fold ( Figure 3D ) . In the presence of both high glucose levels and lactic acidosis , TXNIP was induced up to 18 fold ( Figure 3D ) . Thus lactic acidosis and glucose exposure are both potent inducers of TXNIP with synergistic induction potential , suggesting distinct mechanisms of TXNIP induction by these two stimuli . To define the individual contribution of lactosis vs . acidosis , we tested the level of TXNIP induction under different degrees of lactosis ( 12 . 5 or 25 mM lactate ) and acidosis ( pH 6 . 7 ) ( Figure 3E and 3F ) . We found that acidosis alone , but not lactosis alone , led to the TXNIP induction , consistent with our previous study [18] ( Figure 3E ) . Although lactosis alone did not induce TXNIP , the addition of lactate to the acidosis conditions led to a dose-dependent augmentation of the TXNIP induction ( Figure 3E ) . Importantly , this effect on TXNIP induction was present even under 10 or 12 . 5 mM lactate , a level seen in many human tumors ( Figure 3E and 3F ) . We also measured how lactosis and acidosis affected glucose uptake , and found that acidosis , but not lactosis , led to a reduction in glucose uptake , a pattern consistent with TXNIP induction ( Figure 3G ) . Although TXNIP was initially identified as a protein interacting with thioredoxin and modulating cellular responses to oxidative stresses , it has the ability to inhibit glucose uptake and is being recognized as an important regulator of dysregulated metabolism in diabetes [54] , [55] . Therefore , the upregulation of TXNIP under lactic acidosis make it an attractive candidate contributing to the anti-Warburg effects and inhibition of tumor glycolysis under lactic acidosis [18] . To investigate this , we silenced the TXNIP transcripts with two independent siRNAs and confirmed the successful reduction of the TXNIP protein ( Figure 4A ) . Consistent with the previously known role of TXNIP to inhibit glucose uptake [55] , the silencing of TXNIP by siRNAs led to significant decrease in the reduction of glucose uptake under lactic acidosis culture condition ( Figure 4B ) . Lactic acidosis caused 52% repression of glucose uptake in MCF7 cells transfected with non-targeting siRNAs ( - ) . In cells transfected with two different siTXNIPs ( T1 , T2 ) , glucose uptake was increased while the repressing effect of lactic acidosis was decreased to 39% and 44% respectively ( Figure 4B ) . In addition , we found the silencing of TXNIP in MCF-7 increased both the glucose consumption and lactate productions under normal media , but lactic acidosis significantly reduced both the glucose consumption and lactate production in all treated cells ( Figure S7 ) . Although the TXNIP gene silencing reduced the glycolysis inhibition under lactic acidosis , the effects were modest , which may be due to the remaining level of TXNIP . We further tested the effect of TXNIP knocking out in the TXNIP deficient MEF cells [56] . Lactic acidosis caused 68% reduction in glucose uptake in the wild-type ( WT ) MEF cells . In contrasts , lactic acidosis only reduced the glucose uptake by 28% in TXNIP knockout ( TKO ) MEF cells ( Figure 4C ) . To further test for the role of TXNIP upregulation in the lactic acidosis response , we exposed the wild type and TXNIP deficient MEF cells to the lactic acidosis conditions ( 10 mM lactate , pH 6 . 7 ) and performed gene expression analysis using the Affymetrix mouse 430A2 GeneChip and normalized the expression data by RMA . We first examined the effects of the disruption of TXNIP on the gene expression of MEF cells by performing zero transformation of TXNIP deficient against the average expression in the wild type MEF cells ( Figure S8 ) . We found 798 probe sets whose expression was altered at least 1 . 7 fold in at least two samples ( Figure S8 ) . Among the genes repressed in the TXNIP KO MEF cells were TXNIP , COX2 , collagens and many other genes in the HOX genes involved in pattern specification ( Figure S8 ) . Among the induced in the TXNIP KO MEF cells were genes in the complement activation , AP2 , and Notch3 ( Figure S8 ) . To define the role of TXNIP in the lactic acidosis gene expression , we compared the respective lactic acidosis response of the wild type and TXNIP deficient MEF cells by zero-transformation against the corresponding cells cultured in the control conditions . 1327 probe sets showing with at least 1 . 7 fold changes in at least two samples were selected and arranged by hierarchical clustering according to similarities in expression patterns ( Figure 4D ) . This analysis showed that the loss of TXNIP reduced the induction and repression of a significant number of genes as shown in the three gene clusters ( Figure 4D ) , including the repression of many cell-cycle related genes under lactic acidosis . In addition , the loss of TXNIP led to a statistically significant reduction in the overall degrees of induction and repression ( p<0 . 0001 ) for the average expression of the 1049 repressed and 278 induced genes ( Figure S9A and Figure 9B ) . We also tested the ability of lactic acidosis to repress the proliferation of MEF cells and found that the proliferation of the wild type , but not TXNIP deficient MEF cells , were repressed by lactic acidosis ( Figure 4E ) . This observation is consistent with the reduced repression of cell cycle genes under lactic acidosis seen for the TXNIP deficient MEF cells ( Figure 4D ) . Taken together , these data showed that TXNIP induction contributes to the inhibition of glycolysis phenotypes , metabolic reprogramming , and cell cycle arrest and gene expression under lactic acidosis . Glucose-induced TXNIP transcription depends on a short proximal region of the TXNIP promoter . Specifically , this region includes a well conserved Carbohydrate Response Elements ( ChoRE ) consisting of two E-boxes [53] . We first tested the influence of lactic acidosis on the reporter constructs driven by the promoters of TXNIP and ARRDC4 and found that lactic acidosis could induce the reporter activities of both constructs by more than 6 folds ( Figure 5A ) . Importantly this induction of reporter activity under lactic acidosis was reduced by 61% with mutation in the ChoRE of the TXNIP promoter ( Figure 5A ) , indicating the importance of the ChoRE regions to the transcriptional activation of TXNIP under lactic acidosis . The transcriptional activation through two E-boxes in the promoters of TXNIP upon glucose exposure has been reported to be caused by the binding MondoA:Mlx [57] or Carbohydrate Response Elements-binding protein ( ChREBP ) [58] . MondoA is likely to be more relevant in the observed lactic acidosis response of MCF-7 given its higher expression levels in MCF-7 and the simultaneous upregulation of both TXNIP and ARRDC4 under lactic acidosis [57] . MondoA:Mlx complexes are held latently at the outer mitochondrial membrane ( OMM ) , yet shuttle between the OMM and the nucleus suggesting that they facilitate communication between these two essential organelles . MondoA:Mlx complexes are sensors of intracellular glucose concentration and accumulate in the nucleus following increases in glucose-6 phosphate to occupy the E-box-containing promoters of targets such as TXNIP and ARRDC4 [57] . To directly test the inducible physical binding of MondoA to the promoter regions of TXNIP and ARRDC4 under lactic acidosis , chromatin-immunoprecipitation ( CHIP ) was performed with the antibody against MondoA in MCF-7 cells which have been exposed to lactic acidosis conditions of different acidity . We detected increased specific binding of MondoA to the promoter regions of TXNIP and ARRDC4 with more acidic lactic acidosis environments ( Figure 5B ) . These results indicated that the MonoA became activated to occupy the promoters of TXNIP and ARRDC4 under lactic acidosis and these binding sites were important for their transcriptional inductions . To further determine the role of MondoA in the induction of TXNIP under lactic acidosis , we used two different sets of siRNAs to knock down MondoA by gene silencing ( Figure 5C ) . During lactic acidosis , the level of MondoA protein did not change , while TXNIP was induced significantly ( Figure 5C ) . This induction of TXNIP was significantly reduced at both levels when MondoA was reduced by both sets of siRNAs targeting MondoA ( Figure 5C ) . The silencing of MondoA also led to increased glucose uptake under both control and lactic acidosis condition ( Figure 5D ) , a result similar to the silencing of TXNIP ( Figure 3F and 3H ) . While lactic acidosis caused 57% repression in MCF7 cells transfected with non-targeting siRNA , such repression was reduced to 44% and 40% with MCF7 cells transfected with two different siRNAs targeting MondoA ( M1 and M2 ) . The degree of the lactic acidosis-induced repression of the glucose uptake was lower after the silencing of MondoA by two independent siRNAs ( Figure S10 ) . Although gene silencing led to significant reduction of MondoA , the remaining MondoA level may still account for the slight induction of TXNIP under lactic acidosis . To further examine the effects in the absence of MondoA , we tested the MondoA knockout MEF ( mouse embryonic fibroblasts ) cells created by the cre-loxP system ( CWP and DEA , manuscript submitted ) . In the complete absence of MondoA , both the basal and inducible level of TXNIP under 2-DG or lactic acidosis was completed abolished ( Figure 5E ) . Importantly , the re-introduction of full length MondoA back to the MEF cells fully restored the induction of TXNIP under lactic acidosis ( Figure 5E ) . Taken together , these results demonstrate the critical role of MondoA in regulating TXNIP under both glucose exposure and lactic acidosis . To determine the prognostic significance of TXNIP expression in human cancers , we performed survival analyses using TXNIP expression as the sole predictor of survival time . Breast cancers in the Miller dataset stratified based on TXNIP expression were found to have significant differences in clinical phenotypes; tumors with high TXNIP have better survival and clinical outcomes ( Figure 6A , p = 0 . 00115 ) . Similar results were also obtained in the three other breast cancer datasets . To test whether the in vitro correlation of TXNIP induction with lactic acidosis pathway activity persists in vivo , we compared the predicted lactic acidosis pathway activity with the TXNIP expression levels in the breast cancer data sets and found a positive correlation in all four data sets; although the relationship is of limited predictive value with low R values reflecting a noisy relationship , the positive relationship is statistically significant and consistent across tumor data sets ( Figure 6B ) . Similar results were found for ARRDC4 , another paralouge of TXNIP and downstream target of MondoA . When breast cancers were stratified by the level of ARRDC4 , tumors with higher levels of ARRDC4 had better prognosis and clinical outcome in the both breast cancer data ( Miller and Pawitan ) in which ARRDC4 was measured ( Figure 6C ) . The expression of ARRDC4 was also positively associated with the predicted lactic acidosis pathway activity in these two breast cancer datasets ( Figure 6D ) . These observations demonstrate the high degree of in vivo correlation between the expression of TXNIP and ARRDC4 and the lactic acidosis pathway activity in human cancers . It is also consistent with the known role of TXNIP as a probable tumor suppressor in several cancer types where it has been shown to suppress oncogenic phenotypes and its loss of function linked with cancer development [59]–[61] . Therefore , the expression of both TXNIP and ARRDC4 are potentially important mediators of the lactic acidosis response and suggest the important prognostic significance of molecular pathways driven by Mondo-Mlx in human cancers .
Hypoxia , glucose deprivation and lactic acidosis are all well-recognized tumor microenvironmental stresses . Each of these stresses has distinct features – either limited availability of energy fuel ( glucose ) , depletion of cofactor ( oxygen ) or accumulation of metabolic byproducts ( lactic acidosis ) . Previous studies in various settings have shown that glucose deprivation causes metabolic stress and induces adaptive responses —so-called “starvation response” – manifest by AMPK activation , mTOR inhibition and multiple other biochemical and metabolic changes in cancer cells . Tumor hypoxia has also been reported to induce such metabolic stresses [62] . Here , we have shown that lactic acidosis triggers a “starvation” response similar to glucose deprivation as evidenced by the biochemical and transcriptional responses . Since this occurs even in the presence of abundant oxygen and nutrients , this response may represent “pseudo-starvation” . This is especially unexpected given the presence of high levels of lactate as an additional potential energy source for cancer cells [63] . Adaptive changes during the starvation response often reflect a need to preserve energy homeostasis during energy depletion by switching off cell growth and proliferation and many anabolic processes , such as protein , carbohydrate and lipid biosynthesis . In addition , there is also a simultaneous increase in cellular catabolism and breakdown of various energy sources to increase energy production . Even with this shared energy need , different tumor microenvironmental stresses employ different means to accomplish such adaptations to preserve energy homesostasis . During hypoxia , the essential cofactor ( oxygen ) required for oxidative phosphorylation becomes limited . As a result , the transcriptional factor HIF-1α activates genes encoding glucose transporters and enzymes involved in the glycolytic process shift to the use of glycolysis as the main source of energy generation . In the setting of glucose deprivation , there is a lack of fuel required for glycolysis as energy source . During these energy shortages , AMPK become activated to modify many proteins controlling energy flow to increase the energy generation in mitochondria from other source of fuels , such as oxidative of fatty acid and amino acids [64] . In addition , the inhibition of mTOR turns off many energy-consumption processes ( e . g . , translation , cell growth and proliferations ) in an effort to restore energy homeostasis . The lactic acidosis-induced cell cycle arrest may be caused by many changes , including the induction of AMPK , mTOR inhibition , TXNIP and cell cycle arrest caused by the induction p57 , p21 and other inhibitors of cellular proliferation . Moreover , the expression of many genes involved in glycolysis is affected by changes in histone acetylation of chromatin [65] . These responses triggered by glucose deprivation may be also important for energy maintenance and survival in acidosis-induced apoptosis [66] , [67] . For example , AMPK may help in re-directing cells to utilize non-glucose energy sources ( e . g . , fatty acids and amino acids ) and increased mitochondria activities under lactic acidosis [18] , [68] . In addition , since the AMPK activator AICAR reduces the expression of lactate importer monocarboxylate transporter ( MCT ) -1 but increases the expression of the lactate exporter in MCT4 [69] , AMPK activation under lactic acidosis may also inhibit the uptake of excessive cellular lactate . These stress environments are also likely to select for tumor cells that have developed strategies to survive energy deprivation with better energetic balance . The genetic mutations of many genes involved in these processes of adaptation to hypoxia and glucose deprivation are known to associate with tumor development . These findings highlight the crucial role of biochemical , metabolic control of energy homeostasis in tumor development . There are many possible explanations as to how lactic acidosis leads to AMPK activation , mTOR inhibition and other features of the starvation response . For example , lactic acidosis can reduce ATP generation in cells [18] and thus lead to a high AMP/ATP ratio to activate the AMPK . Extracellular acidosis also triggers an increase in cytosolic Ca++ [9] which may activate CaMKK – an alternative AMPK upstream kinase [70]–[72] – to phophorylate and activate AMPK . Lactic acidosis may also cause intracellular nutrient depletion with reduced uptake of glucose ( this study ) and glutamine through the inhibition of acidosis-sensitive glutamine pumps [73] , [74] . This depletion of intracellular pools of nutrients may in turn repress the energy sensor mTOR , the translation activities and ribosomal biogenesis required for cellular proliferation . In the future , it will be important to further dissect the contribution of each factor and signaling components to the induction of these starvation responses under lactic acidosis to further our detailed understanding of their impact on the metabolisms and phenotypes of cancer cells . In both hypoxia and glucose deprivation , there is an increase in glucose uptake and glycolysis to provide essential fuel through induction ( by HIFs ) , modification of glucose transporters ( by AMPK ) [75] , [76] and histone acetylation via ATP-citrate lyase [65] . In contrast , lactic acidosis presents a different instance of starvation with inhibition of glucose uptake and other glycolysis activities in cancer cells [77] . Through these comparisons , we have also dissected two distinct molecular pathways ( AMPK-mTORC1 , MondoA-TXNIP ) by which various microenvironmental stresses influence cancer metabolic phenotypes . While the AMPK-mTOR response is similar under lactic acidosis and glucose deprivation , the MondoA-TXNIP is affected in opposite directions by these two stresses . Given the continuous need for energy generation with reduced ATP and metabolic substrate from the glycolysis pathways under lactic acidosis , cells are likely to undergo extensive metabolic reprogramming to utilize other nutrients as energy sources . This idea is supported by the increased reliance on mitochondria for ATP generation . Our analysis has highlighted the potential roles of TXNIP and AMPK in this metabolic reprogramming . The induction of TXNIP under lactic acidosis and its ability to inhibit glucose uptake and reduce lactic acidosis production from glycolysis form a negative feedback loop . In addition , the loss of TXNIP leads to many features of Warburg effects and glycolytic phenotypes of cancer cells , opposite to the influences of lactic acidosis [18] . High TXNIP expression is associated with favorable outcomes , consistent with its postulated role as a tumor suppressor gene based on its growth-suppressing activity and the increased occurrences of tumors with deficiency of TXNIP [59] , [60] , [78] , [79] . These findings may provide important insights into the regulatory mechanisms of TXNIP as well as the phenotypic alterations under lactic acidosis . TXNIP may affect the glucose uptake through the suppression of two important regulators of glycolysis , Akt [56] and HIF-1α [80] . In addition , TXNIP is known for other properties , which may explain its effect on the gene expression during lactic acidosis . TXNIP is a negative regulator of cellular oxidative tolerance by binding thioredoxin [81]–[83] and as a feedback regulator of S-nitrosylation [84] relevant in the cellular adaptive response to tumor microenvironmental stresses . For example , repression of TXNIP enhancing the anti-oxidative capacity during glucose deprivation may be required to cope with increased mitochondria oxidative stresses [85] . 2-Deoxy-D-glucose ( 2-DG ) can enhance the cytotoxicity of cisplatin through mechanisms involving increased oxidative stress [86] . Since 2-DG is a strong inducer of TXNIP [57] , a lower anti-oxidative capacity caused by 2-DG exposure may also involve neutralization of the anti-oxidative capacity of thioredoxin by TXNIP induction . TXNIP and ARRDC4 are both transcriptionally regulated by the MondoA:Mlx complex to coordinate the fuel status of cellular metabolism and proliferation [57] , [87] . Our findings identify lactic acidosis as a novel stimulus for the activation of MondoA and induction of TXNIP/ARRDC4 . Even though we have identified the importance of MondoA-TXNIP of one component of the lactic acidosis response , there are still many aspects of lactic acidosis which remained unexplained , such as similarities to the glucose deprivation response . At least two factors are likely to be relevant for the shared gene expression response of glucose deprivation and lactic acidosis . First , these changes may lead to significant changes in the level of acetyl-CoA , which in turn impacts on histone acetylation of the chromatin of many target genes [65] . In addition , AMPK activation under both stresses may also impact on several transcription factors and co-activators to affect gene expression [88] , [89] . It is likely that the anti-tumor activities of AMPK activators ( e . g . , metformin or AICAR ) [90]–[93] may involve similar pathways triggered by lactic acidosis . Many studies on the lactate levels in human cancers have found that tumors with high lactate levels are associated with poorer clinical outcomes , tumor aggression and treatment failure [3]–[5] . How do our finding of the association of the lactic acidosis response with favorable outcomes reconcile with these studies ? There are at least four different ways how our results can be reconciled with the clinical observation . First , our studies mainly focus on the cellular response of cancer cells to short term exposure of lactic acidosis . This result may be different from the long term selection in the high lactic acidosis environments of human tumors . For example , glucose deprivation triggers starvation response with AMPK activation , mTOR inhibition and cell cycle arrest during short term exposure [48] . During the long term exposure to these stresses in human cancers , these undesirable conditions may select for tumor cells with the somatic mutations ( such as K-Ras mutation [94] ) that confers ability to adapt these stress conditions and strong metastasis potential . Similar selection pressure has also been suggested for the lactic acidosis [1] , [15] , [95] . In addition , the high level of lactate in the tumors is the downstream effects of the preferential use of glycolysis pathways due to tumor hypoxia , which may lead to more aggressive tumor behaviors and worse clinical outcomes . While lactic acidosis itself may exert some anti-tumor influences , this may not be enough to counter the influence of these somatic mutations and hypoxia in driving the tumor aggressiveness . It is also important to point out that the experimental evidence suggests that the lactic acidosis response is mainly due to acidosis instead of lactosis . Although we have found that lactosis may further augment the acidosis response , lactosis by itself have relatively little effects on the gene expression . Although the degree of the cellular response to these stresses ( such as CA9 ) can reflect the levels of stresses ( low tumor pO2 ) and serve as “endogenous markers” of such stresses in human cancers , such connection are not absolute and direct [96] . Therefore , the degree of lactic acidosis response in human cancers may or may not correlate directly with the tumor lactate levels . Through the dissection of individual stresses in vitro , we have shown here that lactic acidosis simultaneously triggers two anti-tumor pathways ( AMPK-mTOR and MondoA-TXNIP ) . Therefore , triggering such responses in cancer cells using small molecule compounds may have therapeutic potentials . It is of interest to investigate the mechanisms by which the lactic acidosis response is sensed and triggered in cancer cells . Extracellular lactate enters the cells through the lactate transporter proteins of monocarboxylate transporter ( MCT ) , which may be also important for the lactic acidosis response . Lactate has also been recognized for its role as an energy source [63] and a signaling molecule to affect tumor cell phenotypes and target of cancer therapeutics [97] . Previous studies have shown that lowering the extracellular pH from 7 . 4 to ∼6 . 7 will lead to a slight lowering of intracellular pH ( pHi ) from 7 . 4 to 6 . 9–7 . 0 [98] , [99] . Since many surface or cytosolic molecules exhibit high sensitivity to pH , this drop in pHe and the corresponding slight decrease in pHi may induce conformational changes to trigger signaling events [100] , [101] . For example , it has been postulated that inhibition of the acidosis-sensing glutamine pumps leads to amino acid depletion [73] , [74] and increases in the proteosome activity [102] in muscle cells , also contributing to muscle loss during metabolic acidosis . Low pHi has also been shown to directly enhance both DNA-binding and the TBP-interacting capacity of the transcriptional factor Sp1 [103] . Similarly , it is possible that this slight drop in pHi may destabilize the hydrogen bonds in several histone residues in the basic regions of MondoA to causes its transcriptional activation of target genes [57] . Homeostasis of pHi is mainly regulated by the sodium/hydrogen exchanger ( NHE ) family of proteins [104] , [105] and such regulation is dysregulated in cancer cells [106] . TXNIP is known to be transcriptionally regulated by a variety of stresses and stimuli [84] , [87] , [107] and it is possible that some of these stimuli may act through modulating pHi to affect MondoA . TXNIP has been proposed as an inhibitor of cell growth [107] and the inhibition of the growth-inhibiting TXNIP expression under alkaline pHi may help to explain its permissiveness for cellular proliferation in response to growth factors [108] , [109] . In addition , evidence is accumulating for the role of membrane acid-sensing receptors in either GPR4 family of G-protein coupled receptors ( GPCR ) and Acid-Sensing Ion Channels ( ASICs ) in many cell types . These acid-sensing receptors may also be critical in the observed response in both the AMPK-mTOR and MondoA-TXNIP under lactic acidosis . Importantly , these acid-sensing receptors can be modulated by small compounds pharmacologically [110] , [111] and thus have potential as cancer therapeutics [112] , [113] . For instance , several glycolipids are known natural ligands of these acid-sensing GPCRs and can activate or inhibit receptor responses to acidosis . Small compounds blocking the ASICs have been used to improve the symptoms and severity of stroke after vascular blockage [114] . Similarly , the use of compounds modulating the lactic acidosis response may mimic the tumor suppressor activities of lactic acidosis in novel cancer therapeutics . The definition of key genes and pathways will allow the use of genetic and chemical means to identify how to modulate cellular lactic acidosis for therapeutic purposes and thus potentially improve outcomes for cancer patients .
MCF7 breast cancer cell lines were cultured in DMEM ( GIBCO11995 ) with 4 . 5 g/L glucose , supplemented with 10% fetal bovine serum , 1× non-essential amino acid and 1× antibiotics ( penicillin , 10000UI/ml; streptomycin , 10000UI/ml ) . MEF cells were cultured in DMEM media supplemented with 15% FBS . Lactic acidosis conditions were created with the addition of 25mM or 10mM lactic acid ( Sigma ) to respective media and adjusted to the desired pHs with HCL or NaOH . Glucose deprivation was created by using 0 glucose/L media ( GIBCO11966 ) . Hypoxia was created by lowering the oxygen level to 1% oxygen . To stabilize the pH of DMEM media better , 25mM HEPES was also added into the media . RNAs from MCF7 cells exposed to control or lactic acidosis culture conditions for 1 , 4 , 12 , 24 hr were collected at respective time points and extracted with miRVana kits ( Ambion ) , followed by hybridization to Affymetrix Hu133 plus2 gene chips with standard protocols . RNAs from MCF7 cells exposed to lactic acidosis , glucose deprivation and hypoxia were extracted after 4 hour exposure with miRVana kits ( Ambion ) and hybridized to Affymetrix Hu133 plus2 gene chips in a similar fashion . CEL files were normalized by RMA using Expression console ( Affymetrix ) , filtered by indicated criteria , clustered with cluster 3 . 0 , and displayed with treeview . All the microarray results have been submitted into GEO with accession number GSE19123 . The RNA from the treated MEF cells were interrogated by hybridization to Affymetrix mouse 430A2 GeneChip with standard protocols and processed in a similar fashion . Gene expression signatures associated with 1 , 4 , 12 , and 24 hour lactic acidosis treatments were derived from RMA expression values using standard Bayesian sparse multivariate regression techniques , full details of which appear in previous publications [18] , [115]–[117] For each treatment group ( i . e . 1- , 4- , 12- , and 24-hr lactic acidosis treatment ) , this estimated the probability of differential expression for each Affymetrix probe as well as the fold change of the differentially expressed transcripts , as compared to the control . These values define the in-vitro signature of the treatment . Projection of these signatures into the primary tumor data sets is derived from the weighted inner product of the vector of tumor gene expression values and the vector of regression coefficients associated with the signature , as described in [117]: the score associated with signature k in sample i is defined as , where is the estimated probability of differential expression of probe g in treatment k , is the estimated fold change of the transcript , is the estimated residual variance of probe g , and is the RMA normalized expression value of probe g in sample i . The signature score thus provides a relative measure of the extent to which the pattern of expression described by the signature is present or reversed in a given vector of expression values . P-values associated with the significance of Kaplan Meier curves were derived from survival analyses conducted using a Cox proportional hazards model in MATLAB . Kaplan Meier curves show the optimal stratification of high-risk and low-risk groups as determined by choosing the partition threshold that maximizes the area between curves . Measures of correlation and the associated P-values for the signature score scatter plots were calculated using the Pearson coefficient . Full details of statistical analysis , including data and computer code to replicate the analyses , are available as Dataset S1 . RNAs were reverse-transcribed to cDNAs with SuperScript II reverse transcription kit following the manufacturer's protocol ( Invitrogen ) . cDNAs were then used as the substrate for gene expression level measured by qPCR with Power SYBRGreen PCR Mix ( Applied Biosystem ) and primers specific for TXNIP ( Forward: CTGGCGTAAGCTTTTCAAGG , Reverse: AGTGCACAAAGGGGAAACAC ) , ARRDC4 ( Forward: CCCCCTCCCACATGGTCACA , Reverse: TCCCTGGCTCCCTTCCATGTGT ) , Actin-beta ( Forward: CTCTTCCAGCCTTCCTTCCT , Reverse: AGCACTGTGTTGGCGTACAG ) following the manufacturer's protocol . MCF7 cells were plated in 6-well/12-well plates at the density of 800 , 000/200 , 000cells per well . Once cells were more than 75% confluent , they were washed with 1× PBS twice , followed by application of serum starvation media ( 0 . 1%FBS ) for three hours . Cells were then treated with respective conditions for the desired time . For MEF cells , they were plated in 12-well plates at the density of 100 , 000 cells per well . Once they reached more than 70% confluence , applying respective indicated conditions for four hours . For glucose uptake measurement , cells would then be washed with 37°C KRH buffer twice , followed by adding in 500ul/200ul KRH buffer containing 0 . 5uCi/0 . 2uCi 2-deoxy-D-glucose ( GE Healthcare ) for one hour in 37°C incubator . 20uM cytochalasin B was added for negative controls . After incubation , cells would be washed three times with 1ml/400ul of ice-cold KRH buffer containing 20mM glucose and 0 . 5mM phloretin to quench the glucose uptake . Finally , cells were lysed with 1ml/400ul RIPA buffer and the lysates were subjected to liquid scintillation counting . Protein concentrations were measured with Bradford assay . To measure the glucose uptake of genetically-manipulated cells , glucose uptake was measured 24 or 48 hours after the transfections . MCF7 cells were plated in 12-well plates for the density of 200 , 000 cells per well . Once the cells reached 60% confluence , 100nM siRNAs were transfected by using lipofectamine . To verify the successful knocking down of the intended transcripts , RNAs were collected with miRVana kit 24 hours after the transfection , whereas 48 hours after the transfection , proteins were collected by lysing the cells with RIPA buffer . Proteins were collected with RIPA buffer and their concentrations were measured with Bradford assay . Equal amounts of proteins were loaded for the protein analyses . Primary antibodies of AMPK , S6K ( cell signaling ) , TXNIP ( MBL ) and anti-MondoA antibodies were applied following the manufacturers' protocols or as described previously [57] . Wild type mouse embryo fibroblast ( MEF ) and TXNIP null MEF cells were plated at the density of 25 , 000 cells per ml . The multi-channel pipette was used to plate 100ul of cell suspension evenly into the 96-well cell culture plate . Respective conditions of 10mM or 25mM lactic acidosis were applied the next day and cell number was estimation by using a standard MTS ( 3- ( 4 , 5-dimethylthiazol-2-yl ) -5- ( 3-carboxymethoxyphenyl ) -2- ( 4-sulphophenyl ) -2H-tetrazolium ) colorimetric assay ( Promega ) 48 hours after respective treatments . ChIP studies were performed as described previously [57] using an off-target region located on chromosome 10 to calculate fold enrichment . Primer sequences for the promoters of TXNIP , ARRDC4 , and the off target control of the CHIP analysis are available upon request to D . E . A . MCF cells were plated in six-well plates at a density of 500 , 000 cells per well . The next day , non-targeting siRNAs and siRNAs against TXNIP were transfected as described previously . After 24 hours , fresh media for the respective conditions , including control and 25mM lactic acidosis were applied to cells for 48 hrs when media were collected for glucose ( ACCU-CHECK Aviva , Roche ) and lactate ( ARKRAY ) measurements with respective meters . The results were normalized by cell numbers to obtain the glucose consumption and the lactate production amount per million cells . | Solid tumors usually have many differences in their chemical environments , such as low oxygen , depletion of glucose , high acidity ( low pH ) , and accumulation of lactate , from normal tissues . These changes are usually called tumor microenvironmental stresses . In this study , we have used microarrays to compare the transcriptional response and metabolic adaptation in response to these different stresses seen in the tumor microenvironments . Through these comparisons , we have found that lactic acidosis triggers a starvation response , highly similar to glucose deprivation , even in the presence of abundant nutrients and oxygen . Even the cells seem to be starved; cells under lactic acidosis have decreased glucose uptake . We found this unexpected biological behavior was due to the paradoxical induction of a glucose-sensing Mondo-TXNIP pathway . The activation of this novel anti-tumor pathway under lactic acidosis contributes to the anti-Warburg effect and the restriction of cell growth in tumorigenesis by limiting nutrient availability and its inactivation may be required for tumor progression under these microenvironmental stresses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"genetics",
"and",
"genomics/cancer",
"genetics",
"pathology/molecular",
"pathology"
]
| 2010 | Lactic Acidosis Triggers Starvation Response with Paradoxical Induction of TXNIP through MondoA |
Polarized membrane trafficking is essential for the construction and maintenance of multiple plasma membrane domains of cells . Highly polarized Drosophila photoreceptors are an excellent model for studying polarized transport . A single cross-section of Drosophila retina contains many photoreceptors with 3 clearly differentiated plasma membrane domains: a rhabdomere , stalk , and basolateral membrane . Genome-wide high-throughput ethyl methanesulfonate screening followed by precise immunohistochemical analysis identified a mutant with a rare phenotype characterized by a loss of 2 apical transport pathways with normal basolateral transport . Rapid gene identification using whole-genome resequencing and single nucleotide polymorphism mapping identified a nonsense mutation of Rab6 responsible for the apical-specific transport deficiency . Detailed analysis of the trafficking of a major rhabdomere protein Rh1 using blue light-induced chromophore supply identified Rab6 as essential for Rh1 to exit the Golgi units . Rab6 is mostly distributed from the trans-Golgi network to a Golgi-associated Rab11-positive compartment that likely recycles endosomes or transport vesicles going to recycling endosomes . Furthermore , the Rab6 effector , Rich , is required for Rab6 recruitment in the trans-Golgi network . Moreover , a Rich null mutation phenocopies the Rab6 null mutant , indicating that Rich functions as a guanine nucleotide exchange factor for Rab6 . The results collectively indicate that Rab6 and Rich are essential for the trans-Golgi network–recycling endosome transport of cargoes destined for 2 apical domains . However , basolateral cargos are sorted and exported from the trans-Golgi network in a Rab6-independent manner .
Many cells that make up animal bodies have multiple plasma membrane domains , which are the basis of their specific functions . Polarized vesicle transport is essential for establishing and maintaining these domains . However , the underlying mechanisms are not well elucidated . The Drosophila retina is a good genetics-based model system for studying polarized transport . A single retina contains approximately 800 ommatidia , which comprise 6 outer and 2 central photoreceptors with 4 distinct plasma membrane domains ( Fig 1A and 1B , see also [1] ) . The first domain is the photoreceptive membrane domain , i . e . , the rhabdomere , which is formed at the center of the apical plasma membrane as a column of closely packed rhodopsin-rich photosensitive microvilli . The second is the peripheral apical domain surrounding the rhabdomere , i . e . , the stalk membrane , which is enriched in Crb and beta H-spectrin [2 , 3] . “Eye shut” ( Eys ) is an extracellular matrix protein secreted from the stalk domain of photoreceptors from the early to mid-pupal stages [4 , 5]; its secretion is necessary and sufficient to form the inter-rhabdomeric space ( IRS ) . The third domain is the basolateral membrane , which is separated from the apical membrane by adherens junctions , where Na+K+ATPase localizes , similar to typical polarized epithelial cells [6] . Finally , the fourth domain comprises the axon and synapses , which extend below the retina to the brain . Rh1 , the rhodopsin expressed in R1–6 outer retinal photoreceptor cells , is the most accessible protein for investigating apical polarized transport in Drosophila photoreceptors . Blue light-induced chromophore supply ( BLICS ) releases Rh1 from the endoplasmic reticulum ( ER ) [7] , allowing us to trace synchronous Rh1 trafficking . We previously used this method to show that Rab1 is involved in ER–Golgi transport [8] , that GPI synthesis is essential for Rh1 sorting in the trans-Golgi network ( TGN ) [1] , and that the Rab11/dRip11/MyoV complex is essential for post-Golgi Rh1 transport [9 , 10] . In addition , other studies demonstrate that the exocyst complex tethers post-Golgi vesicles to the base of the rhabdomeres [11] and that Rab6Q71L overexpression inhibits Rh1 transport in the early secretory pathway [12] . Rab6 is a member of the Rab family of small GTPases , which regulate the specificity between donor and acceptor membranes for vesicle budding , docking , tethering , and fusion steps during transport [13–15] . Human Rab6 comprises 4 different isoforms—Rab6a , Rab6a′ , Rab6b , and Rab6c—and is the most abundant Golgi-associated Rab protein [16] . Rab6 regulates retrograde transport from the Golgi complex to the ER [17 , 18]; the retrograde transport of the B subunit of Shiga toxin from early endosomes or recycling endosomes ( REs ) to the Golgi complex [19 , 20]; and the post-Golgi trafficking of herpes simplex virus 1 ( HSV1 ) envelop proteins , tumor necrosis factor ( TNF ) , and vesicular stomatitis virus G-protein ( VSV-G ) to the plasma membrane [21–23] . In the present study , we screened ethyl methanesulfonate ( EMS ) -mutagenized flies to find mutants with defective polarized transport . As one of the mutants identified was a null allele of Rab6 , we reevaluated the function of Rab6 in Drosophila photoreceptors and found Rab6 is essential for transport towards both distinct apical domains ( i . e . , the stalk and rhabdomeres ) but not towards the basolateral membrane . These results indicate basolateral cargos are sorted and exported from the TGN in a Rab6-independent manner before they are sorted into pathways destined for the stalk and rhabdomeres .
To identify the genes essential for rhabdomere morphogenesis and membrane trafficking , we previously performed retinal mosaic screening of P-element–inserted lines maintained in stock centers [24] using the FLP/FRT method [25] with in vivo fluorescent imaging [26] of Arrestin2::GFP [1] . To extend the screening to comprehensive coverage , we are currently engaged in genome-wide , large-scale screening of EMS-mutagenized chromosomes; thus far , we have isolated 233 Rh1-accumulation deficient mutant lines , which stochastically covers more than 80% of genes in more than 60% of Drosophila genome [27] . Flies including Drosophila melanogaster have an open rhabdom system , in which the rhabdomeres of each ommatidium are separated from each other and function as independent light guides . In contrast , bees and mosquitoes have a closed system , in which rhabdomeres within each ommatidium are fused to each other , thus sharing the same visual axis . Without expression or secretion to the IRS of the extracellular matrix glycoprotein Eys , fly ommatidia exhibit a closed rhabdom [4 , 5] . By screening EMS-mutagenized left arms of the second chromosome , we isolated 11 mutant lines with a closed rhabdom phenotype similar to the eys mutant ( Fig 1C and 1D ) ; these lines were further classified into 4 complementation groups . The mutations in one group failed to complement a loss-of-function allele of the eys gene , eysBG02208 [4] , indicating that mutations in this group represent new loss-of-function alleles of eys . On the other hand , the mutations in the other 3 groups complemented the eysBG02208 mutant . Fig 1C and 1D show Arrestin2-GFP localizations in 510U mutant in the eys complementation group as well as 546P in another complementation group; 546P was the sole member of the complementation group , and the homozygotes were lethal in the early larval stage . The closed rhabdom phenotypes of 510U and 546P had 100% penetrance: all adult flies observed for 510U or 546P using Arr2GFP had mutant ommatidia with a fused rhabdomere . We subsequently observed similar high penetrance of the phenotypes in immunostaining at the fly and cell levels unless otherwise noted . Immunostaining of the mosaic retina of these mutants with anti-Eys antibody revealed complete loss of Eys protein in the whole mutant ommatidia of 510U ( Fig 1E ) , which again strongly suggests the complementation group ( which contains 510U ) represents eys . On the other hand , in 546P mutant ommatidia , Eys was readily detected; however , it did not accumulate in the extracellular domain ( i . e . , the IRS ) but rather within undefined cytoplasmic structures ( Fig 1F ) . These results suggest the transport of Eys protein is inhibited in the 546P mutant . We subsequently investigated the localization of several plasma membrane proteins in 546P mutant photoreceptors ( Fig 2 ) . In wild type photoreceptor cells , Rh1 localized at the photoreceptive rhabdomeres . However , Rh1 was mainly detected in the undefined cytoplasmic structures in 546P mutant photoreceptors ( Fig 2A , blue ) . In contrast to Eys and Rh1 , a basolateral membrane protein , Na+K+ATPase localized normally in the basolateral membrane and did not accumulate in the cytoplasm of 546P mutant photoreceptors ( Fig 2A , green ) . Eys and Rh1 were co-localized in the same undefined cytoplasmic structures ( Fig 2B ) . Two other rhabdomeric proteins—TRP and Chaoptin ( Chp ) —and a stalk protein—crumbs ( Crb ) —were also detected in the undefined cytoplasmic structures ( Fig 2C–2E ) . On the other hand , DE-Cadherin ( DE-Cad ) was localized normally at the adherens junctions ( Fig 2E ) . These results indicate inhibition of transport toward the rhabdomeres and stalks but not toward the basolateral membrane . Coinciding with the localizations of plasma membrane proteins , the basolateral membrane in 546P mutant photoreceptors was extended as in the wild type despite the shrunken stalk membrane and small rhabdomeres ( Figs 2A , 2D and 2E and S5 ) . None of the 222 other Rh1 accumulation-deficient lines clearly showed 546P-like phenotypes , closed rhabdomeres , or inhibition of transport toward the rhabdomere and stalk ( but not toward the basolateral membrane ) , indicating these 546P phenotypes are characterized by an apical-specific transport deficiency rather than the result of a delay or loss of general membrane transport . To clarify the details of the transport defects in 546P mutant photoreceptors , we investigated the dynamics of Rh1 transport using BLICS [1 , 8] . Rh1 comprises an apoprotein , opsin , and the chromophore 11-cis retinal . Without the chromophore , opsin accumulates in the ER . Blue light illumination photoisomerizes all-trans retinal to 11-cis , inducing the synchronous release of Rh1 from the ER into the secretory pathway . Prior to BLICS , Rh1-apoprotein accumulated normally in the ER and Rh1 was transported to the Golgi units by 40 min after BLICS in 546P mutant photoreceptors ( Fig 3A and 3B ) . However , in contrast to wild type photoreceptors , Rh1 did not reach the rhabdomeres in 546P mutant photoreceptors but instead localized in large globular structures in the cytoplasm with the late endosome marker Rab7 180 min after BLICS ( Fig 3C and 3H ) . These results indicate Rh1 is normally synthesized in the ER and transported to the Golgi units in 546P mutant photoreceptors but is subsequently transported to the late endosomes rather than the rhabdomeres . To determine how Rh1 is exported from the Golgi units to the late endosomes , we compared Rh1 staining in the Golgi units 30 and 60 min after BLICS in both wild type and 546P mutant photoreceptors ( Fig 3D and 3E ) . There were no apparent differences between wild type and 546P mutant photoreceptors 30 min after BLICS . However , 60 min after BLICS , Rh1 staining became stronger in the Golgi units of mutant photoreceptors than wild type photoreceptors . Moreover , the Golgi units extended Rh1-positive tubular or globular structures ( Fig 3E ) . Next , we triple stained 546P mutant photoreceptors with antibodies against the medial-Golgi marker p120 [28] , late endosomal marker Rab7 [29] , and Rh1 40 and 90 min after BLICS and investigated the Rh1-positive tubular or globular structures extending from the Golgi units ( Figs 3F and S1 ) . Golgi markers p120 , Rab7 , and Rh1 were frequently partially co-localized . Rab7 staining was often adjacent to p120 staining , and Rh1 staining overlapped with both of them ( Fig 3F , arrows; S1 Fig ) . A Z-projection ( Fig 3G ) of 20 optical sections at 0 . 5-μm intervals ( S1 Fig ) showed most Golgi units faced the globular Rab7-positive structure , both of which were positive for Rh1; the Golgi units with Rh1 ( turquoise ) were often adjacent/connected to late endosomes containing Rh1 ( yellow ) . These observations suggest Rh1 is first transported to the Golgi units in a normal manner but is not exported normally and subsequently accumulates there; finally , Rh1 directly exits the Golgi units to Rab7-positive tubular or globular structures . We previously showed that the Rab11/dRip11/MyoV complex is essential for post-Golgi vesicle transport and that deficiency of any component of the complex induces cytoplasmic accumulation of Rh1-loaded post-Golgi vesicles [9 , 10] . The pattern of Rh1-accumulation observed in the 546P mutant clearly differed from those of Rab11 , dRip11 , or MyoV mutants . To investigate the epistatic interaction between mutations of the Rab11/dRip11/MyoV complex and the 546P mutation for Rh1 transport , we observed Rh1 localization in 546P mutant mosaic retina expressing the dominant-negative MyoV C-terminal domain . Rh1 did not accumulate in the cytoplasm of 546P/MyoV double-mutant photoreceptors like the MyoV single mutant but instead accumulated in the globular structures ( Fig 3H ) . This phenotype was indistinguishable from that of cells with only the 546P mutation ( Fig 2A–2C ) . This result indicates the 546P mutation is epistatic to mutations of the Rab11/dRip11/MyoV complex . Thus , in the 546P mutant , the Rh1 in Rab7-positive late endosomes directly comes from the Golgi units before being sorted into post-Golgi vesicles . Kinetic and epistatic analyses of Rh1 transport in 546P mutant cells revealed that the processes between Golgi entry and before/upon post-Golgi vesicle formation are inhibited in the 546P mutant during Rh1 biosynthetic trafficking . As the phenotype of 546P is drastic and important for understanding the mechanism of polarized vesicle transport , we identified the mutation responsible for the phenotype . Rough mapping using meiotic recombination and restriction fragment length polymorphisms ( RFLP ) [30] placed the 546P mutation between RFLP893 and RFLP977 . We subsequently sequenced the whole genome of 546P homozygous animals using whole-genome amplification and a next-generation sequencer [27] and found 4 unique mutations in the mapped area; complementation tests over defined chromosomal deletions showed that only one of these , a nonsense mutation on the Rab6 gene ( Rab6 R73 to a stop codon ) , is lethal like 546P homozygous larvae . As a Rab6 null mutation was found in the mapped area , we subsequently determined if YFP::Rab6wt expression rescues the 546P phenotype . Using the UAS/Gal4 system , we expressed YFP::Rab6wt from early/late pupa in whole bodies using heat shock-Gal4 or in late pupal outer photoreceptors using Rh1-Gal4 ( Fig 4 ) . YFP::Rab6wt expression from 28% pupal development by heat shock-Gal4 completely rescued the 546P phenotype ( Fig 4F and 4G ) , confirming that the Rab6 gene is responsible for the 546P phenotype . Therefore , we designated 546P as a new allele of the Rab6 gene , Rab6546P . Interestingly , YFP::Rab6wt expression starting from 68% pupal development by heat shock-Gal4 or by Rh1-Gal4 showed partial rescue; Rh1 accumulated in fused rhabdomeres lacking normal level of Eys in the IRS ( Fig 4D , 4E , 4I and 4J ) . These results are concordant with previous reports on Eys/spacemaker [5] showing that Eys expression in early pupae is required to form open rhabdomeres . We subsequently investigated whether photoreceptors homozygous for the other Rab6 null allele Rab6D23D have defects in Rh1 and Eys trafficking . Like Rab6546P mutant photoreceptors , Rab6D23D ommatidia exhibited a closed rhabdomere phenotype visualized by phalloidin staining ( S2A Fig ) . Eys accumulated in the globular cytoplasmic structures , although a limited amount localized between photoreceptor apical membranes ( S2A Fig ) . In Rab6D23D photoreceptors , most Rh1 was localized in the cytoplasm , whereas the basolateral membrane protein Na+K+ATPase localized normally in the basolateral membrane ( S2B Fig ) . These results corroborate the notion that the 546P phenotype is due to Rab6 protein deficiency . In yeast and mammalian cells , the Ric1/Rgp1 complex functions as a guanine nucleotide exchange factor ( GEF ) for Rab6 [31 , 32] . The Drosophila Ric1p homolog Rich is a Rab6 effector in Drosophila [33] . Rich contains RIC1 and WD40 domains , both of which bind to fly Rab6 . Although neither the GEF activity for Rab6 nor binding to the Drosophila rgp1 homolog CG1116 has been confirmed , Rich might function as a GEF together with other interacting proteins [33] . Rab6 and Rich interact genetically and are required for synaptic specificity in fly eyes and olfactory receptor neurons [33] . Therefore , we investigated the localizations of Rh1 and Eys in photoreceptors homozygous for a null allele in Rich , Rich1 . Similar to Rab6546P and Rab6D23D photoreceptors , Rh1 and Eys accumulation in the rhabdomeres and IRS , respectively , were limited , and both proteins were detected in cytoplasmic structures ( S2C and S2D Fig ) . On the other hand , the Na+K+ATPase localized normally in the basolateral membrane in Rich1 mutant photoreceptors ( S2D Fig ) . The phenotypes were highly penetrant in Rab6D23D photoreceptors . However , in the case of Rich1 homozygous photoreceptors , some Rich1 mutant cells only exhibited a weak phenotype . Detailed observation of Rh1 transport by BLICS in Rab6D23D and Rich1 mutants showed that Rh1 is synthesized and transported to the Golgi units in a normal manner; however , after leaving the Golgi units , most Rh1 enters the late endosomes instead of the rhabdomeres in the same manner as in Rab6546P photoreceptors ( S3 Fig ) . The localizations of other membrane proteins were investigated in Rab6D23D and Rich1 mutants ( S4 Fig ) . The rhabdomeric proteins Chp and TRP accumulated in the cytoplasm in both Rab6D23D and Rich1 mutants . The localization of the stalk protein Crb in the stalk membrane was limited in both Rab6D23D and Rich1 mutants , but its cytoplasmic accumulation was observed only in Rab6D23D mutants . More severe membrane trafficking defects were observed in Rab6D23D mutants than Rich1 mutants . On the other hand , DE-Cad localized normally in the basolateral membrane in both Rab6D23D and Rich1 mutants . These observations coincide with the Rab6546P phenotype . Quantification of the lengths of the stalk and basolateral membrane in tangential sections revealed that Rab6546P and Rab6D23D mutant photoreceptors had shorter stalk membranes than those of wild type but basolateral membranes of normal length ( S5A–S5F Fig ) . Furthermore , we determined if Rab6 deficiency affects the transport of another basolateral protein , FasIII , using ovarian follicle cells ( S6 Fig ) . FasIII localized normally on the basolateral membrane , whereas the localization of the apical membrane protein , Notch , was diminished in follicle cells . Thus , the involvement of Rab6 in apical transport but not basolateral transport could be common in polarized cells . To investigate the details of organelle structures and plasma membrane domains , we observed thin sections of late pupal wild type , Rab6546P , Rab6D23D , and Rich1 whole mutant ommatidia by electron microscopy ( Fig 5A–5D and S7 Fig ) . A tangential section of a wild type ommatidium contains 7 round rhabdomeres separated by the IRS . However , Rab6546P , Rich1 , and Rab6D23D ommatidia contained miniature rhabdomeres attached to each other , and a narrow IRS ( Fig 5 , purple ) did not separate adjacent rhabdomere , similar to Eys395 and Eys1 mutants [4 , 5] . The stalk membranes in Rab6546P , Rab6D23D and Rich1 photoreceptors were shorter than those of the wild type , whereas the basolateral membranes were normal ( S5G–S5I Fig ) . Similar to the confocal microscopic observations , these phenotypes were less prominent in Rich1 ommatidia than Rab6546P and Rab6D23D ommatidia . Overall , these plasma membrane morphological phenotypes are consistent with the immunohistochemically detected transport phenotypes of Rab6 and Rich mutants . The appearance of the Golgi units in Rab6546P , Rab6D23D , and Rich1 photoreceptors also differed from those in the wild type: both Rab6- and Rich-deficient Golgi units were larger and well developed , and their cisternae were dilated ( Fig 5E–5H ) , similar to previous reports [21–23] . However , in contrast to the report of Storrie et al . , ( 21 ) the vesicle or budding profiles of Rab6 and Rich-deficient Golgi units were not obviously increased . In addition to the differences in these plasma membrane domains , the IRS , and Golgi units , the number and size of multi-vesicular bodies ( MVBs ) were greater in Rab6546P , Rab6D23D , and Rich1 photoreceptors than the wild type ( Fig 5A–5D , blue ) . We quantified the MVBs in dark-adapted homozygous wild type , Rab6546P , and Rich1 photoreceptors , because MVB formation is strongly triggered by light-dependent Rh1 endocytosis even in wild type photoreceptors ( Fig 5I ) . MVBs 300–600 and ≥600 nm in diameter were categorized as small and large , respectively . Adult wild type photoreceptors contained ( mean ± SD ) 0 . 15 ± 0 . 06 small MVBs and zero large MVBs . On the other hand , single Rab6546P photoreceptors contained 0 . 33 ± 0 . 11 and 0 . 27 ± 0 . 09 small and large MVBs , respectively . Similarly , Rich1 photoreceptors contained 0 . 35 ± 0 . 09 and 0 . 27 ± 0 . 06 small and large MVBs , respectively . In addition to the increased number and size of MVBs , MVBs in Rab6- and Rich-deficient photoreceptors appeared different from those in the wild type photoreceptors; Rab6546P , Rab6D23D and Rich1 photoreceptors contained many MVBs densely packed with inner vesicles that appeared partially degraded and were often accompanied by cisternae , some of which were Golgi units ( Fig 5E–5H ) . The consistent sizes and populations of MVBs suggest that MVBs are likely identical to the Rab7-positive undefined cytoplasmic structures that accumulate Rh1 , Eys , TRP and Chp in the 546P mutant ( Figs 2 and 3 ) . Further understanding the function of Rab6 requires knowing the localization of Rab6 protein in fly photoreceptors . In our first attempt , we used UAS-YFP::Rab6wt transgenic flies [34] . YFP::Rab6wt expressed by Rh1-Gal4 forms cytoplasmic foci in R1–6 photoreceptors . Most of the YFP::Rab6wt foci were associated with the cis-Golgi marker GM130 ( Fig 6A , arrows ) , indicating these foci are located in Golgi units . GM130-negative YFP::Rab6wt foci at the base of the rhabdomeres were presumably post-Golgi vesicles bearing Rh1 ( Fig 6A , arrowheads ) . Dispersed YFP signals were also observed in the cytoplasm and the rhabdomeres . The cytoplasmic YFP signal likely reflects GDP-bound form of YFP::Rab6wt associating with GDI . However , the reason for the YFP signal in the rhabdomeres is unclear; Rab6 might associate with some rhabdomeric proteins . To confirm the localization of endogenous Rab6 , we generated antisera against histidine-tagged full-length Rab6 protein in 1 rabbit and 2 guinea pigs ( Rb anti-Rab6 , GP1 anti-Rab6 and GP2 anti-Rab6 ) . Western blotting showed that all antisera recognized a band around 22 kD in wild type head extract as well as a band around 60 kD in genomic Rab6EYFP heterozygous head extracts ( S8A Fig ) . Immunohistochemistry showed cytoplasmic foci only in wild type photoreceptors in Rab6D23D mosaic retinas ( Fig 6B , arrows and S8B Fig ) . Hence , these antisera specifically recognize Rab6 . Rab6-positive cytoplasmic foci are also GM130-positive , suggesting these foci are Golgi units . Similar to YFP::Rab6wt , there were Rab6 foci at the bases of the rhabdomeres , which were GM130-negative , suggesting Rab6 also localizes on post-Golgi vesicles ( Fig 6B , arrowheads ) . In Rich1 homozygous photoreceptors ( Fig 6C , arrows ) , cytoplasmic Rab6-positive foci were not observed , suggesting that Rich is required for Rab6 recruitment to the Golgi membrane . Photoreceptors at earlier stages ( i . e . , ~20–30% pupal development ) contain well-developed Golgi units , often reaching >1 μm in width [10] . To investigate the Golgi–cisternal association of Rab6 , we used these large Golgi units in young pupal photoreceptors expressing a CFP-GalT marker , i . e . , a cyan fluorescent protein fused to the 81 N-terminal amino acids of human β-1 , 4-galactosyltransferase [10] , which localizes at the trans-cisternae of the Golgi units and the TGN [35 , 36] . Double staining of CFP-GalT–expressing retina with anti-Rab6 and anti-GM130 showed that Rab6 localized to the opposite side of CFP-GalT–positive compartment from the cis-Golgi marker GM130 ( Fig 7A ) . Detailed observations indicate the Rab6-positive region includes the TGN and extends more distally ( Fig 7C ) . Most of the Golgi units were accompanied by Rab11-positive puncta adjacent to the trans side ( S9 Fig ) . Unlike the late pupal stages , when Rh1 transport is active and many Rab11-positive puncta localize at the base of the rhabdomeres , there are only a few Rab11-positive puncta apart from Golgi units in the young pupal stages . As Rab11 is typically considered as RE marker and is close to the TGN , these results suggest that these Rab11-positive puncta are likely fly photoreceptor REs emerging from the TGN or transport vesicles going to REs . Double staining of a CFP-GalT–expressing retina with anti-Rab6 and anti-Rab11 showed that Rab6 spreads from the CFP-GalT–positive TGN to the Golgi-associated Rab11-positive recycling endosomes ( Figs 7B and S9 [wide view] ) . These results suggest that Rab6 regulates transport between the TGN and RE . The Clathrin heavy chain ( Chc ) is an essential coat protein of some post-Golgi vesicles , especially for the basolateral pathway in epithelial cells [37] . Therefore , we compared Rab6 localization with Chc by triple staining wild type retinas with anti-Rab6 , anti-Rab11 , and anti-Chc ( Fig 7C ) [38] . Interestingly , Rab6-signals overlapped with both Rab11 and Chc , whereas Chc and Rab11 are largely separate ( Fig 7C , upper panel ) . These results suggest that Rab6 functions in a step close to but before Rab11-dependent post-Golgi trafficking .
In the present study , screening for the failure of Rh1 accumulation in Drosophila photoreceptor rhabdomeres led to the re-identification of Rab6 as an essential molecule for Rh1 transport . Our detailed observations of Rh1 trafficking by BLICS indicate the inhibition of transport in Rab6-null mutants occurs within the Golgi units: Rh1 is transported to the cis-Golgi with normal kinetics but does not exit the Golgi units into the plasma membrane . Consequently , Rh1 greatly accumulates on the Golgi membrane and finally exits directly into the endosomal compartment . Furthermore , in Rab6 mutant photoreceptors , other rhabdomere membrane proteins ( TRP and Chp ) , a stalk membrane protein ( Crb ) , and the apically secreted protein ( Eys ) are reduced and degraded in MVBs together with Rh1 . In contrast , the basolateral membrane proteins Na+K+ATPase and DE-Cad are unaffected . Moreover , Rab6 proteins are distributed from the TGN to Golgi-associated Rab11-positive structures . Many aspects of Rab6 functions have been reported , such as retrograde transport from endosomes to the TGN , intra-Golgi or Golgi to ER trafficking , anterograde transport from the TGN , Golgi homeostasis , and Golgi-ribbon organization [39] . Nevertheless , there is no concrete understanding of all Rab6 functions . Among them , the defective morphology of the Golgi units and the failure of Rh1 transport observed in the present study are consistent with the results of Rab6 knockdown in HeLa cells by Storrie et al . [21] , who report amplification and dilation of Golgi stacks , accumulation of unreleased vesicles on the trans-Golgi , inhibition of Golgi–plasma membrane transport of VSV-G without delayed ER–Golgi transport , and increased MVBs around the Golgi . The present results are also concordant with a study by Januschke et al . , who report that mutants of Rab6b and a Rab6 GEF BicD in Drosophila oocytes induce accumulation of the TGF-α homolog Gurken in “ring-like particles” that have the properties of endolysosomes [40] . Two recent reports on HSV1 production and TNF secretion indicate Rab6 depletion inhibits the post-Golgi transport of HSV1 envelope proteins and TNF , respectively [22 , 23]; these proteins accumulate at a juxtanuclear Golgi-like location and Golgi membranes in Rab6-depleted cells , respectively . These results are concordant with Rh1 localization after BLICS in our studies . Thus , together with the Rab6 localization at the trans-Golgi , the present results indicate that Rab6 is crucial for apical cargo export from the TGN . The constitutively active GTP-locked form of Rab6 overexpressed by Rh1-promoter drastically reduces Rh1 content and retains Rh1 in the ER form in Drosophila photoreceptors [12] . Although the observed phenotype appears specific to rhodopsin and deficiency of ER–Golgi transport rather than post-Golgi transport , the phenotype is different from that in the present study likely because of differences in the methods used: Eys , TRP , and Crb expressions peak during the mid-pupal stage , whereas Rh1-promoter begins to be expressed at the late pupal stage . Gain or loss of Rab6 function causes defects in protein transport via the Golgi; however , they have opposite effects on the structure of Golgi units . Constitutively active Rab6 forces Golgi proteins to relocate to the ER [41] in contrast to the accumulation of the Golgi cisternae in Rab6-knockdown conditions [21] . Therefore , Rh1 may remain in the ER in the Rab6 hypermorph but stall at the TGN in the Rab6 amorph . The results of the present study show that the Rab6-binding protein Rich is involved in rhabdomere and stalk membrane transport . Although its biochemical activity as a Rab6 GEF is unconfirmed [33] , the loss of Rab6 localization to the Golgi in Rich1 mutant indicates Rich is required to recruit Rab6 to the Golgi membrane and likely works as a Rab6GEF like yeast and mammalian Ric1 [31 , 32] . Rich was originally identified as a regulator of DN-Cad trafficking to synapses; Rab6 is also involved in these processes [33] . The present and previous findings collectively suggest Rab6/Rich-dependent export from the TGN is required for the trafficking of some axonal proteins as well as rhabdomeres and stalk proteins . The most important finding of the present study is that in Rab6 mutants , the transport of protein cargos destined for the 2 apical membranes ( i . e . , the rhabdomere and stalk ) is arrested and they accumulate together in the MVBs , while the basolateral cargos are transported normally . Hence , Golgi-localized Rab6 is not required for all plasma membrane-directed transport pathways . These results have implications for previous work on the Drosophila germline cyst [42] suggesting the existence of Rab6-dependent and Rab6-independent exocytic pathways . Although the roles of these multiple exocytic pathways in the polarized membrane domains are not mentioned , the Rab6-independent pathway to the plasma membrane in the germline cyst might correspond to the basolateral transport in photoreceptors . In addition , in embryonic salivary glands , Rab6 and Rab11 are localized on cytoplasmic vesicles containing an overexpressed apical membrane protein Cadherin 99C ( Cad99C ) [43] . Moreover , there is phenotypic similarity between Rab6 null and sec5 null . These findings indicate Rab6 and Rab11 sequentially regulate the apical transport of Cad99C , similar to Rh1 transport . However , in this case , as basolaterally transported truncated Cad99D also associates with Rab6 , Rab6 and Cad99C colocalization merely reflects the association of Cad99C with Golgi units rather than apical transport vesicles . Clathrin and AP1B are involved in basolateral transport [44 , 45] and are required to sort various apical proteins in both mice and C . elegans [46–49] . In fly photoreceptors , the sole fly AP1 is essential for Na+K+ATPase transport towards the basolateral membrane , and the loss of the AP1 subunit , gamma , or mu causes mistransport of Na+K+ATPase to the stalk membrane [1] , similar to zebrafish hair cells lacking AP1β subunits [50] . Clathrin likely plays a role in Na+K+ATPase transport towards the basolateral membrane in fly photoreceptors . The partial colocalization of Chc and Rab6 as well as the separation of Chc and Rab11 at the distal TGN ( Fig 7B and 7D ) indicate that the basolateral transport pathway likely branches off the Rab6-dependent apical transport pathway at the TGN . We previously showed that Rh1 transport from the TGN to the rhabdomere is dependent on Rab11 [10] . In addition , Rab6 localization is not restricted to the TGN but extends further to the Rab11-positive puncta closely associated with the TGN . These results imply that Rab6 is involved in Rh1 transport from the TGN to the closely associated Rab11-positive puncta . As Rab11 is commonly considered as RE marker , these TGN-associated Rab11-positive puncta are likely REs emerging from the TGN or transport vesicles going to REs . These results collectively suggest Rab6 regulates the trafficking of apical cargos from the TGN to REs in fly photoreceptors . The proposed model of polarized transport in Drosophila photoreceptors is illustrated in Fig 8 . Membrane proteins are synthesized on the ER and transported together to the Golgi units . Within the Golgi units , the basolateral cargos are sorted into Chc/AP1 vesicles at the TGN , whereas apical cargos bound for the rhabdomere or stalk membrane are not segregated within the TGN but are rather transported together to REs . Rab6 is essential for the TGN–RE transport of apical cargos . The second sorting event at the RE segregates the cargos to the rhabdomere and stalk membrane , and carrier vesicles bearing the rhabdomere proteins are exported in a Rab11/dRip11/MyoV-dependent manner . The proposed model postulates the second sorting occurs at REs; however , this might occur earlier , for example at the TGN or between the TGN and REs . In addition , the direction of Rab6-mediated transport ( i . e . , anterograde or retrograde ) remains unresolved , because anterograde cargo transport can be due to the retrograde recycling transport of the TGN-resident proteins . The mechanisms of intra-Golgi transport and molecular role of Rab6 remain controversial despite substantial research . Three models of transport between the TGN and REs are based on models of intra-Golgi transport . First , the vesicle transport model posits that Rab6 is involved in the anterograde transport of tubules/vesicles bearing apical cargos from the TGN to REs . Second , in the cisternal maturation model , the TGN gradually maturates into REs as TGN-resident proteins are retrieved by Rab6-dependent retrograde tubule/vesicle-mediated transport [51 , 52] . Third , in the cisternal progenitor model , in which Rab GTPases including Rab6 define the identity of membrane subcompartments , continual fusion fission and the “Rab cascade” achieve cargo transport from the TGN to REs [53 , 54] . Future studies should elucidate the molecular mechanisms by which Rab6 functions in the transport of apical cargos from the TGN to REs .
Flies were grown at 20–25°C on standard cornmeal–glucose–agar–yeast food unless indicated otherwise . Carotenoid-deprived food was prepared from 1% agarose , 10% dry-yeast , 10% sucrose , 0 . 02% cholesterol , 0 . 5% propionate , and 0 . 05% methyl 4-hydroxybenzoate . EMS mutagenesis and F1 or F2 live-imaging screening were performed as described previously ( 27 ) . Starter strain with the second chromosome carrying proximal neoFRT at 40A was isogenized from the Bloomington Stock 5615 ( Bloomington , IN , USA ) , which is used in single nucleotide polymorphism ( SNP ) mapping [30] . The tester line w; Rh1Arr2GFP ey-FLP/TM6B were used for live imaging of mutant lines . Meanwhile , y w ey-FLP; P3RFP FRT40A/SM1 was used for immunostaining . Furthermore , y w 70FLP; Ubi-mRFP . nls FRT40A was used for the mosaic analysis of ovarian follicle cells . To quantify MVBs by electron microscopy , mutant whole-eye clones were made for Rich1 using the EGUF/hid method [55] . For Rab6546P and Rab6D23D , mutant whole-eye clones were not obtained probably because of the cell lethality of the mutants . Therefore , to partially rescue the lethality at early retinal development , YFP::Rab6wt was expressed by ey-Gal4 , and Rab6546P mutant whole-eye clones were obtained as described in the mutant analysis of Sec6 [11] . The following fly stocks were used: Rh1-Gal4 , heat shock-Gal4 , EGUF40A ( Bloomington stock number 5250 , Bloomington , IN , USA ) , UAS-YFP::Rab6wt ( Bloomington stock number 23251 , Bloomington , IN , USA ) , UAS-GFP::MyoVCT [9] , Rab6D23D FRT40A/CyO , Rich1 FRT80/TM3Kr>GFP , Rich2 FRT80/TM3 , Kr>GFP ( from Dr . Bellen , Baylor College of Medicine ) , and YRab6 ( from Dr . Brankatschk , Max Planck Institute ) . Fluorescent proteins expressed in photoreceptors were imaged by the water-immersion technique as described previously [8] . Briefly , late pupae with GFP-positive RFP mosaic retina were attached to the slide glass using double-sided sticky tape , and the pupal cases around the heads were removed . The pupae were chilled on ice , embedded in 0 . 5% agarose , and observed using an FV1000 confocal microscope equipped with a LUMPlanFI water-immersion 40× objective ( Olympus , Tokyo , Japan ) . Fixation and staining methods were performed as described previously by Satoh and Ready [56] . Primary antisera were as follows: rabbit anti-Rh1 ( 1:1000 ) [10] , chicken anti-Rh1 ( 1:1000 ) [1] , rabbit anti-GM130 ( 1:300 ) ( Abcam , Cambridge , UK ) , rabbit anti-NinaA ( 1:300 ) ( a gift from Dr . Zuker , Columbia University ) , mouse monoclonal anti Na+K+-ATPase alpha subunit ( 1:500 ascites ) ( DSHB , IA , USA ) , rat monoclonal anti DE-Cad ( 1:20 supernatant ) ( DSHB , IA , USA ) , mouse monoclonal anti-Notch ( 17 . 9C6 ) ( 1:10 supernatant ) ( DSHB , IA , USA ) , mouse monoclonal anti-FasIII ( 7G10 ) ( 1:10 supernatant ) ( DSHB , IA , USA ) , mouse monoclonal anti-Chp ( 24B10 ) ( 1:20 supernatant ) ( DSHB , IA , USA ) , rat anti-Crb ( a gift from Dr . Tepass , University of Toronto ) , rabbit anti-TRP ( a gift from Dr . Montell , Johns Hopkins University ) , rabbit anti-Rab7 ( 1:1000 ) ( a gift from Dr . Nakamura , Kumamoto University , Kumamoto , Japan ) , mouse monoclonal anti-Eys ( 1:20 supernatant ) ( DSHB , IA , USA ) , chicken anti-GFP ( 1:1000 ) ( Chemicon International Inc . , Billerica , MA , USA ) , rabbit anti-Chc ( 1:500 ) ( a gift from Dr . Kametaka , Nagoya University , Nagoya , Japan ) , mouse anti-Rab11 ( 1:250 ) [10] , and rabbit anti-Rab11 ( 1:300 ) [10] . Secondary antibodies were anti-mouse , anti-rabbit , anti-rat and/or anti-chicken antibodies labelled with Alexa Fluor 488 , 568 , and 647 ( 1:300 ) ( Life Technologies , Carlsbad , CA , USA ) or Cy2 ( 1: 300 ) ( GE Healthcare Life Sciences , Pittsburgh , PA , USA ) . Images of samples were recorded using an FV1000 confocal microscope ( 60× 1 . 42 NA objective lens; Olympus , Tokyo , Japan ) . To minimize bleed-through , each signal in double- or triple-stained samples was imaged sequentially . Images were processed in accordance with the Guidelines for Proper Digital Image Handling using ImageJ and/or Adobe Photoshop CS3 ( Adobe , San Jose , CA , USA ) . CFP was detected using the virtual channel function of the FV1000 , which stacks CFP images taken by the other dichroic mirror on other channel images ( i . e . , Alexa Fluor 488 , 568 , and 647 ) . However , the CFP laser beam axis was not perfectly aligned with the others probably because of the change in the dichroic mirror . Therefore , we scored this misalignment according to anti-GFP antibody staining of CFP with secondary antibodies conjugated with Alexa Fluor 488 and 647 , and calibrated each of the 3 colored pictures obtained . Newly eclosed flies fed carotenoid-deprived food were switched to carotenoid-deprived food with crystalline all-trans retinal ( Sigma , St . Louis , MO , USA ) in the dark . After 1 or 2 nights in the dark , the flies were irradiated with a 405-nm diode laser module at 30 mW for 20 min ( Pepaless , Hyogo , Japan ) to isomerize the all-trans retinal form to the 11-cis form and initiate Rh1 maturation . Electron microscopy was performed as described previously [8] . Samples were observed on a JEM1400 electron microscope ( JEOL , Tokyo , Japan ) , and montage images were taken by a CCD camera system ( JEOL , Tokyo , Japan ) . MVBs were counted in the tangential sections of 12–36 photoreceptor cells of individual flies , and the means and standard deviations were calculated from 4 flies for each allele . Meiotic recombination mapping was performed according to the standard method [57] . Briefly , to allow meiotic recombination between the proximal FRT , the phenotype-responsible mutation , and a distal miniature w+ marker , flies carrying isogenized chromosome 546P were crossed with flies with isogenized P{EP0511} , which carry the miniature w+ marker near the distal end of arm 2L . Female offspring carrying the mutated chromosome and the miniature w+-marked chromosome were crossed with males carrying FRT40A and P3RFP on the isogenized second chromosome and Rh1Arr2GFP on the third chromosome . Live imaging was to judge whether the resultant adult offspring with w+ mosaic ( i . e . , maternally inherited FRT and w+ ) inherited the mutation responsible for the Rh1 transport defect . The recovered flies were individually digested in 50 μL 200 ng/μL proteinase K in 10mM Tris-Cl ( pH 8 . 2 ) , 1 mM EDTA , and 25 mM NaCl at 55°C for 1 h and subsequently heat inactivated at 85°C for 30 min and 95°C for 5 min . Then , 0 . 5 μL digested solution was used as the template for PCR amplification for RFLP analysis according to the method described in the FlySNP database ( [58] , http://flysnp . imp . ac . at/index . php ) . The mutation responsible for the Rh1 transport defect was mapped between SNP markers 893 and 977 defined in the FlySNP database . For the whole-genome resequencing of the 546P mutant , the second chromosome was balanced over a balancer , CyO , P{Dfd-GMR-nvYFP} ( Bloomington stock number 23230 , Bloomington , IN , USA ) , to facilitate the isolation of homozygous embryo . Using the REPLI-G single-cell kit ( QIAGEN , Hilden , Germany ) , the genomic DNAs were independently amplified from four 546P homozygous embryos . The pair-end library was prepared using the Nextera DNA sample prep kit ( Illumina , San Diego , CA , USA ) for each embryo , 2 × 250 bp reads were obtained using the MiSeq v2 kit ( Illumina , San Diego , CA , USA ) . Reads were mapped to release 5 of the D . melanogaster genome using BWA 0 . 7 . 5a . The RFLP-mapped region of 546P was covered by reads with an average depth of 29 . 5x and width of 97 . 8% . Mapped reads were processed using Picard-tools 1 . 99 and the Genome Analysis Tool Kit 2 . 7–2 ( GATK , Broad Institute , Cambridge , MA , USA ) . Single nucleotide variants and indels were called using Haplotypecaller in GATK , and those of the isogenized starter stock were subtracted to extract the unique variants in 546P and annotated using SnpSift [59] In the mapped region , there were 4 unique variants that could cause some defects in gene function ( i . e . , CG31705 P453L , Rab6 R73STOP , ACXE A193T , and nimB1 G255E ) . Complementation tests over defined chromosomal deletions covering the RFLP-mapped region ( i . e . , BSC213 , BSC241 , BSC242 , BSC244 , ED775 , ED761 , ED780 , and ED791 ) revealed that only ED775 and ED761 were lethal to 546P homozygous larvae . The region deleted in both ED775 and ED761 contained the Rab6 R73STOP mutation in 546P , whereas this was not present in the other 3 variants described above . The point mutation of Rab6 R73STOP on 2L:12108583 ( Release 5 ) was verified by capillary sequencing of PCR-amplified fragment using the following primers: 5′-AGCGGAGCGAGAAGAGAGTT-3′ and 5′-GCTCCTGCTGTTGAAAAAGG-3′ . Full-length cDNA encoding Rab6 was amplified from a cDNA clone isolated from the fly retina cDNA library [60] and cloned into pQE60 . The 6xHis-tagged Rab6 protein was expressed in E . coli pG-KJE8/BL21 ( TAKARA ) at 23°C and purified in native conditions using Ni-NTA Agarose ( QIAGEN , Hilden , Germany ) . To obtain antisera , 1 rabbit and 2 guinea pigs were immunized 6 times with 200 μg 6xHis-Rab6 protein ( Hashimoto , Wakayama , Japan ) . We designated the resultant antisera Rb anti-Rab6 , GP1 anti-Rab6 , and GP2 anti-Rab6 . Immunoblotting was performed as described previously [8] . The following antibodies were used: rabbit anti-Rab6 ( 1:2000 concentrated supernatant ) ( made in house ) , 2 guinea pig anti-Rab6 ( 1:2000 concentrated supernatant ) ( made in house ) as primary antibodies; HRP-conjugated anti-rabbit or anti-guinea pig IgG antibodies ( 1:20 , 000 , Life Technologies , Carlsbad , CA , USA ) as a secondary antibody . Signals were visualized using enhanced chemiluminescence ( Clarity Western ECL Substrate; Bio-Rad , Hercules , CA , USA ) and imaged using ChemiDoc XRS+ ( Bio-Rad , Hercules , CA , USA ) . | Cells in animal bodies have multiple plasma membrane domains; this polarized characteristic of cells is essential for their specific functions . Selective membrane transport pathways play key roles in the construction and maintenance of polarized structures . Drosophila photoreceptors with multiple plasma membrane domains are an excellent model of polarized transport . We performed genetic screening and identified a Rab6 null mutant with a rare phenotype characterized by a loss of 2 apical transport pathways with normal basolateral transport . Although Rab6 functions in the Golgi are well known , its function in polarized transport was unexpected . Here , we found that Rab6 and its effector , Rich , are required for multiple apical transport pathways but not the basolateral transport pathway . Our findings strongly indicate that the membrane proteins delivered to multiple polarized domains are not sorted simultaneously: basolateral cargos are segregated before the Rab6-dependent process , and cargos going to multiple apical domains are sorted after Rab6-dependent transport from the trans-Golgi network to the Golgi-associated Rab11-positive compartment , which presumably recycles endosomes . Our finding of the function of Rab6 in polarized transport will elucidate the molecular mechanisms of polarized transport . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"vesicles",
"cell",
"processes",
"social",
"sciences",
"neuroscience",
"membrane",
"proteins",
"golgi",
"apparatus",
"eyes",
"cellular",
"structures",
"and",
"organelles",
"endosomes",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"animal",
"cells",
"sensory",
"receptors",
"head",
"cell",
"membranes",
"signal",
"transduction",
"psychology",
"immunostaining",
"cell",
"biology",
"anatomy",
"secretory",
"pathway",
"neurons",
"photoreceptors",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"afferent",
"neurons",
"sensory",
"perception",
"ocular",
"system"
]
| 2016 | Rab6 Is Required for Multiple Apical Transport Pathways but Not the Basolateral Transport Pathway in Drosophila Photoreceptors |
This cross-sectional study aimed to investigate the current prevalence and risk factors associated with intestinal polyparasitism ( the concurrent infection with multiple intestinal parasite species ) among Orang Asli school children in the Lipis district of Pahang state , Malaysia . Fecal samples were collected from 498 school children ( 50 . 6% boys and 49 . 4% girls ) , and examined by using direct smear , formalin-ether sedimentation , trichrome stain , modified Ziehl Neelsen stain , Kato-Katz , and Harada Mori techniques . Demographic , socioeconomic , environmental , and personal hygiene information were collected by using a pre-tested questionnaire . Overall , 98 . 4% of the children were found to be infected by at least one parasite species . Of these , 71 . 4% had polyparasitism . The overall prevalence of Trichuris trichiura , Ascaris lumbricoides , hookworm , Giardia duodenalis , Entamoeba spp . , and Cryptosporidium spp . infections were 95 . 6% , 47 . 8% , 28 . 3% , 28 . 3% , 14 . 1% and 5 . 2% , respectively . Univariate and multivariate analyses showed that using an unsafe water supply as a source for drinking water , presence of other family members infected with intestinal parasitic infections ( IPI ) , not washing vegetables before consumption , absence of a toilet in the house , not wearing shoes when outside , not cutting nails periodically , and not washing hands before eating were significant risk factors associated with intestinal polyparasitism among these children . Intestinal polyparasitism is highly prevalent among children in the peninsular Malaysian Aboriginal communities . Hence , effective and sustainable control measures , including school-based periodic chemotherapy , providing adequate health education focused on good personal hygiene practices and proper sanitation , as well as safe drinking water supply should be implemented to reduce the prevalence and consequences of these infections in this population .
Intestinal parasitic infections ( IPI ) are still public health problems in many communities , particularly among children in rural areas of developing countries . It is estimated that more than 2 billion people worldwide are infected with IPI and more than half of the world's population are at risk of infection [1] , [2] . These infections are caused by helminth parasites such as soil-transmitted helminths ( Ascaris lumbricoides , Trichuris trichiura , Strongyloides stercoralis , and hookworm ) , Taenia spp . and Hymenolepis nana or by protozoa such as Entamoeba histolytica , Giardia duodenalis , and Cryptosporidium spp . IPI are associated with high morbidity particularly among young children and women of childbearing age , and have been termed as ‘the cancers of developing nations’ by Egger et al . [3] . IPI can occur in silence as chronic infections and infected individuals are either asymptomatic or suffering from mild diseases . However , acute and severe IPI , especially with pathogenic Entamoeba and Giardia , may cause fatal diarrhea especially among children and both are commonly associated with travellers' diarrhea [4] , [5] . Moreover , Entamoeba can cause invasive intestinal infection or disseminate to the liver ( and rarely to the lung and the brain ) causing amebic liver abscess with about 100 , 000 deaths annually , making amebiasis the second leading cause of death from protozoal diseases , after malaria [6] , [7] . On the other hand , opportunistic IPI such as Cryptosporidium , Isospora belli , Microsporidia , and Strongyloides infections are commonly reported among immunocompromised individuals with significant morbidity and mortality [8] , [9] . There is a general acceptance that severe IPI are likely to result in failure to thrive and poor growth in children [10]–[12] , vitamin A deficiency [13] , [14] , iron deficiency anemia [15] , [16] , and poor educational performance [17] , [18] . Moreover , recent studies highlighted the impact of polyparasitism on the hosts' immunity and showed that polyparasitism is associated with higher mortality rates and may increase the susceptibility to other infections relative to infection with a single parasite [19]–[21] . In Malaysia , despite sustained socioeconomic and infrastructural development , IPI are still highly prevalent , especially among impoverished rural communities . Several previous studies have investigated the prevalence of IPI and reported the findings either on a specific parasitic infection , monoparasitism [22]–[24] , or groups of parasitic infections such as soil-transmitted helminths ( STHs ) and intestinal protozoa [25]–[29] , or on the overall prevalence of IPI [30] , [31] . These studies revealed high prevalence rates of IPI with prominent morbidity among Orang Asli and other rural populations in Malaysia . However , there has been limited information on polyparasitism among these populations . The present study was , therefore , carried out to determine the prevalence of polyparasitism and to investigate its associated risk factors among Orang Asli school children in Lipis district , Pahang , Malaysia by comparing children who were infected with multiple parasites with those who had only single parasitic infection or were not infected at all .
The present study was carried out according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Medical Ethics Committee of the University of Malaya Medical Centre , Malaysia ( reference number: 932 . 7 ) . Permission was also obtained from the Department of Orang Asli Development ( reference number: JHEOA . PP . 30 . 052 Jld . 6 ) and the Department of Education , Pahang [reference number: JPNP . SPS . UPP . 600-2/6 ( 80 ) ] . Before the commencement of the study , meetings were held with the heads of villages , headmasters , and teachers of both schools to provide information about the objectives and protocol of the study and their consents were obtained . During fieldwork , the purpose and procedures of the study were explained to the children and their parents/guardians . Moreover , they were informed that their participation was voluntary and they could withdraw from the study at any time without citing any reason whatsoever . Written and signed or thumb-printed informed consent was obtained to conduct the study from parents or guardians on behalf of their children before starting the survey , and these procedures were approved by the Medical Ethics Committee of the University of Malaya Medical Centre . All the infected children were treated with a 3-day course of 400 mg albendazole . Each child chewed the tablets with chocolate flavoured biscuits before swallowing them while being observed by a researcher and medical officer ( direct observed therapy ) [2] . Albendazole is considered as the drug of choice for Ascaris , Trichuris , and hookworm infections and it is also effective for Giardia infection [24] . Moreover , all results were submitted to relevant authorities for further follow-up . This cross-sectional study was carried out between January and April 2012 , in the Lipis district of Pahang state , located at the center of Peninsular Malaysia , about 200 km northeast of Kuala Lumpur . The climate is equatorial with hot-humid conditions and rainfall throughout the year . Most of the houses in the area are made of wood or bamboo with inadequate sanitary facilities . More than half of the houses had no piped water supply . However , the villages are located alongside rivers which are the main source of water for daily activities . The majority of the people in the villages work as labourers in palm oil and rubber plantation or are engaged in the selling of forest products , to earn a living . This study was conducted in two Orang Asli areas namely Kuala Koyan ( two villages ) and Pos Betau ( 18 villages ) in Lipis district ( Figure 1 ) . There are two primary schools at these areas namely , Sekolah Kebangsaan Kuala Koyan ( SKKK ) and Sekolah Kebangsaan Pos Betau ( SKPB ) . The villages and schools were selected from the available official village list in collaboration with the Department of Orang Asli Development ( JAKOA ) . The selection criteria were as follows: ( i ) school located in a rural area , ( ii ) easy access from the main roads , and ( iii ) school enrolment of more than 100 pupils . Orang Asli translated as ‘original or first people’ are the Aboriginal minority peoples of Malaysia; representing 0 . 7% of the country's total population . Although the total enrolment of the schools was 783 pupils ( 167 SKKK and 616 SKPB ) , only 650 were present during sampling visits . Of these children who were present , 498 children aged 6–12 years ( 252 boys and 246 girls ) had agreed voluntarily to participate in this study and had met the inclusion criteria ( written consent signed by the guardian , completed questionnaire and delivered stool samples for examination ) . The minimum sample size required for this study was calculated according to the formula provided by Lwanga and Lemeshow [32] . At a 5% level of significance and a 95% confidence level , the minimum number of participants required for the study was estimated at 288 , assuming that the prevalence of intestinal parasitic infection among Orang Asli children was about 75% as previously reported [28] , [30] . The study flow chart and the participation of children are shown in Figure 2 . The demographic , socioeconomic and environmental information , personal hygiene practices , history of receiving anthelmintic treatment and health status of the participants were collected by using a pre-tested questionnaire . The questionnaire was designed in English and then translated into Malay . Two research assistants from the Department of Parasitology , University of Malaya were trained for the purpose of this study on how to administer the questionnaire . The children and their parents were interviewed in their home settings . During the interviews , observations were made on the personal hygiene of the children ( e . g . , cutting fingernails , wearing shoes when outside the house , and hands and clothes cleanliness ) , household cleanliness and the availability of functioning toilets and piped water . The children were given a clearly labeled , wide mouth and screw-caps containers and were instructed to bring their early morning stool samples the next day . The collected samples were transported ( within 5 hours of collection ) in suitable cool boxes at temperature between 4 and 6°C for examination at the stool processing laboratory in the Department of Parasitology , Faculty of Medicine , University of Malaya . The samples were examined by using six different techniques; namely , direct smear , formalin-ether sedimentation , Kato-Katz , Harada Mori , trichrome stain , and modified Ziehl Neelsen stain techniques . All the samples were screened first by direct smear technique . Then , formalin-ether sedimentation technique was used to increase the detection rates , especially when the samples were negative by direct smear [33] . For the estimation of intensity of STH infections , egg counting was done by using the Kato-Katz technique [2] . Harada Mori culture technique was done to detect hookworm larvae in light infections as described elsewhere [34] . The larvae were collected and examined to distinguish between hookworm and S . stercoralis by the characteristic morphology of the larvae ( i . e . , size of buccal cavity and presence of genital primordium ( rhabditiform larvae ) or presence of notched tail ( filariform larvae ) . For STH infections , positive samples were recorded according to species and the intensity of infection was recorded as eggs per 1 g of stool ( EPG ) and was graded as heavy , moderate , or light according to the criteria proposed by the World Health Organization [2] . A suitable amount ( approximately 10 g ) of faeces was mixed thoroughly and fixed in polyvinyl alcohol ( PVA ) for the detection of intestinal protozoa ( Giardia and Entamoeba ) using trichrome staining technique [35] . Moreover , fecal smears were prepared and stained with modified Ziehl Neelsen stain , according to Henriksen and Pohlenz [36] , for the detection of Cryptosporidium oocysts . Overall , the samples were considered as positive if the eggs/cysts/trophozoites/oocysts were detected using at least one of these techniques . For quality control , duplicate slides were prepared from 20% of the samples for each diagnostic technique and the slides were read by two different microscopists . Data were reviewed and double-checked before and after data entry by two different researchers . Only those participants who had complete data records , including results of intestinal parasites by the different methods and complete questionnaire , were retained for the final analyses . For descriptive analysis , prevalence of infections and illnesses was expressed in percentage , while mean ( standard deviation; SD ) or median ( interquartile range; IQR ) was used to present the quantitative data and results were presented in tables . All quantitative variables were examined for normality by Kolmogorov-Smirnov Z test before analysis . To assess polyparasitism , a variable termed “infection status” was created to denote conditions of non-infected , monoparasitism ( infection with a single parasite species ) , or polyparasitism ( co-infections with two or more parasite species ) . Pearson's χ2 test was used to investigate the association between polyparasitism as the dependent variable and demographic factors ( age , gender , and household size ) , socioeconomic factors ( parents' educational and employment status , family monthly income , source of drinking water , presence of toilet in the house , having domestic animals in the households , and presence of infected family member ) , and personal hygiene practices ( washing hands before eating and after defecation , washing fruits and vegetables before consumption , wearing shoes when outside , eating soil ( geophagy ) , boiling drinking water , cutting nails periodically , and indiscriminate defecation ) as explanatory variables . All variables in the survey were coded in a binary manner as dummy variables . For example , polyparasitism ( positive = 1 , negative = 0 ) ; gender ( boys = 1 , girls = 0 ) ; presence of toilet in the house ( no = 1 , yes = 0 ) , and washing vegetables before eating ( no = 1 , yes = 0 ) . Family size was categorized into two groups ( ≥7 and <7 members ) , and age of participants was categorized into two groups that were below 10 years and ≥10 years according to previous studies conducted among Orang Asli [13] , [24] , [28] . Odd ratios ( ORs ) and 95% confidence intervals ( CIs ) were computed for all variables . To adjust for multiple comparisons , we calculated critical significance thresholds for each factors group using the sequential Bonferroni correction [37] . Multivariable logistic regression model was performed to identify risk factors that were significantly associated with intestinal polyparasitism; coded as 1 = polyparasitism , 0 = mono-parasitism and uninfected . In order to retain all possible significant associations , variables that showed an association with P≤0 . 25 were used in the logistic regression model , as suggested by Bendel and Afifi [38] . Moreover , gender variable was also included in the multivariable analysis as it has been considered as an important behavioral modifying factor [39] . Overall , 15 variables met our inclusion criteria to the final model . Many explanatory variables were included separately in this study as there is limited preceding evidence to support inclusion of one factor before or instead of others . Moreover , the different personal hygiene practices are included separately in order to reflect distinct parasite transmission routes . Population attributable risk fraction ( PARF ) was calculated for significantly associated risk factors [40] . Data analysis was done by using SPSS for WINDOWS ( version 13 . 0; SPSS Inc , Chicago , IL ) , and significance was set at P<0 . 05 .
The demographic and socioeconomic characteristics of the participants are shown in Table 1 . Four hundred and ninety eight school children aged between 6 and 12 years , with a median age of 9 years ( IQR 8–11 years ) had participated in this study . The children were from 20 villages in Lipis , Pahang . Of these children , 50 . 6% were boys and 49 . 4% were girls . In general , poverty prevails in these communities and about two thirds of the families had low monthly income ( <RM 500 , 156 US$ ) ; the poverty income threshold in Malaysia [41] . Moreover , about half and 40 . 2% of the mothers and fathers , respectively , had no formal education . A majority of the mothers and two thirds of the fathers are not working . Those working were mainly engaged in agriculture ( rubber and oil palm plantations ) , forestry , fishing , and related occupations . Almost half of the houses are without toilets and it was found that Orang Asli people preferred to defecate at the site of the streams . The data from the questionnaire survey indicating none of the children had received anthelmintic treatment in the 6 months prior to the study . Fecal samples were screened by different techniques for the presence of intestinal parasites and the results are presented in Table 2 . Overall , 98 . 4% ( 490/498 ) of the children were found to be infected with at least one intestinal parasite species . The results showed that T . trichiura was the predominant species with a prevalence rate of 95 . 6% ( 476/498 ) , followed by A . lumbricoides ( 47 . 8% ) , G . duodenalis ( 28 . 3% ) , and hookworm ( 27 . 9% ) . The results also showed that almost two thirds and 62 . 0% of the T . trichiura and A . lumbricoides infections , respectively , were of moderate-to-heavy intensity ( mean EPG faeces of ≥5 , 000 for A . lumbricoides and ≥1 , 000 for T . trichiura ) , while all hookworm infections were of light intensity ( ≤2 , 000 EPG ) . The detection rates of T . trichiura , A . lumbricoides , hookworm , G . duodenalis and E . histolytica/E . dispar by formalin-ether sedimentation method were found to be 93 . 4% , 46 . 6% , 21 . 9% , 25 . 9% and 12 . 0% , respectively . On the other hand , lower detection rates were noted with Kato-Katz technique; T . trichiura ( 86 . 7% ) , A . lumbricoides ( 45% ) and hookworm ( 10 . 6% ) . However , the detection rate of hookworm was substantially higher ( 22 . 5% ) by using Harada Mori technique while almost similar detection rates for G . duodenalis ( 23 . 5% ) and E . histolytica/E . dispar ( 11 . 4% ) were observed by using trichrome staining technique when compared with the results obtained by the formalin-ether sedimentation method . The age-associated prevalence of different reported IPI is illustrated in Figure 3 . The age- associated patterns of infection prevalence were generally similar among the different parasite species . T . trichiura was the most prevalent infection in all ages followed by A . lumbricoides and G . duodenalis infection . All IPI , except Cryptosporidium , occurred in all ages with the highest prevalence noted among children aged 9 years . When we compared the results according to gender , boys were noted to have higher prevalence of hookworm and G . duodenalis while other IPI were higher among the girls . However , the differences were not statistically significant ( statistical analysis not shown ) . On the other hand , Cryptosporidium was not reported among children aged 6 years and the highest prevalence was reported among girls aged 9 years ( 17 . 9% ) . Of the infected children , 71 . 4% had polyparasitism , while the remaining 28 . 6% were infected with a single parasite species . Of those who had polyparasitism , 189 ( 54 . 0% ) and 88 ( 25 . 1% ) had double and triple infections , respectively . Moreover , 5 . 4% of them harbored five different parasite species concurrently . T . trichiura and A . lumbricoides were the most prevalent co-infection representing 54 . 0% of the polyparasitism prevalence , followed by the combination of T . trichiura and G . duodenalis ( 28 . 4% ) . Moreover , 18 . 6% of the infected children had the three STH species ( T . trichiura , A . lumbricoides and hookworm ) . Fecal samples were also examined for the presence of other intestinal parasites and the children were found to be positive for Entamoeba coli ( 15 . 5% ) , Blastocystis sp . ( 15 . 1% ) , Iodamoeba bütschlii ( 6 . 8% ) , Chilomastix mesnili ( 4 . 8% ) , and Endolimax nana ( 2 . 6% ) . On the other hand , S . stercoralis larvae were not detected among these children . Table 3 displays the associations of polyparasitism with demographic , socioeconomic , environmental and personal hygiene factors . The results showed that the prevalence of polyparasitism was significantly higher among children aged <10 years ( 73 . 8% ) when compared with those aged ≥10 years ( 65 . 4% ) . With regard to socioeconomic factors , children born to non-educated mothers ( i . e . <6 years formal education ) had significantly higher prevalence of polyparasitism ( 76 . 3% ) as compared to those of educated mothers ( 64 . 3% ) . It was also found that polyparasitism was higher among children from families with low household monthly income ( <RM 500 ) ( 74 . 4% ) when compared with children who belong to families with higher household monthly income ( 63 . 0% ) . The results further showed that children who live in houses without toilets ( 80 . 1% ) and those who use unsafe sources for drinking water ( 80 . 2% ) had higher prevalence of polyparasitism when compared to those who live in houses with functioning toilets ( 61 . 8% ) and those who use piped water ( 60 . 8% ) . Examining the association of polyparasitism with personal hygiene practices among these children showed that the prevalence of infection was significantly higher among children who do not wash their hands before eating ( 76 . 7% ) and also after defecation ( 74 . 3% ) , those who walk barefooted ( 75 . 4% ) and those who do not cut their fingernails periodically ( 77 . 5% ) when compared to their counterparts . Moreover , children who do not wash fruits and vegetables before consuming had significantly higher prevalence of polyparasitism compared to those who wash the fruits and vegetables before consuming ( P<0 . 001 ) . Besides these variables , the presence of other family members infected with IPI was significantly associated with higher rates of polyparasitism among these children ( P<0 . 001 ) . Using the Bonferroni adjustment , a series of χ2 analyses revealed no association between polyparasitism and age groups , not washing their hands after defecation , and not wearing shoes when outside the house . Table 4 shows that the multiple logistic regression model identified seven variables as significant risk factors of polyparasitism . Hosmer–Lemeshow test , used for the inferential goodness-of-fit test , showed that the model was fit to the data well ( χ2 = 9 . 2; P = 0 . 320 ) . The results of the logistic regression model including all 15 factors confirmed that children who use unsafe sources for drinking water and/or live in houses without proper toilets were at a two times higher odds of having polyparasitism when compared with their counterparts . Likewise , the presence of other family members infected with multiple parasitic infections increased the children's odds for the polyparasitism by 1 . 7 times . It was found that not washing vegetables before consumption increased children's odds for polyparasitism when compared with always washing vegetables by 2 . 5 times . Furthermore , children who do not wash their hands before eating had 1 . 6 times odds while those who do not cut their nails periodically and/or walk barefooted were at twice the odds of polyparasitism when compared with their counterparts . When stratified by age group ( <10 and ≥10 years ) , similar risk factors were identified among both groups . PARF analysis showed that the number of polyparasitism cases would be reduced by 18 . 0% , 13 . 6% , 11 . 9% , and 6 . 8% if all children had good standards of personal hygiene; namely , washing vegetables before consumption , washing hands before eating , cutting nails periodically , and wearing shoes when outside the house , respectively . Moreover , 13 . 5% , 12 . 1% , and 10 . 3% of the polyparasitism cases could be reduced when the children in this population had a provision of clean and safe drinking water , had toilet facilities at home , and had no other family members infected with polyparasitism , respectively .
The findings of the present study reveal that polyparasitism is very common among Orang Asli school children in rural Malaysia . Hence , there is an urgent need to implement an innovative and integrated control program to reduce the prevalence and intensity of these infections significantly and to save these children from the negative impact of IPI as a part of the efforts to improve the quality of life of Orang Asli population . Based on the findings , there is a great need for a proper health education regarding good personal hygiene practices and community mobilization to enhance prevention and instil better knowledge on IPI transmission and prevention in these communities . Moreover , sustainable school-based deworming , providing proper and adequate sanitation and safe water supply should be considered in the control program targeting this population . | Intestinal parasitic infections ( IPI ) are still a major public health problem worldwide , with more than 2 billion people infected with at least one parasite species . Despite efforts to improve the quality of life of the Orang Asli population in rural Malaysia , IPI are still highly prevalent and of serious concern in this population , especially among children . We screened 498 school children in Lipis district , Pahang , Malaysia for the prevalence and risk factors of polyparasitism ( concurrent infection with multiple parasite species ) . Overall , 98 . 4% of the children were found to be infected by at least one parasite species , with 71 . 4% of them having polyparasitism . Using an unsafe water supply as a source for drinking water , presence of another family member infected with IPI , not washing vegetables before consumption , absence of a toilet in the house , not wearing shoes when outside , not cutting nails periodically , and not washing hands before eating were significantly associated with intestinal polyparasitism among these children . Our findings revealed an urgent need to implement an effective and integrated control program to reduce the prevalence of IPI as a part of the efforts to improve the quality of life in Orang Asli communities . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"cryptosporidiosis",
"helminth",
"infections",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"amebiasis",
"parasitic",
"intestinal",
"diseases",
"giardiasis",
"neglected",
"tropical",
"diseases",
"tropical",
"diseases",
"protozoan",
"infections",
"soil-transmitted",
"helminthiases",
"parasitic",
"diseases"
]
| 2014 | Epidemiology of Intestinal Polyparasitism among Orang Asli School Children in Rural Malaysia |
From genomic association studies , quantitative trait loci analysis , and epigenomic mapping , it is evident that significant efforts are necessary to define genetic-epigenetic interactions and understand their role in disease susceptibility and progression . For this reason , an analysis of the effects of genetic variation on gene expression and DNA methylation in human placentas at high resolution and whole-genome coverage will have multiple mechanistic and practical implications . By producing and analyzing DNA sequence variation ( n = 303 ) , DNA methylation ( n = 303 ) and mRNA expression data ( n = 80 ) from placentas from healthy women , we investigate the regulatory landscape of the human placenta and offer analytical approaches to integrate different types of genomic data and address some potential limitations of current platforms . We distinguish two profiles of interaction between expression and DNA methylation , revealing linear or bimodal effects , reflecting differences in genomic context , transcription factor recruitment , and possibly cell subpopulations . These findings help to clarify the interactions of genetic , epigenetic , and transcriptional regulatory mechanisms in normal human placentas . They also provide strong evidence for genotype-driven modifications of transcription and DNA methylation in normal placentas . In addition to these mechanistic implications , the data and analytical methods presented here will improve the interpretability of genome-wide and epigenome-wide association studies for human traits and diseases that involve placental functions .
Functional genomic approaches using combinations of genomic , epigenomic and transcriptomic data can define regulatory landscapes and provide a basis for biomarker discovery and insights into disease etiology . Advances in sequencing technologies have resulted in a surge in the number and complexity of multi-omics datasets , providing unprecedented opportunities to map the regulatory landscape , while imposing substantial analytical challenges . Increasingly , researchers acknowledge the limited usefulness of genome-wide approaches that examine a single component of transcriptional regulation in isolation and appreciate the growing importance of “multi-omic” analyses . Indeed , genomic sequence variation , epigenetic and post-translational regulators are interdependent and jointly contribute to the normal functioning or dysfunction of a tissue , calling for integrative approaches to fully appreciate the interactions between these different layers of regulation . Originally two main lines of investigation were pursued—association studies and molecular quantitative trait loci ( QTL ) analysis . Association studies aim to identify traits associated with genetic variants ( Genome Wide Association Studies , GWAS ) and more recently traits associated with changes in DNA methylation ( Epigenome Wide Association Studies , EWAS ) . Over the past decades , thousands of common genetic variants associated with specific diseases or phenotypes have been cataloged [1–3] . Building on the GWAS model , EWAS rapidly became a fast-growing area of research [4] . However , these discovery approaches provide limited insights into causal mechanisms . To address these limitations , QTL analyses have been implemented linking genetic variants with changes in expression ( eQTL ) , DNA methylation ( mQTL; alternatively abbreviated as meQTL ) or other transcriptional regulatory mechanisms . Catalogs of thousands of tissue-specific and shared regulatory eQTL variants at high resolution , such as in the Genotype Tissue Expression ( GTEx ) catalog , provide insights into the diversity and regulation of gene expression across various tissues [5] . Although cytosine methylation varies with sex , age or exposure to environmental factors [6] and changes in methylation patterns have been associated with many common diseases [7] , DNA methylation is also under strong genetic influence , and locus-specific methylation levels are often correlated in related individuals [8 , 9] . Twin studies provide additional evidence of the underlying genetic effect on DNA methylation pattern [9 , 10] and it has been estimated that more than 30% of the methylation variance can be attributed to genotype [11] . More directly , genetic variants and haplotypes have been shown to influence local DNA methylation patterns , most often in cis . Loci , where DNA methylation is under genetic control , are known as methylation quantitative trait loci ( mQTL ) . The physical counterpart of cis-acting mQTLs , which can be directly scored by bisulfite sequencing ( bis-seq; methyl-seq ) is haplotype-dependent allele-specific methylation ( hap-ASM ) . We ( CD , BT ) have previously characterized mQTLs and hap-ASM in multiple human tissue types , including T lymphocytes and brain cells , as well as a small series of placentas [12 , 13] . In that work , we noted that an analysis of a larger independent series of placentas was in progress , which is represented by the current study . Unlike most other human tissue types , the number of genome-wide–omic level studies involving placentas is relatively limited with placenta being absent from most initiatives such as the GTEx consortium or the GWAS database . However , in view of its critical role , there is growing interest in better understanding the impact of placenta malfunctions throughout fetal life . During a key developmental window , the placenta controls fetal access to nutrients , hormone production and mitigation of adverse effects from the environment , with placental dysfunction resulting in chronic diseases such as heart disease , type 2 diabetes or cancer [14–17] . Evidence suggests that such perturbations of the intrauterine environment alter the appropriate genetic programming and disrupt placental and fetal development [18–20] . Although there is strong support for changes in DNA methylation to be involved , a direct association between environmental conditions , methylation alterations , and gene expression is difficult to confirm . The unique function of the placenta is reflected through its unique transcriptome and methylome profiles . In a previous study , Storvik et al . [21] showed a relatively poor correlation in term of transcriptome between placenta and 34 other tissues based on microarray data with many mRNA and miRNA species specific to the placenta . Similarly , as previously shown by our group and others , the mean genome-wide level of DNA methylation across placentas is lower than that found in other tissues , with an increased representation of low and intermediate levels [13 , 22] . Such singularities emphasize the need for specific investigations of the placenta regulatory landscape . Along these lines , Peng et al . [23] performed an eQTL study using 159 placentas . They examined correlations between GWAS data and the newly identified eQTLs and found evidence for eQTLs potentially driving postnatal disease susceptibility , supporting the Developmental Origins of Health and Disease ( DOHaD ) hypothesis . These findings confirm the need for a thorough characterization of the placenta regulatory landscape , preferably not limited to genetic variant and gene expression . Among the potential limitations of mQTL and eQTL studies are subpopulation effects , cell specificity , genetic background or genomic context , as many factors that need to be accounted for to achieve reproducibility and accuracy . Further challenges include the integration of multiple types of data calls for quality control assessments , independent controls for both false-positive and false-negative findings , and a limited understanding of stochastic variation in the different signals . Finally , the availability of a reference dataset is key to support ongoing functional genomics analysis . However , the cell dependent aspect of transcriptional regulation does not allow for direct carryover of findings from one tissue or cell type to another . Lack of ethnic and racial diversity can also be a limitation , with the most comprehensive studies up to date involving predominantly Caucasian populations [24–26] . These considerations have motivated the present study where we set out to generate essential reference data sets of genotypes , gene expression , and DNA methylation patterns , based on a representative population of placentas; an understudied yet important tissue . In addition , we aimed to develop stringent analytical methods to serve as a blueprint for future studies . We intensively characterized the effects of genetic variation on gene expression ( eQTLs ) and DNA methylation ( mQTLs ) and developed a comprehensive approach to identify methylation-sensitive transcription ( abbreviated here as expression-relevant Quantitative Trait Methylation; eQTM ) in normal human placentas . We emphasize technical caveats , independent validations , and consideration of both false-positive and false-negative findings . Our integrative multi-omic approach provides a multilayered view of gene regulation influenced by common genetic variants in the human placentas , thus providing a basis for future advances in individualized medicine .
Actual functional genomics studies are suffering from the overrepresentation of Caucasian backgrounds , which limit the reproducibility of the findings in more heterogeneous and representative populations . Aware of this limitation , it was important to characterize the ancestral history of our population to confirm the mix-population background of our cohort and to identify outliers that should be excluded from further analysis . We performed local ancestry deconvolution using PCA analysis based on the reference population using genotype data from the 1000 Genomes Project from Caucasians ( CEU ) , Africans ( YRI ) and East Asians ( CHB + JPT ) . Our data showed a clear clustering to populations , with some enrichment for mixed populations represented as expected ( S1C Fig ) validating the use of our cohort as a representative cohort for control placenta samples . To assure the relevance of our human placenta cohort as representative of ‘normal’ individuals , we used quality control measures to assess the impact of variability in the datasets . Sample heterogeneity was estimated running Pearson’s correlation across expression and methylation data . Global correlation among our samples ranges between 0 . 92 and 0 . 99 for expression values and between 0 . 93 and 0 . 99 for DNA methylation values ( S1A and S1B Fig ) . As correlation can be artificially over-estimated when interrogating large datasets with extreme values , further analysis aimed at selecting those genes and loci with greater variability . Variability was estimated using median absolute deviation ( MAD ) as previously described [27] . After assigning a variability score to each interrogated CpG or transcript , variability was divided using quartile distribution creating 4 categories: no variability , low , medium and high variability . Loci and transcripts from the 4th quartile of variability ( high ) were used to run correlation using a similar approach and range from 0 . 75 to 0 . 96 for expression values and from 0 . 64 to 0 . 98 for DNA methylation values . Principal Component Analysis ( PCA ) was then used to describe variability across samples and link variability to known covariates including batch effects , biological and clinical cofactors . As expected , the batch of sequencing is contributing to most of the variability in both expression and methylation datasets ( S2 Fig ) . Interestingly , when focusing on the principal component accounting for most of the variability , we also found a significant association with preeclampsia status and mode of delivery in both datasets ( S2 Fig ) . However , our approach still suffers from limitations as it only provides us with information on known confounders . For example , even though all samples were taken from the same region of the placentas ( Methods ) , heterogeneity in tissue composition likely exists and could influence inter-sample variability but is not reflected in this analysis . The placenta presents a unique transcriptome and methylome profile as demonstrated by the poor correlation in term of transcriptomic profiles between placenta and 34 other tissues based on microarray data [21] . Among these tissues , the lung was the most like placenta and liver showed the weakest correlation . Because microarrays do not query the full extent of the transcriptome , we decided to perform a similar correlation analysis using RNAseq data comparing placenta transcriptomes from our study with transcriptomes of the 53 available tissues from the GTEx consortium . For each tissue , we considered median TPM values across samples and only processed genes that were expressed ( TPM>0 . 1 ) across each tissue type . The correlation scores ranged from 0 . 02 to 0 . 27 reflecting a high diversity between placenta and the other tissues ( S1 Table ) . Interestingly , among the tissues overlapping between the GTEx consortium and the one tested by Storvik et al . , we confirmed lung as being closest to the placenta and liver to be the one with the weakest correlation . The mean genome-wide level of DNA methylation across placentas is lower than that found in other tissues , with an increased representation of low and intermediate levels [13 , 22] . High variability in DNA methylation has been studied in CD34+ hematopoietic stem and progenitor cells [27] and proposed as a marker that can help localize key regulatory elements involve in cell-lineage commitment and cell specific functions . Therefore , we decided to further investigate genes with increased variability in expression or DNA methylation levels across samples and to look for associated pathways . As described above , gene and CpG were classified based on their variability across samples using MAD . We then focus on the top 500 variable genes or CpGs and performed pathways enrichment analysis ( S2 and S3 Tables ) . Traditional gene set enrichment analysis ( GSEA ) does not take into account the physical characteristics of the gene and has been shown to be biased by factors such as the length of the gene [28] . To address this , we used the Bioconductor package GOseq [29] developed to control for variability of the length of genes . For each CpG , the closest associated gene was considered . For the CpG analysis , we adapted the original version of GOseq to control for the number of probes by gene instead of gene length , as length is not relevant in this case . Interestingly , in both datasets , the majority of the pathways ( S4 Table ) were related to the immune response involving genes from the HLA family suggesting that heterogeneity among samples could be driven by a different level of inflammations in each placenta . Looking only at expression data , the top pathway is KEGG "ECM-receptor interaction" . ECM related genes have been previously associated with placental development [30] and have been shown to be correlated with oxygen tension and nutrient availability [31] . Therefore , variability in the expression for these pathway-associated genes may reflect various in utero exposures with impact on placenta development and potential long term consequences even in the absence of clear phenotype at birth . To generate maps of associations between genetic variants and gene expression or CpG methylation patterns specific to the human placenta , we performed eQTL and mQTL analysis across 1 , 374 , 581 SNPs , 23 , 003 genes , and 485 , 578 CpG sites respectively . We focused on identifying cis-eQTLs within windows of 1 Mb between the gene transcription start site ( TSS ) and each SNP . QTL tools identified 28 , 906 significant associations with 1 , 916 preserved after permutations , which was further reduced to 985 eQTLs after exclusion of associations that failed the quality threshold ( Methods and Fig 1 , S3 Fig ) . A complete list is of the 985 eQTLs used for further analysis is provided in S5 Table . For mQTL analysis , we focused on 150 kb windows ( 75 kb upstream and downstream of the CpG site ) as it has been previously published that mQTLs correlate best with SNPs within 1–2 kb and that most mQTLs are located within 100 kb in several tissues including placenta [13 , 32 , 33] . 471 , 852 significant associations were called using our pipeline , only 105 , 031 associations pass the permutation test with 4 , 342 mQTLs conserved for further analysis ( Fig 1 , S3 Fig ) after stringent exclusions based on quality threshold ( see Methods ) and potential probe artifacts . The 4 , 342 mQTL associations kept for further analysis are recapitulated in S6 Table . Some genetic variants that appear as individual differences in methylation can act trivially via SNPs that create or abolish CpGs ( CpG-SNPs ) [34] or via technical artifacts due to incomplete probe binding [34] or probe-cross reactivity [35] referred to here as “at risk-probe” . Incomplete probe binding may suggest the presence of a genetic variant within the probe binding site with greater impact in the proximity of the 3' end of the binding probe with differences in signal intensity persisting for up to an approximately 20 bp [36] . Therefore , to make our lists more stringent , we identified and excluded associations involving probes that contain known SNPs , along with probes previously identified as cross-reactive . A list of excluded probes is in S7 Table . As we illustrate by independent validations below , this high level of stringency is expected to reduce false-positives . Additionally , it can simultaneously give rise to false-negative findings , particularly for loci queried by probes that contain internal SNPs with modest minor allele frequencies and/or those without strong effects on probe binding , which can yield accurate readouts in most samples . By using a reference-based approach , identification of probes sensitive to CpG-SNPs may be subject to tissue and sample specificity . Recently , a methodology has emerged to identify “at risk” probes independently of reference datasets , based on a clustered distribution of methylation scores [36] . Using this approach , the authors were able to identify a significant number of “gap probes” that largely overlap with CpG-SNP probes . However , as SNP-affected probes can exist without gap signatures , the implementation of this reference free approach should be considered in complement with the reference-based approach . To be as inclusive as possible , we applied the “gap-hunter” [36] approach to our datasets to identify remaining probes and associations subject to possible technical artifacts . By doing so , we were able to detect 10 , 971 “gap probes” . From these , 3 , 235 are overlapping with previously identified CpG-SNPs probes and 205 with cross-reactive probes . Finally , 432 of the remaining “gap probes” overlap with our 4 , 342 mQTL associations and 158 overlap with our eQTM associations . These probes were not excluded from further analysis but were annotated following the recommendation of Andrews et al . [36] ( S6 and S7 Tables ) . Interestingly , after FDR correction , about 22% of the mQTL ( 23 , 043 out of 105 , 031 ) involved an association between an at-risk probe and genetic variant . We are aware that this proportion reflects in part the stringency of our criteria for exclusion . However , these findings are strong enough to encourage for stringent criteria when running genome-wide analysis . To take into consideration false-negative results , it is particularly attractive to use annotation instead of direct exclusion , as we have used previously [13] , and in part adopted here . To assess the robustness of our approach we were first interested in looking into the overlap between our newly identified associations and published data on the human placenta . Because identification of the causal genetic variant is still challenging due in part to linkage disequilibrium , overlaps based on genetic variants are likely to not be as informative . Therefore , we decided to look at overlap driven by gene or CpG for eQTL and mQTL associations respectively . A recent study published by Peng et al . [23] has identified 3 , 218 eQTLs associations based on 159 human placenta tissues . These data represent a unique opportunity to assess data reproducibility by examining the number of associations preserved across different cohorts . Our 985 eQTL associations represent 615 unique genes and 62% of them overlap with genes previously identified by Peng et al . ( S8 Table ) . As noted above , we ( CD , BT ) previously reported a list of mQTLs in placentas from a smaller group of samples [13] , independent of the current series . Using more lenient criteria for probe exclusions , we identified 866 mQTLs ( n = 665 with similar exclusion criteria ) . Among these mQTLs , 319 ( ~48% ) were also found in the current series , with many of the non-overlapping “hits” reflecting probes on the 450K methylation arrays that were excluded from the current study ( S9 Table ) . The significance of these overlaps was further confirmed using a permutation test ( n = 1 , 000 ) . The significance of the enrichment was defined by the overlap between observed versus expected distribution . Here , the null distribution represents a random sampling from the total number of annotated Refseq genes or CpGs interrogated by the Illumina 450K platform . Random permutations failed to replicate such overlap with a maximum of 17% overlapping genes and 2% overlapping CpGs . Therefore , we feel confident in the strength of our study and believe that these newly identified associations represent a solid foundation for further exploration . After demonstrating a significant overlap for placenta in different cohorts , we were interested in evaluating the overlap across tissue types . The GTEx Project , assessing and cataloging eQTLs across more than 40 tissues , is a valuable resource for studying human gene expression regulation and its relationship to genetic variation across tissue types . However , among the list of tissues in the GTEx project , the placenta has so far been missing . Therefore , our current study represents a key opportunity to expand the GTEx exploratory work toward this crucial fetal organ . To evaluate to what extent the placenta eQTLs replicated the significant SNP-gene pair eQTLs found in the GTEx study ( at FDR<5% ) we used the π1 statistic . The π1 statistic can be interpreted as the fraction of eQTL associations shared between the placenta and the other tissues available through GTEx . This approach allowed us to identify placenta-specific eQTLs , and identify tissues with similarity to the placenta in this respect . Levels of significance were collected across the different tissues for our list of significant eQTL association from the placenta . π1 value range from 0 . 31 for the brain cerebellar hemisphere tissue to 0 . 69 for the transformed fibroblast cells ( S4 Fig , S10 Table ) . It is interesting to note the absence of interdependence between gene expression correlation and number of conserved eQTLs . Indeed , no significant overlap was found ( Fisher’s exact test , p . value = 0 . 5935 ) when overlapping the most similar tissues to the placenta ( n = 10 ) based on π1 value ( S10 Table ) with the most similar tissues to the placenta ( n = 10 ) based on gene expression correlation value ( S1 Table ) . Among the 985 eQTL associations identified in placenta , 63 appear to be specific to placenta as not found significant in any other GTEx tissues ( S11 Table ) . When looking at the genes associated with these 63 eQTLs we were not able to identify enriched pathways . Unfortunately , similar curated reference doesn’t exist for mQTL associations with only a limited number of tissue types studied ( brain , blood , lymphoblastoid cells , T-cells and lung [37–42] ) . Contrary of the GTEx analytical design , there was no consensus on how associations were defined increasing variability across studies , a likely confounder when assessing the biological relevance of the overlap . In this study , we used a conservative approach to minimize false positive by excluding probes with potential cross-reactivity or mapping of common SNPs within 20 bp of the queried CpG , which might affect the probe hybridization . This stringent approach will reduce false-positives , but it can also lead to false-negative findings in which bona fide mQTLs are discarded . To investigate this aspect directly , we applied targeted bisulfite sequencing ( bis-seq ) to assess allele-specific CpG methylation ( ASM ) , which is the physical counterpart of mQTLs . For this purpose , we chose three loci that passed our current stringent criteria for inclusion of the 450K probes and showed mQTLs ( ranking from 82 to 1900 in strength , see S6 and S7 Tables ) , and four loci for which the probes were excluded by the current criteria but were included by our more lenient criteria and showed mQTLs in our prior study of the smaller placental series [13] . While the net methylation at the index CpGs showed more inter-individual variability in the bis-seq data , the average net methylation matched that observed in 450K arrays ( +/- 5% ) suggesting no or mild hybridization issues with the four excluded Illumina probes . Further , statistically significant ASM , affecting from 22% to 45% of the heterozygous samples , was found at the amplicon level for all 7 loci ( S12 Table and Fig 2 , S5 and S6 Figs ) , thus validating these loci as genuine mQTLs . These data indicate that our stringent criteria for excluding Illumina probes , while minimizing false-positives , can also lead to false-negative findings from exclusion of probes that are , in fact , legitimately informative . Based on these findings , we conclude that an optimal approach for harvesting and interpreting mQTLs using Illumina arrays is to annotate each probe as meeting either stringent or lenient criteria for predicted reliability , and include this information as a descriptor in comprehensive lists of mQTLs . Here we adopt this approach for reporting mQTLs ( S6 and S13 Tables ) , but for our bioinformatic enrichment and pathway analyses , we have used only the mQTLs that pass our stringent probe selection criteria . Characterizing gene expression and DNA methylation patterns that are influenced by genetic variants can potentially provide functional information about disease-causing genes and pathways . Therefore , we were interested in better characterizing genes ( n = 615 ) and CpG sites ( n = 4 , 324 ) associated with eQTLs and mQTLs . First , we aimed to assess correlations between level of expression or level of DNA methylation with genetic variants . Genes were first assigned to different categories of expression from not expressed to highly expressed using quartiles based on our placental RNA-seq dataset . A similar approach was applied for DNA methylation levels . Enrichments for specific quartiles of expression or DNA methylation were assessed by permutation as described in Methods . The significance of the enrichment was defined by the overlap between observed versus expected distribution . In these cases , null distribution represents a random sampling from the total number of annotated Refseq genes or CpGs interrogated by the Illumina 450K platform . eQTL associations were found significantly enriched for genes with intermediate and high level of expression ( Fig 3 ) . Such enrichment has been previously reported in other tissues than placenta by the GTEx consortium [5] . By comparing the profile of expression between genes associated and genes non-associated to eQTLs , they showed a shift toward high expression for associated genes across tissue types . mQTL associations were enriched for CpGs with intermediate levels of DNA methylation ( Fig 3 ) . This pattern is likely to reflect on the placenta specific methylation profile with enrichment for low and intermediate DNA methylation levels . This can also explain the limited range of variation in DNA methylation across the different genotypes found in our selected mQTLs ( with a regression slope <5% for 87% of the association ) . Due to the nature of the DNA methylation signal , intermediate levels of methylation are likely to reflect cellular heterogeneity [27] . Therefore , it is possible that mQTLs associations help to identify genetic variant impacting cell differentiation mechanisms , expanding downstream functional consequences associated with mQTLs . To lay a groundwork for future studies on disease susceptibility , identification of biological pathways associated with eQTLs and mQTLs are of interest . In this context , it is valuable to identify pathways involving genes with greater sensitivity to genetic background . Therefore , we looked at gene set enrichment for genes associated with an eQTL and the nearest gene from CpG sites associated with a mQTL using the Bioconductor package GOseq [29] . For both eQTLs and mQTLs , we found enrichment for inflammation related pathways ( S14 Table ) suggesting that the variability captured by our associations may , in fact , reflect different degrees of inflammatory response across our collected samples . It is also interesting to note that the enrichment analysis outcomes ( S14 Table ) rely mainly on genes part of the major histocompatibility complex ( HLA-DRB1 , HLA-DBR5 , HLA-DQA1 , HLA-DQB1 , HLA-C ) . These genes are known to be highly polymorphic with 13 , 840 different HLA alleles reported in the IMGT/HLA database [43] suggesting polymorphism mechanism as a possible bias in QTL and GWAS analysis . Finally , we decided to look into the overlap between the gene associated with eQTLs and gene associated with mQTLs where CpG was associated to the closest gene . We found 111 genes overlapping between the 2 associations ( S15 Table ) with pathway enrichment only relying on HLA related genes . Genomic context is known to play a cell type-specific role in regulating transcription . To determine whether genetic variation was occurring at regulatory sites , we took advantage of public chromatin mapping data for placenta generated by the Roadmap Epigenomics Program ( S16 Table ) . The DNase hypersensitivity and ChIP-seq data create combinatorial patterns that can help to define functional elements in the genome ( S7 Fig ) [44] . However , because placenta specific reference datasets are limited and do not include data from term placenta , we were compelled to build our genomic annotation based on data from the pre-term placenta which may affect the accuracy of our annotation especially in a highly dynamic tissue like placenta . For relative enrichment studies , the observed overlap between genetic variants and a genomic feature was compared with the expected overlap given the total coverage of the annotation . For permutation tests , we compared the observed overlap of SNPs significantly associated with eQTLs ( n = 608 ) or mQTLs ( n = 3 , 022 ) and a particular genomic interval ( for example gene promoters , CpG islands ) or feature from ChromHMM analysis with the distribution of the same overlap under the null hypothesis . SNPs associated with eQTLs and mQTLs were enriched in candidate promoters ( Feature 1 ) , candidate active and poised enhancers ( Feature 2–4 ) and enriched at regions likely to be transcribed ( Feature 7 ) ( Fig 4A and 4B ) . When looking at the distribution of the distance between SNP and gene for eQTL association ( n = 985 ) or SNP and CpG for mQTL association ( n = 4 , 342 ) , we found enrichment for closest interactions ( within 10Kb , S8 Fig ) suggesting that the associated genetic variant has an effect via its presence in proximal cis-regulatory elements . We then explored the genomic context for CpGs ( n = 4 , 342 ) associated with mQTLs and found , as in our previous study of multiple human tissues [13] , that they are enriched in candidate active and poised enhancers ( Feature 2 and 3 ) ( Fig 3C ) . This finding suggests a possible cis-regulatory impact of mQTLs on gene expression . Using combined genetic and epigenetic profiling applied to other human tissues , we and others have found evidence that some of the epigenetic effects of genetic variation are mediated by SNPs in insulator elements and transcription factor binding sites [13 , 45–47] . Therefore , we decided to analyze the impact of genetic variation on the integrity of the binding motif in correlation with predicted binding affinities . Due to linkage disequilibrium , the index SNP may not be causal , so we decided to also investigate the impact of previously identified SNPs in proximity to the eQTL and mQTL associated gene TSS or CpG respectively ( regardless of their linkage disequilibrium status ) . This approach allows us to expand our discovery to transcription factors likely to be affected by linked genetic variants . We compared the predicted binding affinities between the reference and alternative alleles within a window of 39 bp around each targeted genetic variant using the motif-based sequence analysis tools FIMO from the MEME suite [48] . To assure the accessibility of the binding site , only those overlapping with DNAse hypersensitive regions ( ENCODE data , release 3 ) were considered for further analysis . The analysis was performed at the motif level . The list of altered binding sites was further pruned down based on 2 criteria ( 1 ) the significance of the binding with the reference allele ( q . value<0 . 05 ) and ( 2 ) the difference in binding affinity between the reference and alternative allele . Distributions of the difference of binding affinity between the reference and alternative sequence were used to define the cutoff for predicted allele-specific binding ( difference in binding >8 , S9 Fig ) . The significance of the alteration from our filter list of genetic variants and transcription factor binding motif was assessed using atSNP [49] . atSNP performs computations for between-allele score differences [50] and outputs p . values for each targeted SNP and selected transcription factor binding motif . The complete list of affected transcription factor binding sites is in S17 Table . Interestingly , when looking at transcription factor binding sites associated with mQTLs a significant number have been previously associated with alterations of the epigenetic landscape ( CTCF , REST , HDAC2 , EWSR1 , FLI1 , SP4/1 and BHLHE40 ) . CTCF ( CCCTC-binding factor ) previously identified as associated with mQTLs and haplotype-dependent allele-specific methylation ( hap-ASM ) by Do et al . [13] is one of the best documented transcription factors affecting DNA methylation . CTCF , which acts to anchor high-order chromatin loops and binds at some of its recognition sites in a CpG methylation-dependent manner , is known to have an essential role in imprinting control achieving allele-specific gene regulation [51] and CTCF binding locally influences DNA methylation [52] . REST ( REI-silencing transcription factor ) inactivation has been shown using a transgenic approach to induce de novo methylation and conversely , the expression of REST was sufficient to restore the unmethylated state of the DNA sequence bound [52] . Direct association between HDAC2 ( Histone deacetylase 2 ) , EWSR1 ( Ewing sarcoma breakpoint region1 ) , FLI1 , SP4/1 and BHLHE40 ( Basic helix-loop-helix family member E40 ) and changes in DNA methylation have not yet been demonstrated but they have been previously associated with changes in histone methylation or acetylation [53–57] . Numerous evidence supports a complex interplay between histone modification and DNA methylation ( for review [58 , 59] ) . However , if in numerous contexts a bidirectional association between changes in DNA methylation and histone modifications has been reported , the exact mechanism involved still needs to be identified . Interestingly , these associations seemed to not only rely on direct interaction between DNMT enzymes and histone methyltransferases but also on the recruitment of a variety of intermediate factors such as member of the SET domain family ( G9a ) or member of the Polycomb complex ( PRC1 ) suggesting that a large number of histone modification factors could indirectly affect DNA methylation . Among the different transcription factor binding sites linked to eQTLs , we also found EWSR1 , FLI1 , and SP4/1 , suggesting that the effect of transcription factors on gene expression may be in part mediated via epigenetic modifications . This list of transcription factors provides candidate mechanisms to explain the association between genetic variants and gene expression or DNA methylation . The GWAS approach identifies disease-related enrichment for specific genetic variants . This approach has considerably improved marker discovery but , because of linkage disequilibrium among many SNPs in each chromosomal region , it has a limited ability to pinpoint disease genes and regulatory sequence variants . “Post-GWAS” approaches that integrate eQTLs and mQTLs to GWAS datasets can help to overcome these challenges [12] . For our current analysis , we overlapped the genetic variants listed in the GWAS catalog [60] with those associated with eQTLs ( n = 608 ) and mQTLs ( n = 3 , 022 ) . Only 7 genetic variants associated with the placental eQTLs were previously found in GWAS peaks . Of these , three validated the previously reported association between the genetic variant and gene in the GWAS database ( S18 Table ) . Two mQTLs shared a genetic variant with statistical peaks in the GWAS database ( S19 Table ) . The GWAS database references a limited number of studies involving trait directly associated with placenta function or dysfunction . We identified only 2 studies related to preeclampsia and 6 related to birth weight and none of them appears in the overlaps with eQTLs and mQTLs . Nonetheless , among these hits , some are found associated with traits that have been previously linked to fetal growth , like obesity and type II diabetes where the placenta is likely to play a role . rs35694355 overlaps with our eQTLs and have been previously associated with obesity . Interestingly , this genetic variant is associated with CRACR2B gene in the GWAS database but with IRF7 in our eQTL study . Contrary to CRACR2B , evidence for a direct association between IRF7 and obesity have been shown [61] with IRF7 knockout mouse being protected from gain weight after high-fat diet exposure . IRF7 is also highly expressed in placenta and it though to play a role in the immune barrier between the fetus and the mother [62] . Due to the tight correlation between inflammatory response and obesity , IRF7 may appear as a better candidate to link rs35694355 with disease sensitivity . From our mQTL analysis , rs623323 has been previously associated with Type 2 diabetes in the GWAS database . In both datasets , rs623323 was associated with NXN gene . Interestingly in a recent publication , increase of DNA methylation and decrease of expression for NXN has been characterized in placenta from women with diabetes during pregnancy [63] suggesting that the presence of rs623323 may influence susceptibility to type II diabetes via alteration of NXN function thus providing with a great example of how integrative QTL analysis can provide with evidence to better understand disease susceptibility . Being able to establish transcript-CpG specific associations ( eQTMs ) based on gene expression and CpG DNA methylation at a genome-wide scale in a tissue specific manner will significantly improve further interpretations of epigenetic modifications when considering their functional implications . Using our transcriptomic and epigenomic data , we decided to identify CpG methylation and gene expression with a significant correlation . To do so , we applied a similar approach to the one used for eQTL analysis . Looking at different combinations of CpG sites and transcripts with linear regression in a 100 kb window ( matrixEQTL ) , we were able to identify 55 , 492 eQTM associations , representing 47 , 140 unique CpG sites and 14 , 352 unique genes . After permutation and quality control , 2 , 655 eQTM associations remained with 2 , 538 unique CpG sites and 1 , 269 genes ( S20 Table ) . 515 associations were located in the sex chromosomes . To make the analysis less computationally intensive , we decided to sample gene expression profiles instead of CpG methylation profiles during permutation . A genomic context-dependent correlation between DNA methylation and expression has previously been shown [64] . Negative correlation , when an increase in DNA methylation is associated with a decrease in expression , was found for CpG sites located in the promoter region . A positive correlation , when an increase in DNA methylation is associated with an increase of expression , was found for CpG sites located in the gene body . We first assessed if such profiles were also found among our associations . Both negative and positive associations were enriched for proximal interaction ( Fig 5A ) . We then examined enrichment between our newly annotated features specific to placenta and our eQTM candidates stratified based on positive ( n = 1 , 009 ) and negative ( n = 1 , 646 ) correlations . We found the negative association to be enriched for feature 1 ( candidate promoter ) and 2 ( candidate active enhancer ) and associations with positive correlation to be enriched for feature 5 ( Fig 5B ) . Unfortunately , feature 5 does not show a clear significant enrichment pattern for the interrogated histone marks ( S7 Fig ) . However , looking at the distribution over Refseq annotations , we confirm the enrichment for gene body region . This validates the inversely correlated relationship between DNA methylation and gene expression in promoter regions as well as the positive correlation when targeting the gene body region . To identify potential mechanisms underlying the correlation between DNA methylation and expression , we assessed the role of transcription factors , looking at enrichment for specific binding sites overlapping the CpG site from the eQTM associations ( n = 2 , 655 ) using a similar approach as described above . The null distribution represents the sampling of CpGs ( n = 1 , 000 ) from the population of all assayed CpGs overlapping with a specific TFBS . Associations with negative correlation were found to be globally enriched for transcription factor binding sites suggesting a general mechanism where increase DNA methylation at a given binding site will alter binding abilities and affect transcription . Interestingly , when looking at associations that showed a positive correlation between expression and DNA methylation , we found enrichment for only one transcription factor , ZNF217 ( S21 Table ) . The ZNF217 transcription factor has been previously recognized as a human oncogene and is known to be part of a complex that contains several histone-modifying enzymes strongly associated with gene repression [13] . Thus , alteration of the binding of this transcription factor is predicted to disturb the formation of a gene repression complex and lead to an increase of expression for the targeted gene by diminishing its inhibition . There is also evidence for a more distal action of ZNF217 with only 2% of the binding site within a 1 kb distance from TSS [13] . When looking at the pattern for correlation between DNA methylation and expression , we identified two distinct profiles . We found eQTM associations with a linear distribution between methylation and expression representing a range of continuous values for both DNA methylation and gene expression and , at other loci , associations with a bimodal distribution with no intermediate values for both DNA methylation and gene expression ( Fig 6 ) . Because DNA methylation is a binary variable as mentioned before , having eQTM associations with bimodal distribution suggest that every interrogated cell within the placenta sample share the same DNA methylation and expression profiles , implying that the association is non cell- or state-specific . On the contrary , eQTM associations with linear distribution suggest a cell- or state-specific mechanism where the continuous signal is explained by the distribution of cells being fully or not methylated among the placenta sample ( S10 Fig ) . To assign eQTMs to linear or bimodal distribution we implemented a Bayesian model for bimodal distribution detection . Each gene expression or DNA methylation distribution was analyzed separately and eQTMs were called as linear if both expression and methylation represented a linear distribution and called bimodal if both expression and methylation show a bimodal distribution profile ( see Methods ) . Using our model , we classified 118 associations as bimodal including 29 unique genes and 108 unique CpG sites , and 1 , 201 as linear which represent 877 unique genes and 1 , 140 unique CpG sites ( Fig 6 , S22 Table ) . The greater occurrence of linear correlations can be attributed to the high cellular heterogeneity of the placenta . Both linear and bimodal associations were enriched for proximal interactions , with linear showing a more spread distribution ( Fig 7A ) . Interestingly , linear and bimodal distributed associations were not found enriched for the same genomic context , linear distributed eQTMs were enriched for feature 2 and 3 , our candidate active and poised enhancer and bimodally distributed eQTMs where enriched for feature 1 our candidate promoter ( Fig 7B ) . Looking at transcription factor binding sites across bimodal and linear eQTM associations , we identified a specific pattern where bimodal but not linear associations were globally enriched in transcription factor binding sites ( S21 Table ) . Interestingly , the transcription factor previously identified as part of the bHLH family ( MAX , MAZ , MXI1 , MYC , and SIN3aK20 ) where highly enriched for bimodally distributed associations while depleted for linear associations ( S11 Fig ) . This suggests that distribution not only reflect cell heterogeneity but also provide information on the different factors involved in DNA methylation regulation of expression . These observations , if not sufficient to propose a definitive mechanism of action , will definitely contribute as key resources to further analysis when linking DNA methylation alteration to phenotypes by offering a curated list of gene sensitive to methylation alterations specific to the placenta .
The results presented here highlight the importance of genetic-epigenetic interactions in human placentas and provide a resource to further characterize the dynamic of these interactions in the context of human diseases . Indeed , because GWAS and EWAS analyses are statistical approaches that only report associations between genetic variants and interrogated phenotypes , an approach in which GWAS data are combined with reference datasets detailing associations between genetic variants and gene expression or CpG methylation can help to pinpoint genes and regulatory DNA sequence variants linked to disease susceptibility . By providing a curated list of eQTL , mQTL , and eQTM associations , with careful consideration of potential false-positive and false-negative findings , we have started to address these issues as they pertain to human placentas . Our data represent an initial basis for understanding how genetic variation in human placentas influences epigenetics marks and expression and allows the identification of genes and CpG sites with greater sensitivity to genetic variants , which should be considered when interpreting disease-focused studies such as GWAS and EWAS . Indeed , a significant amount of research today is built on the dichotomy between treated versus non-treated or disease versus non-diseased with less consideration toward understanding what constitutes normal or healthy . Thus , our ability to more fully define a ‘normal’ molecular phenotype will have a notable impact on the interpretation and significance of such studies . Lastly , a better understanding of the interplay between the multiple functional layers at a systems level under normal conditions is needed to fully benefit from the ever-increasing amount of multi–omic data . Epigenome-wide association studies , primarily focusing on DNA methylation , have been used for examining the impact of environmental exposures and early biomarker discovery [13] . The increasing number of studies and challenges in this area have been reviewed [65] . Among these challenges is the ability to adequately correlate DNA methylation and expression profiles to identify reliable and causal associations in a tissue-specific context . Here , we provide a list of candidate genes and CpG sites with high genetic-epigenetic correlations in human placentas , revealing some general principles of these interactions . Associations found in promoter regions represent a categorical/bimodal signal with enrichment for transcription factor binding sites , while enhancer interactions represent a linear additive profile with no significant enrichment for transcription factors . These differences can be related to cell heterogeneity . Indeed , promoter regions show a more constitutive pattern across cell types while enhancers are more likely to be cell type specific . Therefore , enhancer regions will present a cell specific signal depicted as a continuous range of values for gene expression and DNA methylation which transcribe as a linear distribution when looking at methylation-expression associations . Cell subpopulation effects are now recognized as a major source of variability and support the recommendation for single cell type analysis in EWAS [4] . The placenta is a highly heterogeneous tissue and demonstrates differences in gene expression and DNA methylation , which may be due to changes in the proportional distribution of different cell types ( for example , as a result of infection , inflammation or other conditions ) that can later confound methylation-disease associations [66] . Heterogeneity may be corrected using statistical deconvolution techniques [67 , 68] where a subset of targets is used to represent distinct DNA methylation or gene expression profiles for each cell type and assess subpopulation distribution . The accuracy of these approaches is highly dependent on the availability of reference epigenomes and transcriptomes and the ability to select representative targets for the tissue considered . Single cell data for the placenta are not available . However , genes and CpG sites with linear distribution can be further used to assess and control for cell subtype population in human placenta . There are several strengths to our study , the most important of which is the comprehensive nature of our analysis . By sampling more than 300 individual placentas from an ethnically diverse and representative population , we present here the most extensive placenta-specific genome-wide analysis published to date . Based on our heterogeneous population we believe that our findings will be applicable to populations with diverse backgrounds including minorities that have been so far underrepresented in both GWAS and EWAS studies . Another key factor of our study is the resolution of our approach . We combine from the same cohort genome-wide transcriptome , epigenome and genetic variation information which allows for improved understanding of the intricacy of expression regulation mechanisms in an understudied tissue by overcoming the limitations inherent to each individual dataset when considered separately . Finally , the ability to extract meaningful information is highly dependent on the quality of the analytical approach . Thus , we describe here stringent methods to incorporate different datasets and extract robust correlations . We also emphasize the limitations intrinsic to the different techniques , which will help to assure reproducibility . A major gap in the field has been the lack of placenta-specific annotations . This situation also makes our findings valuable as they will promote the creation of a placenta-specific data repository . We believe that the massive amount of high-quality data generated will serve as a reference and benefit to further–omics placenta specific studies and contribute to a new perspective in term of disease susceptibility and placenta development .
This study has been approved by the Columbia University IRB Exp . The protocol number is IRB-AAAD9014 . Written informed consent were obtained for all women at time of the enrollment . Subjects and samples were identified from the Eunice Kennedy Shriver National Institute of Child Health and Human Development ( NICHD ) Fetal Growth Study , which was a prospective , multicentered , observational study including 3 , 000 pregnant women [69] . Healthy , non-obese , low risk pregnant women across four race/ethnicity groups , who conceived spontaneously and had no obvious risk factors for fetal growth restriction or overgrowth were eligible for inclusion in the study . Specifically , self-identified non-Hispanic white , African-American , Hispanic , and Asian/Pacific Islander women with a singleton pregnancy less than 13 weeks and 6 days of gestation were enrolled . All women were between 18 and 40 years old with BMI between 19 . 0 to 29 . 9kg/m2 with no confirmed or suspected fetal congenital structural or chromosomal anomalies . Multiple exclusions were identified to assure a low risk population and included but was not limited to , cigarette smoking in the past six months , use of illicit drugs in the past year , consumption of at least 1 alcoholic drink per day , chronic hypertension , diabetes mellitus , HIV or AIDS , and history of gestational diabetes in a prior pregnancy . To assure correct dating , all pregnancies had first trimester ultrasound screening consistent with gestational age . Placentas were collected from pregnancies with a predicted birthweight by ultrasound between the 10th and the 90th percentile ( S23 Table ) . Placental samples were obtained and processed within one hour of delivery by trained research personnel . Placental parenchymal biopsies measuring 0 . 5 cm x 0 . 5 cm x 0 . 5 cm were taken from the fetal side of the placenta just beneath the fetal membranes and were placed in RNALater and frozen for molecular analysis . After delivery , neonatal anthropometric measurements were taken . These included: birth weight , head circumference . Information on maternal weight gain , antenatal and intrapartum complications , and neonatal outcomes were extracted from the medical record . RNA from 80 placenta biopsies ( 42 males and 38 females ) was isolated using TRIZOL reagent ( Invitrogen , MA , USA ) . Poly-A pull-down was used to enrich for mRNAs , and libraries were prepared using the Illumina TruSeq RNA kit . Libraries were pooled and sequenced on an Illumina HiSeq2000 machine with 100 bp paired-end reads . RTA ( Illumina , San Diego , CA , USA ) was used for base calling and bcl2fastq ( version 1 . 8 . 4 ) for converting BCL to FASTQ format , coupled with adaptor trimming . The reads were mapped to the human reference genome ( NCBI/build37 . 2 ) using Tophat ( version 2 . 0 . 4 ) with 4 mismatches ( —read-mismatches = 4 ) and 10 maximum multiple hits ( —max-multihits = 10 ) . The relative expression level of genes was estimated by FPKM ( Fragments Per Kilobase of transcript per Million mapped reads ) using cufflinks ( version 2 . 0 . 2 ) with default settings . FPKM values were used in log2-transformed scale after quantile normalization . Data are available through dbGap repository , accession number: phs001717 . v1 . p1 . Genomic DNA from 303 placental biopsies ( 151 males and 152 females ) , 500 ng , was used as per the manufacturer’s instructions for HumanMethylation 450 Beadchips ( Illumina ) , with all assays performed at the Roswell Park Cancer Institute ( RPCI ) Genomics Shared Resource . Data were processed using Genome Studio , which calculates the fractional methylation ( AVG_Beta ) at each queried CpG , after background correction , normalization to internal control probes , and quantile normalization . All probes mapping to the X or Y chromosome were removed . As recommended by Illumina , AVG_Beta values with a detection p-value>0 . 05 were excluded from the analysis and replaced by missing values . Probes that queried CpGs directly overlapping the positions of known common DNA variants as reported in 1000 Genomes Project Phase 3 ( allele frequency >1% ) and within a 20 base pair window upstream of the interrogated CpG were excluded since these CpGs may be abolished by the SNP itself or the binding of the probe itself could be altered [34] . Additionally , the accuracy of the Illumina 450K can be affected by cross-reactive probes that were also excluded from our analysis [35] . Data are available through dbGap repository , accession number: phs001717 . v1 . p1 . The DNA samples ( n = 303 , 151 males , and 152 females ) were genotyped on Illumina HumanOmni2 . 5 Beadchips , followed by initial data processing using Genome Studio . SNPs were annotated using dbSNP138 . Quality control exclusionary measures for subjects were: genotype call rates <95% , marked departure from Hardy-Weinberg equilibrium ( p<0 . 001 ) and low minor allele frequencies <5% after QC filtering genotype was conserved for 1 , 374 , 581 SNPs . Data are available through dbGap repository , accession number: phs001717 . v1 . p1 . To test for association between genotype and expression ( each SNP vs each transcript , eQTL ) or genotype and DNA methylation ( each SNP vs each methylation site , mQTL ) we used a linear regression model implemented in QTLtools [70] . For the association between expression and DNA methylation ( each transcript vs each methylation site , eQTM ) we used a linear regression model ( LINEAR ) implemented in Matrix eQTL [71] . For eQTL and eQTM , the linear regression model was corrected using principal component analysis ( PCA ) and PEER normalization ( K = 10 ) based on expression data [72] . For PCA analysis , only the principal components ( PC ) significantly contributing to variability were added into the model . In our case , PC1 and PC2 were included as contributing for up to 32% of variability ( S2 Fig ) . For PEER analysis , covariate contribution was estimated using Bayesian approaches to infer hidden determinants and their effects from gene expression profiles by using factor analysis methods [72] . For each association , FDR was estimated by permutations a 1 , 000 times after correction for multiple comparisons using Benjamini & Hochberg [73] approach implemented in R package ( p . adjust ) . The same random indexes were applied to the PEER factors and covariates . For eQTLs and mQTLs , permutation is implemented in the QTLtools package . For expression-methylation associations ( eQTMs ) , we permuted ratio for each transcript with significant association ( n = 14 , 353 ) 1 , 000 times . For eQTL analysis , the correlation was run between expression and genotype across 80 samples ( 42 male , 38 female ) , for mQTL analysis association between methylation and genotype was calculated across 303 samples ( 151 male , 152 female ) and for eQTM analysis through 74 samples ( 38 male , 36 female ) . Associations reaching the significance threshold were retained to further analysis if having at least a shared genotype in 5% or more individuals ( AA , AB , BB , where A is the reference and B the alternate allele ) and regression slope >5% . Associations were further prioritized based on their regression slope and distance between a genetic variant and associated gene or CpG to encompass the linearity and the magnitude of the progression between the different genotype . Briefly , we generated a score for each association following this equation [score = ( 1-P . value ) *abs ( regression slope ) /log ( abs ( distance ) +2 ) ] . “At-risk” probes were defined following 3 distinct criteria ( 1 ) the presence of common SNPs in a 20 bp window from the 3’ end of Illumina probe ( SNP-probe ) [34] , ( 2 ) previous identification as cross-reactive probes [35] and ( 3 ) based on a specific methylation profile using Gap-Hunter algorithm ( gap-probe ) [36] . Criteria 1 and 2 were sufficient to exclude the probe from further analysis while criteria 3 only resulted in a flagging of the identified probes . We first identified all CpG-SNPs interrogated by the Illumina 450k assay based on 1000 Genomes Project Phase 3 dataset using an allele frequency >1% . We then used the same reference to identify probe with known genetic variants within the 20 bp upstream of the interrogated CpGs including the single base extension for probe of type I . We decided to consider 20 bp because of the absence of consensus for exclusion and because we did not detect a differential enrichment from 20 bp to the 5’ end of the probe ( S12 Fig ) . This list of affected probes was further pruned down by excluding probes containing a genetic variant with reference and alternative alleles involving a C or a T which will not affect the binding of the probe as we are considering bisulfite-treated reads . The list of cross-reactive probes is publicly available [35] . We used the function “gaphunter” from the minfi package ( v1 . 18 . 4 ) to identify probes with a gap in a beta signal . Such probe signal has a tendency to be driven by an underlying SNP or other genetic variants . Beta values were provided as input and the function outcomes a data frame listing for each identified gap signal , the number of groups and the size of each group . Only gap signals not driven by outlier ( cutoff = 1% ) were considered . A complete list of SNP-probe , cross-reactive probe as well as gap-probe is available as S7 Table . S7 Table also includes the list of association involving a gap-probes . Finally , the non filtered list of mQTLs with annotated probes ( SNP/cross-reactive/Gap-probes ) is available at S13 Table . For an observed overlap X between our candidates and a list of intervals of interest Y of size n from a total population of assayed sites Z , we compared the observed value of X with the distribution of the same overlap n under the null hypothesis using permutation tests . To obtain the null distribution , we employed a random sampling approach where we randomly sampled 1 , 000 times from the population of all assayed sites ( Z ) . Thus , with Y , n , and an observed overlap with a genomic interval , X , we sampled n loci from Z and found the overlap of the random sample with the interval , Yk , for k in 1 , 2 , … , 1000 ( that is 1000 random samples ) . Next , we compared X with the distribution of simulated overlaps , Y1 , 2 , … , 1000 . If the resulting null distribution , Y1 , 2 , … , 1000 , contains the observed overlap , X , then we can conclude that there is no significant enrichment . Conversely , when the null distribution , Y1 , 2 , … , 1000 , excludes the observed overlap , X , then we can conclude that there is significant enrichment beyond that of random chance . The significance is further assigned as follow: the number of time when Yk > X divided by k . The π1 statistic was used to quantify overlap between eQTL associations found in placenta and the tissue cataloged by the GTEx consortium . The non–model-based pairwise analysis method identifies significant SNP-gene pairs in a first tissue , and then uses the distribution of the P values for these pairs in the second tissue to estimate π1 , the proportion of non-null associations in the second tissue . It provides an estimate of the fraction of true positive eQTLs , π1 = 1-π0 , where π0 = estimated fraction of null eQTLs , estimated from the full distribution of p-values ( Storey and Tibshirani q-value approach [74] ) . The π1 statistic considers also sub-threshold placenta eQTL p-values below the FDR<5% cutoff . Using the 1000 Genomes Project Phase 3 dataset , we identified all genetic variants within 2 kb windows centered on the gene transcription start site ( eQTL ) or CpG ( mQTL ) using a minor allele frequency >1% . The transcription factor dataset was from ENCODE ChIP-seq experiments , together with DNA bindings motifs identified within these regions as displayed by the ENCODE Factorbook repository [75] . We first compare transcription factor binding affinity between the reference and alternative allele using the FIMO tools in the MEME suite [48] . FIMO is a software tool for scanning DNA or protein sequences with motifs described as position-specific scoring matrices . The program computes a log-likelihood ratio score for each position in a given sequence database , uses established dynamic programming methods to convert this score to a P-value and then applies false discovery rate analysis to estimate a q-value for each position in the given sequence . FIMO output a list of significant binding site for both reference and alternative allele . However , FIMO only provides with qualitative information and does not allow for significant quantification of the difference in binding affinity between the two alleles . Therefore , we used atSNP [49] to quantify the difference between the reference and alternative allele for the binding site identified by FIMO . It uses ENCODE motifs and JASPAR motifs to evaluate the regulatory potential of the SNPs . It outputs for each SNP the significance of the match to each position specific matrix with both the reference and the alternative allele and also the significance of the change in these match scores . atSNP relies on sampling algorithm with a first-order Markov model for the background nucleotide sequences to test the significance of affinity score and SNP-driven changes in these scores . atSNP is an R package . For pathways analysis , we used the Bioconductor package GOseq [29] combined with the KEGG annotation database [76] . GOseq is based on gene ontology analysis but allows for correction such as gene length that has been shown to bias gene set enrichment analysis . For eQTL associations , we control for gene length as implemented in the reference manual and for mQTL associations , we generated a list of the number of probes per genes using Illumina 450K and Refseq annotations . This list was then used as input to control for gene overrepresentation bias . Placenta specific genome annotation has been generated using publicly available data from the Roadmap in Epigenomics project . Chromatin immunoprecipitation followed by massively parallel sequencing ( ChIP-seq ) was obtained from fetal placenta primary tissue from several donors ( S16 Table ) . Annotation involved processing the raw data for chromatin accessibility ( DNase hypersensitivity ) along with ChIP-seq data for six histone modifications ( H3K4me1 , H3K27me3 , H3K27ac , H3K4me3 , H3K36me3 and H3K9me3 ) , followed by the use of the ChromHMM algorithm [77] to predict seven features as previously described . ChromHMM is based on a multivariate hidden Markov model that models the observed combination of chromatin marks using a product of independent Bernoulli random variables , which enables robust learning of complex patterns of many chromatin modifications . ChromHMM outputs both the learned chromatin-state model parameters and the chromatin-state assignments for each genomic position . Based on enrichment and genomic localization , feature 1 was annotated as candidate promoter ( H3K4me3 ) , feature 2 as candidate active enhancer ( H3K4me1 and H3k27ac ) , feature 3 and 4 as candidate poised enhancer ( H3K4me1 and H3k27me3 ) and feature 7 as transcribed sequences ( H3K36me3 ) . Finally , features 5 and 6 did not show enough specific enrichment to be annotated ( S7 Fig ) . Targeted bis-seq was performed for the validation of 3 mQTL regions and to assess the effect of common SNPs in 4 Illumina probes as described in Do et al . [13] . Briefly , the primers ( S24 Table ) were designed using MethPrimer . Bisulfite-converted DNA was amplified by PCR , followed by Nextgen sequencing ( Illumina MiSeq ) . The PCR and library preparation were performed using Fluidigm Access Array system . PCRs were performed in triplicate and pooled to ensure sequence complexity . ASM was assessed when the coverage was at least 100 DNA fragments . While the absolute differences between methylation of the two alleles are not exaggerated by ultra deep sequencing , the p-values tend to zero as the number of reads increases . Therefore , the Wilcoxon non parametric test was performed using bootstrapping ( 1000 random samplings , 20 reads per allele ) to minimize an artificially low p-value due to ultra deep coverage . The significance of the allele asymmetry was defined by p<0 . 05 and an absolute methylation difference between allele >20% . For the graphical representations of the bis-seq , one representative random sampling of the reads was shown per allele . Samplings and bootstrapping were performed using R . First , we assessed unimodal and bimodal distribution for expression and methylation data . For expression , a Bayesian model of normal distribution was used to estimate the parameters of the normal distribution fitting the data . The likelihood function is given by a normal distribution: expi~dnorm ( μ , τ ) Here , expi is the expression value for the ith sample . μ denotes the mean of the normal distribution , τ is the precision of the normal distribution . Standard deviation σ=1τ . The prior for the parameter μ and τ are described by two non-informative priors as: μ~dnorm ( 0 , 0 . 0001 ) I ( 0 , ) τ~dgamma ( 0 . 01 , 0 . 01 ) Here , the prior for the μ is distributed on the right side of a normal distribution with mean 0 and standard deviation 100 . The prior for the precision is described as a gamma distribution with mean 1 and standard deviation 10 . To assess the bimodal distribution of expression data , a mixture of two normal distributions model was used to estimate the means , standard deviation and the mass of the bimodal distribution . The likelihood function can be written as: expi~ω1dnorm ( μ1 , τ ) +ω2dnorm ( μ2 , τ ) Here , ω1 and ω2 represent the mass of each mixture . μ1 and μ2 are the means of each mixture . We assumed that the two mixtures have similar variance , τ . The prior for the mass ω1 is described as: ω1~dbeta ( 1 , 1 ) [0 . 1 , 0 . 9] ω1+ω2=1 Here , the prior of the mass for the first mixture follows a uniform beta distribution ranging from 0 . 1 to 0 . 9 , which means the proportion for any mixture has to be larger than 10% . And the two mixtures add up to 1 . The prior for the means μ1 and μ2 are described the same way as the mean in a unimodal model . Because the methylation value for CpG site ranging from 0 to 1 , we used a beta distribution to describe the unimodal distribution of the methylation data instead of the normal distribution . Likelihood: methyi~dbeta ( α , β ) Here , methyi is the methylation value for the ith sample . The parameters of the beta distribution α and β are determined by μ , the mean of the methylation data , and κ , the variation of the methylation data , as: α=μ*κ β= ( 1-μ ) *κ The priors for μ and κ are modeled as: μ~dbeta ( 1 , 1 ) κ~dgamma ( 0 . 01 , 0 . 01 ) Here , μ follows a non-informative uniform beta distribution , and κ follows a gamma distribution . For bimodal distribution of methylation data , a mixture of two beta distributions was used . The likelihood: methyi~ω1dbeta ( α1=μ1κ , β1= ( 1-μ1 ) κ ) +ω2dbeta ( α2=μ2κ , β2= ( 1-μ2 ) κ ) Here , ω1 and ω2 represent the mass of each mixture . μ1 and μ2 are the means of each mixture . We assume that the two mixtures have similar variance , κ . The prior for the mass ω1 is described as: ω1~dbeta ( 1 , 1 ) [0 . 2 , 0 . 8] ω1+ω2=1 Here , the prior of the mass for the first mixture follows a uniform beta distribution ranging from 0 . 2 to 0 . 8 , which means the proportion for any mixture has to be larger than 20% . And the two mixtures add up to 1 . The prior for the means μ1 and μ2 are described the same way as described in unimodal methylation data . Having both the model and data , we used the Gibbs Sampler JAGS ( Martyn Plummer 2003 ) to sample the posterior distribution through the “rjags” package ( Martyn Plummer 2016 ) . Finally , we compared the goodness of fit of a unimodal and bimodal model for the data by Bayesian Information Criteria ( BIC ) . The unimodality or bimodality of the data is determined by the model with better BIC . The mean and standard deviation of the data is estimated from the posterior distribution sampled by JAGS . Codes used during our analysis are available here: [https://github . com/fabiendelahaye/Placenta_analysis . git] | The placenta is a critical organ playing multiple roles including oxygen and metabolite transfer from mother to fetus , hormone production , and vascular perfusion . With this study , we aimed to deliver a placenta-specific regulatory map based on a combination of publicly available and newly generated data . To complete this reference , we obtained genotype information ( n = 303 ) , DNA methylation ( n = 303 ) and expression data ( n = 80 ) for placentas from healthy women . Our analysis of methylation and expression quantitative trait loci ( QTLs ) and correlations between methylation and expression data were designed to identify fundamental associations between genome , transcriptome , and epigenome in this key fetal organ . The results provide high-resolution genetic and epigenetic maps specific to the placenta based on a representative ethnically diverse cohort . As interest and efforts are growing to better understand the etiology of placental disease and the impact of the environment on placental function these data will provide a reference and enhance future investigations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"genome-wide",
"association",
"studies",
"gene",
"regulation",
"regulatory",
"proteins",
"dna-binding",
"proteins",
"alleles",
"dna",
"transcription",
"mathematics",
"genome",
"analysis",
"transcription",
"factors",
"epigenetics",
"dna",
"dna",
"methylation",
"chromatin",
"discrete",
"mathematics",
"combinatorics",
"genomics",
"chromosome",
"biology",
"proteins",
"gene",
"expression",
"chromatin",
"modification",
"dna",
"modification",
"genetic",
"loci",
"biochemistry",
"permutation",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"genetics",
"of",
"disease",
"human",
"genetics"
]
| 2018 | Genetic variants influence on the placenta regulatory landscape |
Plant morphogenesis is strongly dependent on the directional growth and the subsequent oriented division of individual cells . It has been shown that the plant cortical microtubule array plays a key role in controlling both these processes . This ordered structure emerges as the collective result of stochastic interactions between large numbers of dynamic microtubules . To elucidate this complex self-organization process a number of analytical and computational approaches to study the dynamics of cortical microtubules have been proposed . To date , however , these models have been restricted to two dimensional planes or geometrically simple surfaces in three dimensions , which strongly limits their applicability as plant cells display a wide variety of shapes . This limitation is even more acute , as both local as well as global geometrical features of cells are expected to influence the overall organization of the array . Here we describe a framework for efficiently simulating microtubule dynamics on triangulated approximations of arbitrary three dimensional surfaces . This allows the study of microtubule array organization on realistic cell surfaces obtained by segmentation of microscopic images . We validate the framework against expected or known results for the spherical and cubical geometry . We then use it to systematically study the individual contributions of global geometry , cell-edge induced catastrophes and cell-face induced stability to array organization in a cuboidal geometry . Finally , we apply our framework to analyze the highly non-trivial geometry of leaf pavement cells of Arabidopsis thaliana , Nicotiana benthamiana and Hedera helix . We show that our simulations can predict multiple features of the microtubule array structure in these cells , revealing , among others , strong constraints on the orientation of division planes .
It is well known that the cortical microtubule ( hereafter abbreviated to MT ) cytoskeleton in plant cells plays a decisive role in controlled cell expansion and oriented cell division , which together drive the plant morphogenesis [1] . In the absence of MT organizing centers like centrosomes [2] , higher plants establish an ordered array of MTs at the cell cortex , the so-called cortical array ( CA ) [3] . Recent studies have revealed that cell shape may have influence on the orientation of the CA [4] . On the other hand , the orientation of the CA controls cell expansion and cell anisotropy , by guiding the deposition of cellulose synthase complexes along the MTs [5–8] . Through this coupling , the CA in turn can influence the cell shape , essentially setting up a morphogenetic feedback loop . This loop is possibly also amplified by a mechanical feedback mechanism discussed in [9] . Therefore , understanding both anisotropic cell expansion and oriented cell division , requires understanding the formation of the ordered CA from an initially disordered state just after cell division [10] . For non-growing cells , the influence of cell shape on the formation of the CA will be static in nature and thus determined by geometrical features alone . A significant variety in these features is observed between different cell types ( see Fig 1 ) . For growing cells , the evolving cell shape may generate a corresponding dynamic influence on the CA formation process . If growth is slow compared to the collective dynamical time scale of the MTs , a quasi-static approach which samples cell shapes at different time points may still be a reasonable approximation . MTs are highly dynamic and filamentous protein polymer aggregates , and form one of the principal components of the plant cytoskeleton [11] . MTs have structurally two distinct ends—a minus-end and a plus-end . The plus-end can dynamically switch from a growing state to a shrinking state or vice-versa . Switching of a MT plus-end from a growing state to a shrinking state is called catastrophe while the reverse switching of a shrinking state to a growing state is called rescue . This phenomenon of reversible switching of MT plus-ends between two states is called dynamic instability . On average , the minus-end of an unstabilized MT continually is in a shrinking state . Thus , the combination of overall growth at the plus-end and shrinkage at the minus-end seemingly moves a MT as a whole . This motion is called treadmilling and has been observed in both in vitro [12–14] and in vivo [15] . In contrast to animal cells , plant cells do not have a well defined MT organizing center . Instead MT activity is dispersed over the whole cell cortex , driven by the localized nucleation of new MTs by γ-tubulin complexes [16] . The cortical MTs are confined to a thin layer of cytoplasm just inside the plasma membrane of the plant cell and are attached to the cell envelope , ensuring that MTs do not translate or rotate as a whole [17 , 18] . In spite of their fixed attachment to the cell cortex , cortical MTs do show mobility which is due to treadmilling motion [17 , 19] . Two dimensional attachment to the cell cortex allows MTs to interact with each other via collisions , which occur when the polymerizing plus-end of a growing MT encounters a pre-existing MT . Depending on the value of the collision angle , three different possible events are observed [20] . For shallow angles ( ≲ 400 ) , a growing MT bends toward the direction of the MT encountered and this kind of adaptive event is called zippering . For steeper angles ( ≳ 400 ) , the encounter may lead to a so-called induced-catastrophe , where the initially growing MT switches to a shrinking state . Alternatively , the growing MT may slip over the one encountered , leading to a crossover event ( see Fig 2 ) . In vivo imaging of cortical MTs has revealed that they nucleate at the cortex , either from isolated nucleation complexes or from pre-existing MTs , and gradually develop from an initially sparse and disorganized state into a final ordered array over a time period of an hour [17 , 18 , 21–24] after the previous cytokinesis . MTs have a finite lifetime and ultimately disappear by shrinking to zero length . Plant cells within tissues typically have well defined relatively flat faces , bordered by cell-edges of significantly higher curvature [25] . In root epithelial cells microtubules have been observed to undergo catastrophes when they encounter these cell-edges with a probability that increases with increasing curvature of the cell-edge [26] . We will denote these as edge-catastrophe events . While existing experimental studies inform us about the molecular events and key parameters that are involved in CA formation [17 , 19 , 24 , 27 , 28] , it represents a complex emergent phenomenon that is the macroscopic result of a large number of stochastic microscopic events involving large numbers of MTs . For a full mechanistic understanding of biological processes of this type , mathematical and computational modelling has by now been recognised as indispensable [29] . From the outset , starting with the seminal work of Dixit & Cyr [20] , attempts have been made to model CA formation ( for reviews on the different approaches involved please consult [30 , 31] ) . From this work a consensus hypothesis has emerged to explain the spontaneous order observed in the CA . This mechanism , also dubbed ‘survival of the aligned’ [32] , can account for the statistically robust spontaneous alignement of MTs under well-characterized conditions , and appears to be consistent with observations . We therefore adopt this hypothesis as the starting point in the present work . As far as the influence of cell geometry on CA formation is concerned , modelling studies have so far been limited to two dimensional ( 2D ) approaches where the effect of a closed shape in three dimension ( 3D ) was mimicked by boundary conditions [33] , or geometrically simple shapes where boundaries are readily described by mathematical expressions , such as cubes [34] and cylinders [35] . Also , some of the simulations reported relied on finite time step brownian dynamics , at most using a Gillespie-type event-sampling algorithm [36] for the spontaneous changes in MT dynamical state . This severely limits the time-efficiency of these simulations , and makes the acquisition of results for wide ranges of parameters with sufficient statistics a very difficult task . Tindemans et al . [35] therefore developed a fully event-driven algorithm ( for a general introduction see [37] , Chapter 6 . 1 ) , in which not only the spontaneous stochastic state changes but also the collisions between MTs are performed by directly sampling from a dynamic list of potential future collision events and their associated waiting times , which is constantly updated . This achieves a speed up of at least three orders of magnitude with respect to standard finite time step algorithms , and eliminates errors in the location of the collision points as well . Here we show how the algorithm of [35] can be implemented efficiently on arbitrary triangulated surfaces , such as those obtained by segmenting 3D reconstructed confocal microscopy images of cells . Strikingly , the additional computational overhead incurred by having to determine the correct passage from triangle to triangle in a manner consistent with the 3D path , is almost completely balanced by the speed up of the collision predicting algorithm due to the strong localization of the search area . In this way the new implementation is almost as fast as the original implementation , and allows the rapid exploration of many different parameter settings and/or geometries each sampled independently many times to obtain statistically reliable averaging over the stochastic ensemble . In addition , the MT dynamical parameters can be chosen independently within each triangular domain and passage probabilities can be associated with the shared edges between the triangular domains . This allows us to easily implement biologically relevant effects such as the observed edge-catastrophes at locations of strong curvature , or potential differences in local MT dynamics due to developmental distinctions between cell-faces , e . g . between faces created by recent cell divisions and faces that have expanded by growth .
In Fig 3 , we illustrate the procedure of transforming experimentally obtained confocal image into simulation input . First , using the image processing software morphographX [38] , we segment the experimentally obtained confocal images and extract the different cell shapes . For segmentation , confocal image stacks ( tif ) were Gaussian blurred using sigma value 0 . 3 μm , subsequently we applied the ITK watershed auto-seeding with level threshold value in the range 500–1500 . Then , we approximated the segmented cells by creating triangulated surface meshes using marching cubes 3D with cube size in the range 0 . 5–1 . 0 μm . Using the surface mesh processing software meshLab [39] , we tag the different cell-edges and cell-faces , which are represented by appropriately coloured associated triangles . Finally , all this information is stored in a single mesh file , where each triangle has a cell-edge tag with the values of the associated cell-edge angles and cell-face tags ( see S1 File , Sec . SI . 1 ) . Confinement of the dynamics of cortical MTs to the surface ( cortex ) of plant cell effectively reduces the system to two spatial dimensions . In our modelling framework , we exploit this advantage by treating the cell shape as a 2D surface , embedded in the ambient 3D space . Because we are working with a triangulated approximation of surfaces , we developed an algorithm to establish a connectivity graph of the triangles [40] in order to appropriately propagate MT dynamics between adjacent triangles . Using a combination of rotations , implemented by quaternions [41] , and z-axis translation , we first transform the 3D-triangulated surface from the 3D x − y − z space to the 2D x − y plane ( see Fig 4 ) . Each individual triangle has a unique quaternion operator which is fully determined by its normal vector in 3D , which operates on all its vertices simultaneously . Different triangles have different normals , and hence end up in different , typically disjoint , 2D positions after rotation , irrespective of whether they had shared edges or not , but this position is irrelevant to the dynamics within each triangle . We simulate MT nucleation , growth , shrinkage and collision events within each individual triangle ( see Fig 5 ) . In order to nucleate new MTs on the surface , we first randomly select a triangle from the entire set and generate uniformly distributed points within the selected triangle [42] . The initial MT growth direction is made isotropic by choosing a random angle of nucleation in the range [0 , 2π] . The persistence length of a MT is much longer than its average length , typically of the order of a few μm’s , [43 , 44] and therefore much larger than the typical dimensions of our triangles ( well below 1μm ) , allowing us to model the individual MT within a triangle as an elongating straight line segment . The location of the collision point between two MTs is then readily determined by calculating the intersection point of their respective trajectories [45 , 46] . To deal with the various MT events , we implement an event-driven simulation technique [35] . The events are separated into two categories: stochastic events and deterministic events . The stochastic events associated with MTs are independent and result from Poissonian random processes specified by the nucleation rate rn , the spontaneous-catastrophe rate rc and the rescue rate rr . The deterministic events associated with the MTs include collisions , disappearance , intersection with triangle-edges , and simulation control events ( i . e extracting simulation output at fixed time intervals ) . In between events , the plus-ends of MTs either grow or shrink with velocities v+ and v− respectively , while the minus-end retracts with the treadmilling speed vtm , so that all length changes between events are readily computed . We use triangle-edge to triangle-edge links ( obtained from the connectivity graph ) to propagate MT dynamics from one triangle to its relevant neighbour . For example , in Fig 6 ΔABC is connected with ΔADC through the shared edge AC , which are linked . The proper transition of a MT from ΔADC to ΔABC through the edge AC on the 3D surface requires , in the 2D plane , Using an affine transformation [47] on the coordinates of the growing MT plus-end tip at point P′ ( x′ , y′ ) of edge AC ( ΔADC ) , we translate ( see Fig 6 left panel ) the tip to the edge AC ( ΔABC ) . Next , we rotate ( see Fig 6 right panel ) the MT trajectory with rotation angle , θ R = arccos ( A C → ( Δ A B C ) . A C → ( Δ A D C ) | A C → | 2 ) ( 1 ) This rotation in the 2D plane implements the assumption that a MT grows along the straightest possible path on the 3D surface , also known as a geodesic . Although the mechanical energy of bending could conceivably also play a role in determining the path of the MT , potentially causing a more complex coupling to the geometry , we have , in the absence of sufficient experimental data discriminating these effects , chosen to opt for this arguably simplest rule . During transition of a MT from one triangle to another triangle , we implement the probability of local edge-catastrophe based on the local value of triangle-edge angle between the adjacent triangles ( see S1 File , Sec . SI . 2 ) . We also use a cell-face tag , which is assigned to the triangles during the input surface mesh creation ( see S1 File , Sec . SI . 3 ) , to implement local differences in stability of MTs e . g . by changing the local spontaneous-catastrophe rate . More generically , our simulation framework allows arbitrary domain specific parametrisation , with the smallest domain being a single triangle . This flexibility allows full freedom to incorporate experimental observations on local MT dynamics or test additional relevant hypotheses . Finally , whenever required we apply the inverse quaternion transformation on the set of 2D triangles and the MT segments present on them , to reconstruct the 3D surface with the MT configurations realised . Through this transformation , segments of a MT which appear to be disconnected in the set of 2D triangles , get ( re ) connected on the 3D surface . As illustrated in Fig 7 , the segment M 1 M 2 ¯ is attached to ΔABC and the segment M 2 M 3 ¯ is attached to ΔADC in the 2D plane . After the inverse quaternion mapping , these two segments get connected in the 3D surface and represent a single MT with end points ( M1 , M3 ) . To measure the degree of order and the orientation of the MT array , we define two order parameters . The first parameter is a scalar ( Q ( 2 ) ) , which measures the average degree of ordering of the MTs , and the second a vector Ω ^ , which indicates the global orientation of the array . Both these parameters are derived from a single tensorial order parameter Q ( 2 ) ( see S1 File , Sec . SI . 4 ) , Q ( 2 ) being the absolute value of the smallest eigenvalue of this tensor and Ω ^ the corresponding normalized eigenvector , i . e . | Ω ^ | = 1 . For completely random orientation of MTs Q ( 2 ) ≈ 0 and for well organised MTs that form an array , Q ( 2 ) ⪅ 1 . The direction of Ω ^ is perpendicular to the average local orientation of the MTs ( see Fig 8 ) , i . e . perpendicular to the plane onto which the total projection of individual MTs is maximal on average . In Fig 9 , we present an overview of the entire simulation approach . In order to obtain a suitable set of MT dynamical parameters for our simulations , we make use of the control parameter G , introduced and used in [48 , 49] . This parameter encapsulates the effect of all six single MT dynamics parameters ( speeds v+ , v− and vtm , dynamical instability rates rc , rr and nucleation rate rn ) into a single magnitude that controls the frequency of MT interactions which are required to ensure a spontaneously ordered state . Its explicit form identifies it as the ratio between two length scales G = − ( 2 ( v + − v t m ) ( v − + v t m ) r n ( v + + v − ) ) 1 3 ( r r v − + v t m − r c v + − v t m ) − 1 = − l 0 l a v g . ( 2 ) Here l0 is the MT-MT interaction length , and lavg is the average length of MTs in the absence of any interaction effects . The statement that G is the single control parameter was only strictly derived for spatially homogeneous and open , boundary-free domains . To allow for possible effects of a closed cell geometry , we chose to independently vary the two factors involved: l0 by varying the MT nucleation rate rn and lavg through varying the spontaneous-catastrophe rate rc . Moreover , and unless explicitly mentioned , we implement homeostatic control of the MT growth speed v+ by implementing a finite tubulin pool ( see S1 File , Sec . SI . 6 for details ) , as this speeds up the relaxation towards a steady state . The remaining simulation parameters are taken from [35] . We tested the parameter choice on a spherical surface with radius r ≈ 6 μm , inspired by the typical dimensions of a Arabidopsis thaliana embryonic cell . This is at the smaller end of the spectrum of plant cell sizes and therefore the most stringent test on finite size effects . We found that we could obtain robustly ordered arrays with Q ( 2 ) ≳ 0 . 70 at a state point with l0 = 2 . 05 μm and lavg = 417 . 5 μm , corresponding to G ≈ −0 . 005 ( see Fig 10 ) . The fact that G < 0 ensures bounded growth regime , where the length of MTs is always finite . All parameters used are summarised in S1 File , Sec . SI . 5 ( Table 1 ) . It is important to note that lavg is the theoretically expected value of average MT length assuming absence of any interaction effects . However , in simulation MTs will interact with each other leading to changes in effective values of MT dynamics parameters , i . e . modification in the values of over all catastrophe rates due to effect of induced-catastrophe . These dynamic changes in MT dynamics parameter will also modify the average MT length to a value l ¯ different from lavg . In the coming sections , average MT length will correspond to the simulated value of average MT length , i . e . l ¯ . For further clarification on this point , variation in the value of l ¯ with respect to lavg is described in S1 File , Sec . SI . 6 . Finally , we did not take any cell cycle related variations in MT dynamics into account , as this work focusses on understanding the relative contributions of the basic MT-MT interaction rule , shape anisotropy , edge-catastrophe and differential MT stabilization at developmentally different cell-faces on CA formation in interphase . We note , however , that such time-dependent effects could readily be implemented within our framework . The simulation code is written in C++ , and compiled using GNU GCC 4 . 8 . 4 compiler . Development and incidental runs were done on a standard PC equipped with Linux . Production runs were performed on the Dutch National Grid infrastructure comprising 10000 distributed compute nodes . Individual runs of 10 hours of biological time took between 20 seconds to 120 seconds wall clock time , depending on the parameter settings .
We performed a set of simulations validating our computational framework . We first tested to what extent the triangulation of otherwise smooth surfaces influences the final results . To do so we considered the case of approximating a perfect sphere , varying both the number of triangles employed , as well as the triangulation method . The key desiderata in this case are ( i ) that in steady state , due to the spherical symmetry , there should be no bias in the orientation Ω ^ of the ordered array , and ( ii ) that the value of the scalar order parameter Q ( 2 ) does not depend on the nature of the triangulation . Next , in order to test the system as a whole against a known , and geometrically non-trivial case , we reproduced previously reported results on a cubic surface [50] . We are now in a position to use our framework to study the role of cell shape on MT array formation . We first do this in a geometrically simple setting , which allows us to systematically disentangle the role the various shape-related factors involved . Then we turn real cell shapes , provided by the leaf pavement cells of Arabidopsis thaliana , Hedera helix and Nicotiana benthamiana .
Arabidopsis thaliana and Nicotiana benthamiana plants were grown as described in [76] . Leaves from these plant species were harvested and incubated in Renaissance staining as described in [77] , samples were then mounted in water to visualized their cell shape . 35S::MAP65-GFP expression in Arabidopsis thaliana leaves was analysed using 3 days old seedlings and 4 weeks old plants respectively . Expression assays in Nicotiana benthamiana leaves were performed using 4 weeks old plants . Agrobacterium containing 35S::TUB-mCHERRY were infiltrated as described in [78] . Arabidopsis thaliana leaves were stained with Propidium idodide ( 10μg/ml ) and imaged using a C- Apochromat 40 X/1 . 20 W Korr water immersion objective of a LSM 880 airyscan Zeiss confocal microscopes . GFP was excited using an argon laser 488nm , and fluorescence emission was detected from 500 to 540 nm . Renaissance staining for Nicotiana benthamiana leaves was detected at 415–440 nm excitation and a 405 nm beam splitter . For 35S::TUB-mCHERRY localization in Nicotiana benthamiana , infiltrated leaves were mounted in water and mCHERRY was detected with 543 nm excitation and 488/543/633 beam splitter . | In contrast to animal cells , plant cells are encased in a rigid cell wall providing among others the necessary rigidity to contain the turgor pressure that allows the plant as a whole to raise itself against gravity . In order to maintain this mechanical integrity while growing requires plant cells to rigorously control their expansion . In this they are aided by a plant-unique intracellular structure called the cortical array , an organised set of long filamentous protein polymers called microtubules attached to the inside of the cell membrane . Experiments and modelling studies have revealed how collisions between these highly dynamic polymers promote the spontaneous alignment of microtubules observed in cells . However , it was hitherto unknown how the shape of cells influences this self-organisation process and helps in determining the overall orientation of the cortical array . The latter is key in deciding the axis of cell expansion and the orientation of cell division planes . Here we develop , validate and apply a simulation framework for studying the dynamics of interacting microtubules on arbitrary triangulated surfaces that can be extracted from three-dimensional reconstruction of real cells from confocal microscopy images . This allows us to start unravelling the delicate interplay between cell shape , effects of sharp cell-edges and developmental cues on the organisation of the cortical array . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"plant",
"anatomy",
"microtubules",
"pavement",
"cells",
"cell",
"cycle",
"and",
"cell",
"division",
"plant",
"cell",
"biology",
"cell",
"processes",
"brassica",
"condensed",
"matter",
"physics",
"plant",
"science",
"model",
"organisms",
"experimental",
"organism",
"systems",
"plants",
"flowering",
"plants",
"cellular",
"structures",
"and",
"organelles",
"cytoskeleton",
"research",
"and",
"analysis",
"methods",
"arabidopsis",
"thaliana",
"nicotiana",
"leaves",
"physics",
"eukaryota",
"plant",
"cells",
"plant",
"and",
"algal",
"models",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"nucleation",
"organisms"
]
| 2018 | A computational framework for cortical microtubule dynamics in realistically shaped plant cells |
A mutated KRAS protein is frequently observed in human cancers . Traditionally , the oncogenic properties of KRAS missense mutants at position 12 ( G12X ) have been considered as equal . Here , by assessing the probabilities of occurrence of all KRAS G12X mutations and KRAS dynamics we show that this assumption does not hold true . Instead , our findings revealed an outstanding mutational bias . We conducted a thorough mutational analysis of KRAS G12X mutations and assessed to what extent the observed mutation frequencies follow a random distribution . Unique tissue-specific frequencies are displayed with specific mutations , especially with G12R , which cannot be explained by random probabilities . To clarify the underlying causes for the nonrandom probabilities , we conducted extensive atomistic molecular dynamics simulations ( 170 μs ) to study the differences of G12X mutations on a molecular level . The simulations revealed an allosteric hydrophobic signaling network in KRAS , and that protein dynamics is altered among the G12X mutants and as such differs from the wild-type and is mutation-specific . The shift in long-timescale conformational dynamics was confirmed with Markov state modeling . A G12X mutation was found to modify KRAS dynamics in an allosteric way , which is especially manifested in the switch regions that are responsible for the effector protein binding . The findings provide a basis to understand better the oncogenic properties of KRAS G12X mutants and the consequences of the observed nonrandom frequencies of specific G12X mutations .
The small GTPase protein KRAS is a signal-transducing protein , which binds GDP in its inactive state and GTP in its active state [1] . The gene KRAS is frequently mutated in various human cancers . The mutation is most often , in about 86% of the cases [2] , found at G12 . In fact , every missense mutation at G12 ( G12X ) is oncogenic . The oncogenic properties associated with KRAS G12X mutation are characterized by the deficiency of the intrinsic GTPase activity and the insensitivity for GTPase-activating proteins ( GAPs ) [3 , 4] . These alterations lead to increased KRAS signaling , as there is more active GTP-bound protein present . Still , the mutant KRAS undergoes GDP–GTP cycling [5] . The basis of the specific G12X mutation frequencies has remained unclear , except for the G12C transversion mutation ( c . 34G>T ) associated with smoking in lung cancer [6 , 7] . An interesting discrepancy among KRAS G12X mutants is observed in their intrinsic GTPase activity [8] . The G12A mutation exhibits the most hindered intrinsic hydrolysis ( ~1% compared to the wild-type ) , whereas the G12C mutation displays the least hindered activity ( ~72% ) . All G12X mutants , however , show insensitivity to GAPs that accelerate hydrolysis [8] . Importantly , not only RAS G12X mutants exhibit a discrepancy in GTP hydrolysis , but they also give rise to differences in the preferred signaling pathway ( in terms of effector protein binding ) [9 , 10] . This behavior was first observed in NSCLC cell lines [9] , where KRAS G12D showed activation of PI3K and MEK signaling , while G12C and G12V mutants exhibited activated RalGDS-pathway and diminished growth factor-dependent Akt activation . Furthermore , an NMR study revealed different binding preferences for mutant HRAS G12V compared to wild-type HRAS , with various effector proteins [10] . Here , HRAS G12V showed reduced interactions with Raf and enhanced binding with RalGDS . However , given that the non-hydrolysable GTP-analog GNP was used in the study , the difference is not due to impaired hydrolysis . Similarly with HRAS , KRAS G12X mutants exhibit reduced affinity to Raf compared to wild-type [8] . The G12D , G12R , and G12V mutants display highly reduced affinity to Raf , while the affinity of G12A is only moderately reduced . Interestingly , the affinity of the G12C mutant is similar to that of wild-type . To bind RAS , the effector proteins use a ubiquitin ( UB ) -like fold: a RAS-binding domain ( RBD ) or a RAS-association domain ( RA ) [11 , 12] . While KRAS has not been co-crystallized with any of its effector proteins , distinct effector proteins have been resolved in complex with HRAS: RalGDS ( PDB ID: 4G0N ) [13] , Raf-1 ( PDB ID: 1LFD ) [14] , PI3Kγ ( PDB ID: 1HE8 ) [15] , PLCε ( PDB ID: 2CL5 ) [16] , RASSF5 ( PDB ID: 3DDC ) [17] and AF-6 ( PDB ID: 6AMB ) [18] . These effector proteins bind to HRAS on top of its switch regions: switch-I ( residues 30–40 ) and switch-II ( residues 58–72 ) , and the binding conformation of HRAS is almost identical in all of the complexes ( S1A Fig ) . Given this , and since the G12X mutation is far from the binding interface ( S1B Fig ) , Smith and Ikura [10] proposed that the discrepancies in the effector protein binding profiles of the mutants are due to altered switch dynamics . Overall , switch-I displays highly dynamic characteristics manifested as two different states when GTP is bound to RAS , and the distribution between these states is altered in mutants [19–22] . Given that the switch regions in HRAS and KRAS are identical ( S1C Fig ) , their expected binding mode to their effectors is alike . A model of KRAS in complex with A-Raf-RBD tethered to a lipid-bilayer nanodisc suggested by NMR data agrees with this binding mode [23] . At the cellular level , the isoform specificity to effector proteins is primarily determined via membrane interactions [24] , but the differences among RAS isoforms’ absolute effector protein binding affinities rise from allosteric effects [25] . It was observed that even a single point mutation in RAS ( Q61L ) has long-range effects on dynamics and alters effector protein interactions [13] . Previous molecular dynamics ( MD ) simulation studies of KRAS at microsecond timescales have mainly focused on the dynamical differences between the three wild-type RAS isoforms ( HRAS , KRAS , NRAS ) [26] , differences among selected KRAS and HRAS mutants [27 , 28] , the role of the hypervariable region ( HVR ) [29] , KRAS’s membrane association or orientation [30–32] , and KRAS oligomerization on the membrane [33] . The total simulation times of these studies were in the range of 1–8 μs , which is reasonable but likely not sufficient to unravel long-time dynamics associated with slow conformational changes . More importantly , there is a lack of comprehensive atomistic MD simulations of all KRAS G12X mutants with extensive simulation times , allowing a reliable analysis for the differences in structure and dynamical behavior between the wild-type and the mutants , especially in the effector protein binding interface . What is the underlying cause for the broad range of different G12X mutations ? How do these distinctly different mutations manifest themselves in the structure , dynamics , and function of KRAS ? This knowledge is crucial to understand KRAS oncogenesis and to develop future therapies targeting mutant KRAS harboring tumors . Therefore , in the present study we first assessed to what extent G12X mutation frequencies are explained by mutation probability . Intriguingly , an outstanding mutational bias emerged from the data . We next employed state-of-the-art atomistic MD simulations ( total simulation time 170 μs ) to study the dynamical behavior of KRAS with its natural ligands ( GDP , GTP ) bound , both in the wild-type KRAS and with all existing oncogenic G12X mutations . The results provided compelling evidence that mutations alter the dynamics of KRAS , that the alteration is mutation specific , displays allosteric characteristics , and that the alteration is manifested especially in the effector protein-binding interface . Furthermore , our data suggest that the observed mutational bias and the oncogenic properties of the individual KRAS G12X mutants are caused , at least in part , by mutation-specific altered dynamics .
First , to perceive up-to-date data of KRAS G12X missense mutation frequencies , we compiled data from the Catalogue of Somatic Mutations in Cancer [2] . A total of 32 , 654 tumor samples identified with a KRAS G12X missense mutation were found from the database . For our analysis , we included only tissues that exhibited these mutations >10% . This status is displayed in eight tissue types , which in total comprised 31 , 251 positive samples ( 95% of all KRAS G12X mutations in the database ) . The large intestine ( 18 , 174 ) , the lung ( 5 , 640 ) , and the pancreas ( 5 , 528 ) were observed to have numerous positive samples , whereas the biliary tract , the endometrium , the ovary , the peritoneum , and the small intestine comprised altogether only 2 , 085 positive samples . A point mutation in KRAS G12X may result in one of six possible missense mutations ( Fig 1J ) . However , instead of being evenly distributed , these specific mutations display considerable variation ( Fig 1A ) . Overall , G12D ( 42% ) , G12V ( 28% ) , and G12C ( 14% ) mutations are very common , whereas G12A , G12R , and G12S are less popular . When the relative fractions of these mutations are considered in different tissues , they are readily observed to vary significantly ( Fig 1B–1I ) [2] . For instance , the G12R mutation is observed in the pancreas with a probability of 13% , while in the small intestine it appears in less than 2% of the cases . The predominating mutations are G12D and G12V , the lung being an exception with G12C standing as the most abundant mutation . In a G12X missense mutation , the guanine ( G ) base in c . 34G or c . 35G is substituted to adenine ( A ) , cytosine ( C ) , or thymine ( T ) ( Fig 1J ) . This base-substitution type exhibits variation ( Fig 1K ) . G>A and G>T mutations ( 47 . 4% and 42 . 1% , respectively ) occur very often , while the G>C mutation ( 10 . 5% ) takes place more seldom . These occurrences display some variation in different tissues . Particularly the lung differs from other tissues with a higher G>T fraction and a diminished proportion of G>A mutations ( P < 0 . 001 ) . Meanwhile , in the pancreas the probability of the G>C mutation is increased ( P < 0 . 001 ) . Moreover , as all of the G12X mutations occur in the first or the second guanine of the codon ( c . 34G , c . 35G ) ( Fig 1J ) , we ascertained if there is a mutational bias between these positions . Interestingly , 76 . 6% of the G12X mutations are c . 35G>X mutations ( G12A , G12D , G12V ) and only 23 . 4% are c . 34G>X mutations ( G12C , G12R , G12S ) ( Fig 1L ) . In fact , all tissues , except for the lung ( 55 . 3% ) , display 77–90% of c . 35G>X mutations . The positional mutation preference for c . 35G>X seems to be the highest with a G>A mutation ( >7x ) , whereas G>C or G>T exhibit nearly twofold preference , 1 . 75- and 2 . 03-fold , respectively ( Fig 1M ) . A few exceptions in the c . 35G>X preference , however , appear in specific tissues . In the pancreas , the G>C mutation occurs nine times more often in c . 34 than in c . 35 ( Fig 1O ) . As for the G>T mutations , the lung is the only tissue where c . 34 is preferred ( >1 . 5x ) ( Fig 1P ) . All tissues , interestingly , exhibit over fivefold c . 35 preference in G>A mutations ( Fig 1N ) . Above all , the pancreas ( >28x ) , the peritoneum ( >37x ) and the ovary ( >41x ) exhibit the most prominent preference for the G>A mutation in c . 35 . We evaluated how random the occurrences of the specific G12X mutations are . To this end , we used the transition:transversion mutation ratio as a figure of merit , and compared this figure to a value of 2 . 3 , which is the ratio for missense mutations observed in large-scale genomic analyses [34 , 35] . If the mutations would take place randomly , G12D and G12S mutations should be the most abundant mutations as they are transition mutations ( S2 Fig ) . G12D mutation is consistent with this view , as it occurs very often in all tissues . Meanwhile , G12S is not consistent with this behavior at all , as it occurs in tumors , perhaps surprisingly , very rarely . Also , regardless of the tissue type , the G12V mutation is overexpressed compared to values expected based on the assumption of random occurrences . Concluding , the mutations’ probabilities of occurrences are not consistent with a transition:transversion mutation ratio based on a random process . Since local DNA-sequences have clearly a major influence on the mutation probability , a sequence-dependent basis for the observed non-random mutations may exist . The TGGT sequence lacks a typical hotspot mutation region , such as a CpG site [36] . However , an adjacent GG region is a susceptible site for a mutation [37 , 38] . The oxidation of guanine by endogenous reactive oxygen species may also result in DNA mutation [39] . Both guanines , the 5’G and the 3’G in a GGT-sequence , are found to act as sites for frequent one-electron oxidation reactions , and they exhibit only a minor difference ( 0 . 05 eV ) in their vertical ionization potential [37] . The oxidation can further transform guanine to 7 , 8-dihydro-8-oxoguanine ( 8oxoG ) , which promotes especially the G>T transversion mutation [40] , and the G>T mutations take place on a regular basis ( Fig 1K ) . Interestingly , studies of DNA-adduct formation by exogenous agents have resulted in adduct formation only in c . 34G , and not in c . 35G [41 , 42] . Finally , cigarette smoking promotes G12C mutations exhibited regularly in the lung tissue ( Fig 1D and 1K ) [7] . Concluding , there are several potential mechanisms able to alter the mutation profile of guanine , thereby leading to the data we discussed above . To understand how G12X mutations affect KRAS functionality , we conducted a total of 170 μs ( 85 x 2 μs ) atomistic MD simulations of wild-type KRAS and all G12X missense mutants , with GDP and GTP . Each individual system was replicated five to ten times starting from different initial conditions ( S1 Table ) . In the simulations , we observed no differences in the dynamics of the residue 12 ( or in its vicinity ) , which appeared to be extremely stable . In contrast , the switch regions ( switch-I and switch-II ) exhibited highly dynamic behavior demonstrated by the root-mean-square fluctuation ( RMSF ) analysis , which revealed major fluctuations in the protein in these regions ( S3 and S4 Figs ) . Nevertheless , there were no evident differences between the different systems , as generally the individual replicas displayed variation as much as the different systems . Only with GDP , the G12A and the G12S display a different RMSF profile in the switch-I region . To gain better insight into the protein dynamics , we conducted principal component analysis ( PCA ) [43] with an objective to find the most significant large-scale motions of KRAS . PCA revealed that the greatest dynamic movements in the protein occur in the switch regions ( Fig 2 , S5A Fig ) . Furthermore , the most significant principal components 1–3 ( PC1-3 , see S5B Fig for contributions ) highlight that there are strong differences between GDP-bound and GTP-bound systems . PC1 of the GTP-bound systems displayed movement only in the switch regions , whereas PC1 of the GDP-bound systems exhibited additional movement also in the α3-helix . PC2 of GDP-bound KRAS revealed the movement in the switch regions and also extensive motion in the α3-helix , the hairpin loop between the β2- and β3-sheets , and the P-loop ( see S1D Fig ) . PC2 of GTP-bound KRAS in turn brought out the movements observed in GDP-bound systems , and further also the motion in the α1- and α4-helices , and in the turn near the SAK-motif . These observations indicate that the key to resolve the changes in protein dynamics is the γ-phosphate . Notably , the α4-helix motion is only observed with GTP bound systems ( PC2 ) . This observation is in agreement with the experimental results by Mazhab-Jafari et al . [23] . They observed that the GDP-bound KRAS drives the protein in the “exposed” configuration on the membrane , where the α4-helix is located in close proximity to the membrane ( PDB ID: 2MSC ) . This would indicate that the dynamical stability of the helix is important for this state . In order to ascertain dissimilarities between the different systems , we next generated score plots for the principal components PC1-3 ( Fig 3 , S5C and S5D Fig ) . The results highlight dissimilarities between the wild-type KRAS and the mutants , as well as between GDP- and GTP-bound proteins . Interestingly , in all of these systems , only the G12R and G12S mutants with GDP appear to reside in the closed switch-I conformation , whereas all other systems eluded this conformation . Even more interestingly , both of these mutants evaded this conformation when they were bound to GTP . The fully open conformation of switch-I appears to be more accessible to the systems , especially with the G12D mutant with GDP . Moreover , wild-type in both GDP- and GTP-bound systems seems to have a unique state with a high-scoring value ( +3 and +4 ) in PC1 and a low-scoring value ( -2 and -0 . 5 ) in PC2 . Taken together , the results show for all the mutants that the profile of their large-scale motions differs from wild-type regardless of the bound ligand , and that the profile is also unique to each mutant . We extended the analysis by carrying out PCA for each system to illustrate the differences in their dynamics ( S6–S9 Figs ) . The individual PCA analyses highlight not just variation in the switch region movements among the systems , but they also show that specific systems display more dynamical behavior in the α3-helix , hairpin loop between the β2- and β3-sheets , the α4-helix , the loop between β5-sheet and α4-helix , and in the SAK-motif ( residues 145–147 ) regions . For example , in GTP-bound systems only the G12A , G12D , G12R , and G12V mutants exhibited movement in the α4-helix in their PC1 or PC2 . Interestingly , these are also the systems that exhibit clearly diminished Raf affinity [8] . Also , the G12R mutant with GTP displayed notably reduced movement in the switch-II region in both PC1 and PC2 . Even though there are no additional direct interactions from the mutated side-chains of G12X , a mutation in this position inflicts a change to the dynamics in the distant sites of KRAS that were highlighted by the PCA analysis . To investigate this , and to identify possible interaction network routes in KRAS , we conducted an interaction network analysis [44] . Interestingly , we identified a hydrophobic hub network in KRAS that indeed connects the distant sites in the structure and is able to convey these effects in an allosteric way ( Fig 4 ) . Therefore , a change in KRAS dynamics in one of the hydrophobic hubs could traverse through this network even to the distant sites . This hub network is comprised of 11 hubs: V14 , M72 , F78 , L79 , F90 , I100 , V114 , A146 , A155 , F156 and L159 . One of these hubs , V14 , is located in the P-loop , in the close proximity of G12X . This hub interaction network is highly distorted in G12A and G12S mutants ( S10 and S11 Figs ) . The distortion in these mutants is further not ligand dependent . For example , in the V14 hub , the G12A and G12S mutants lack the interaction to A81 ( <2% vs . wild-type 26 . 9% and 39 . 7% , with GDP and GTP , respectively ) , and also display highly diminished interaction to A11 compared to the wild-type KRAS ( S10A Fig ) . From the V114 hub , these mutants lack interactions to A155 and L79 , but instead have a strengthened interaction to I142 and a totally new interaction to L113 , which is not displayed by other systems ( S11A Fig ) . From the hub A146 , they lack the interactions to A18 and L19 ( S11B Fig ) . From the hub A155 , both lack the interactions to V114 , L79 , and I142 , but instead they have an elevated interaction to F156 , and the G12A has an additional interaction to V152 ( S11C Fig ) . In contrast to G12A and G12S , the other mutants ( G12C , G12D , G12R , and G12V ) seem to follow more closely the wild-type’s interaction patterns . However , selected interactions are shifted even with these mutants , although not that extensively as observed with G12A and G12S . For instance , in the hub M72 located in the switch-II region the interaction patterns are shifted with G12C , G12R , and G12V ( S10B Fig ) . Interestingly , the GTP-bound G12D mutant displays almost identical interactions with the wild-type . We also noticed that the frequency of the salt-bridge between the residues D154–R161 was altered in different systems ( S11F Fig ) . Both of these residues are located in the α5-helix , in the close proximity of three hydrophobic hubs: A155 , F156 and L159 ( Fig 4 ) . With the wild-type KRAS this salt-bridge is more stable with GDP ( 69 . 1% ) than with GTP ( 46 . 4% ) . Meanwhile , again with the G12A and G12S mutants this salt-bridge is highly distorted ( 4 . 5%–20 . 6% ) , regardless of the bound ligand . As discussed above , the PCA and the interaction network analysis suggest that the protein dynamics is altered among the systems , yet in some obscure manner . To gain better insight into the origin of these differences in wild-type and mutant GTP-bound systems ( active KRAS ) , we analyzed the simulation data by constructing Markov state models ( MSMs ) [45 , 46] to explore the long-time statistical conformational dynamics of KRAS . The goal here was to identify the clusters of highly identical protein conformational states , here called metastable states , and to explore how the conformations of the wild-type and the mutant proteins are distributed between these metastable states . For the analysis , we selected the wild-type KRAS together with the most abundant mutants G12D and G12V , and the intriguing G12R mutant , which displays a highly variable distribution in the different tissues ( Fig 1 ) . The MSM analysis identified seven metastable states represented schematically in Fig 5 . Overall , all systems populate frequently two of the states: the states VI and VII ( 77–87% ) . The less populated metastable states I-IV are specifically represented among the systems . The metastable state I is quite unique for G12R ( 6% ) and the state III for G12V ( 6% ) , whereas the other systems are mostly absent from these two states . The metastable state IV is only present in wild-type ( 3% ) and in the G12V mutant ( 3% ) . The moderately populated metastable state V , where switch-II appears in a fixed conformation , is rarely observed with the G12V mutant ( 1% ) , whereas it is similarly represented among the other mutants and the wild-type ( 12–16% ) . In fact , the switch-II conformation appears to be closed in the effector protein complexes ( S1A Fig ) . However , none of the observed metastable states corresponds to this specific switch-II binding end-point conformation . The states can be further divided in four groups based on their switch conformations ( Table 1 ) . The states I and V as combined form the first group , where switch-I appears to be in a fully open conformation and switch-II in a fixed conformation . This group is frequently occupied by G12R ( 24% ) , whereas it is mostly absent from G12V ( 1% ) . The states VII and VI form individually the second and the third groups , in respective order . In the state VII , switch-II is in a fixed conformation and switch-I is more closed compared to the first group . This group is more frequently populated by G12R . In the state VI , switch-I is open and switch-II is in a mixed conformation between the fixed and perpendicular conformations . This state is clearly less populated by G12R compared to the other systems . The fourth group , where the switches appear in a perpendicular conformation , is frequent with G12V ( 12% ) , whereas the other mutants rarely visit this state ( 1% ) . Of all the mutants , G12D displays the most similar metastable state population distribution compared to wild-type ( Fig 5 , Table 1 ) . This is most evident in the most populated states ( states IV , VI , and VII ) , yet G12D also differs from wild-type in the less populated states ( I–IV ) . In contrast , the conformations of G12R are clearly shifted towards the fixed switch-II states , whereas G12V is shifted away from these states towards the perpendicular states . The results suggest that for G12V the shift among the states is due to the mutant’s lipophilic character , which may cause changes in solvent organization preventing specific switch-II conformations . Finally , it is exceptional that while wild-type does not populate the metastable states I and III at all , there are mutants ( G12R , G12V ) whose population in these states is significant ( about 6% ) . This summarizes the main message: the conformation distribution of KRAS mutants includes conformations not occupied by wild-type , and these conformations are also mutation specific .
Although frequently observed in cancer , not only is the basis for the specific frequencies of KRAS G12X mutations poorly understood [47] , but also the effects of these specific mutations on a molecular scale are not clear . To the best of our knowledge , this is the first study to assess KRAS G12X mutation probabilities , and to understand how they are associated with the observed mutation frequencies . Generally , the mutation frequencies have an explanatory basis . For instance , chemical characteristics of c . 35G explain the enrichment of the G12V mutations by oxidation . However , complex mutation distributions are displayed by the tissues , and we conclude that some of the observed frequencies cannot be explained simply by the mutation probability . For example , there is no clear explanation why , on average , a mutation occurs five times more probably in , e . g . , c . 35G than in c . 34G . One plausible explanation is that the 3D-environment in the DNA-sequence may aid the c . 35G mutations to evade DNA-repair mechanisms . In fact , the structures of DNA in complex with N-glycosylase/DNA lyase ( OGG1 ) , which is a base-excision repair enzyme for 8oxoG , exist in a catalytically active form for 8oxoG that is adjacent to guanine only in the 5’-position in -AGGT- sequences ( S12 Fig ) [40 , 48–51] . Correspondingly , this 5’G position in the KRAS sequence ( -TGGT- ) represents the c . 34G position , thus suggesting that the c . 34G position is more susceptible for DNA-repair . Nevertheless , this observation holds true only for the G>T transversion mutation . For the other mutations and their repair mechanisms , the positional bias needs to be investigated , especially for the G>A transition mutation , which holds the strongest bias in favor of c . 35G mutations ( >5x in all tissues ) . Furthermore , exceptions or an enhanced preference in c . 35G for specific mutations in particular tissues were observed . For instance , it seems that either the advantage for G12D or the disadvantage for G12S , or both , exists in the pancreas , given that there is a 28-fold preference for G>A mutations for c . 35G over c . 34G . Similarly , the G12R mutation displays an advantage in the pancreas , while G12A is perhaps disadvantageous , given that there is a 9-fold preference for c . 34G over c . 35G in the mutation probability in the G>C mutations . Altogether , these data suggest that specific mutations are advantageous or disadvantageous depending on the cellular and tissue environments . Therefore , we hypothesize that the biochemical and biophysical differences among mutants , resulting in signal-transducing differences , may explain , at least in part , the observed mutational bias . To gain insight into these observed discrepancies among the mutants on a molecular level , we carried out a comprehensive all-atom MD simulation study of all KRAS G12X missense mutants . We found that mutations have a profound effect on the dynamics of KRAS . In particular , we observed that the switches are highly dynamic . This conformational flexibility revealed through atomistic simulations is consistent with 31P-NMR spectroscopy studies of RAS proteins [21 , 52] , while the published KRAS crystal structures do not unlock this behavior . Even in our extensive analysis of the long-timescale simulation data the differences in the dynamics were not readily visible . This is not surprising given that even though the binding affinity of a specific mutant toward an effector protein is increased or diminished , the ability to bind still exists [8] . This suggests that the changes in protein dynamics are quite subtle and difficult to quantify . In our work , we unlocked this issue through the analysis of the simulation data using PCA , interaction network analysis and MSMs that indeed revealed the differences , not only between the wild-type and the mutants , but also between the mutants . In order to capture the subtle differences among the mutants , we kept our MD simulation systems realistic but sufficiently simple , enabling the extended simulation times close to 200 μs in total . Even though the HVR and the cell membrane were absent from our simulations , we recognize that these elements have a substantial influence on KRAS dynamics [53 , 54] and signaling [55] . We therefore cannot deduce whether some of the observed mutational effects attenuate or amplify through these factors . However , effects related to the cell membrane remain to be explored in follow-up studies . Importantly , we identified the hydrophobic hub interaction network that is able to convey the shifts in KRAS dynamics throughout the whole structure in an allosteric manner . The crystal structures of KRAS G12X mutants display only minor differences , but the lack of structural differences does not exclude the allosteric effect of the mutation [56] . The shift in the dynamics by G12X is able to occur via the closest hydrophobic hub V14 . As the wild-type KRAS has a flexible glycine residue in this position , a G12X mutation alters the dynamics of the neighboring residue A11 or the whole P-loop ( including both A11 and V14 ) . As this hydrophobic hub V14 is connected to a hydrophobic network , a local alteration in KRAS dynamics can be conveyed via the hydrophobic network to the other remote structural regions in KRAS in an allosteric manner . Supporting the fact that V14 is an important hub in the KRAS hydrophobic interaction network , a mutation in this position , V14I , is found to be one of the responsible mutations for the Noonan syndrome [57 , 58] and may also predispose tumor development [59] . Whereas the V14I mutation does not change the GTPase activity of KRAS , it displays similar affinity to RAF1 , as does also the G12V mutant [60] . Therefore , a mutation that has an influence in the dynamics of these hydrophobic hub interactions may have a dramatic influence in overall KRAS dynamics and thereby KRAS signaling . The most altered interaction pattern within the hydrophobic interaction network in all the hubs is observed with the G12A and G12S mutants . Surprisingly , the other mutations ( G12C , G12D , G12R and G12V ) are not radically different compared to the wild-type , although some alterations in the hub interactions are evident . However , even though a mutant , such as G12D , displays the same interaction frequencies as the wild-type , the characteristics of the interactions may still differ , as the exact characteristics of these interactions , unfortunately , cannot be derived from this analysis , only their frequencies . To highlight that the alteration in KRAS dynamics is also present with the mutants that display a minor shift in the hydrophobic hub interaction network compared to the wild-type , is the observed variability in the distribution among the metastable states of exceptional importance . The MSM confirmed the indirect effect of the mutation on the switch-region protein conformations and dynamics . As for MSMs one needs to have extended simulation data , we focused on the most important KRAS G12X mutants ( G12D , G12R and G12V ) . In crystal structures these dynamic metastable states are not observed . This is due to the fact that in the structures the switch regions are disordered , if there are no crystal contacts to the switches . Based on the MSMs , the G12D mutant follows the dynamics of the wild-type more closely than of the G12R and G12V mutants . Intriguingly , this is in line with the findings of the interaction network analysis , where the G12D displayed the most similar profile with the wild-type . In particular , our MD results show that the effects of KRAS G12X mutants are mutation-specific , and suggest that the observed changes in protein conformations and dynamics may alter protein activity [61] . We consider that the difference in the mutant dynamics , for instance the G12V dynamics with its inability to reach the metastable states I and V ( Fig 5 ) , may reflect the differences observed in the RAS effector protein binding [8 , 10] . In fact , simple protein complexes assemble generally via a single pathway [62] , and the observed metastable states may correspond to the first steps in the effector protein binding process . These states may be important for specific effector protein binding and pathway activation . However , based on the simulation data we were unable to distinguish if a putative effector protein ( s ) or a particular signaling pathway ( s ) is related to a specific metastable state . It needs to be clarified , if these states act as intermediate steps in the KRAS–effector protein association and play a role in the macromolecular recognition process , effecting the association kinetics of the complex formation [63] . Furthermore , multiple other aspects related to the altered dynamics may also affect KRAS mediated signaling . The altered dynamics may cause a conformational change of KRAS on the membrane , resulting in occluded conformation from a specific effector [23] , alter the dimer formation [64] , or affect KRAS nanoclustering [65] . Furthermore , the altered dynamics may affect the stability of a KRAS–effector complex itself that may lead to a more stable complex , resulting in binding with a longer lifetime , or conversely to a more unstable complex , resulting in faster dissociation . Altogether , this implies that the altered protein dynamics has an influence on the KRAS binding partner selection . As a crucial factor for KRAS dimerization , the intermolecular D154–R161 salt-bridges between the dimers were recently identified [66] . The KRAS dimerization is a GTP-dependent process [67] . Here we observed that with the GTP-bound wild-type , there is a shift in the dynamics of the putative α4-α5 dimer interface region , manifested by the more unstable intramolecular D154–R161 salt-bridge . This suggests that the more stable intramolecular salt-bridge within the monomer residues could hinder the dimer formation in GDP-bound KRAS , whereas due to the change in the dynamics on the site in GTP-bound KRAS , the more unstable intramolecular salt-bridge could promote the formation of intermolecular salt-bridges among these residues , and thus dimer formation . This intramolecular salt-bridge is again , regardless of the bound ligand , heavily distorted with the G12A and G12S mutants that have a major effect on the hydrophobic interaction networks . In general , the mutational frequency data combined with the observation from the simulations suggest that at least in the pancreatic cancer , where a KRAS mutation is a key initiator [68] , a major shift in KRAS dynamics is not tolerated . This fact is manifested by the low frequencies of G12A and G12S mutants in the pancreas . Moreover , the distorted dynamics could also offer an explanation why the G12S mutant is rarely observed even though it is a transition mutation . It has been clearly shown that the KRAS G12X mutation has an effect on the intrinsic GTPase activity and that it causes insensitivity for GAPs . However , it seems that the interpretation of the mutation effect on the oncoprotein’s behavior has been oversimplified . First , in specific tissues G12X mutation frequencies exhibit an inexplicable individual bias . Furthermore , the mutation inflicts individual changes in the protein dynamics , affecting the allosteric communication network that conveys the shift in dynamics to the remote sites within KRAS . Finally , this shift in protein dynamics may lead to modulated KRAS mediated signal transduction . We therefore suggest that altered dynamics among KRAS G12X mutants may promote the observed non-random frequencies in specific tissues . In order to establish successful therapies against mutant KRAS-harboring tumors , these discrepancies between the G12X mutants need to be reconsidered thoroughly . Concluding , KRAS G12X mutants are not equal , they are unique .
The simulations were conducted using the GROMACS package v . 4 . 6 with the ( all-atom ) OPLS-AA force field [69–72] . For the simulations , a high resolution ( 1 . 24 Å ) truncated ( 169/188 residues ) GDP-bound wild-type KRAS structure ( PDB ID: 4OBE ) [73] was selected , where most of the HVR is absent ( see S1 Fig ) . Mutant KRAS structures were generated from the wild-type structure using Maestro [74] . For GTP-systems the GDP was replaced with GTP . As a model for GDP and GTP ligands , the default OPLS-AA parameter set was used and the geometry optimization for GDP and GTP was conducted with Gaussian [75] , using the Hartree-Fock method and the 6–31+G** basis set . The partial charges for both ligands were derived from the electrostatic potential by performing a RESP fitting procedure with R . E . D . Tools IV [76 , 77] . Co-crystallized water molecules from the crystal structure were hold intact , with an exception of GTP-systems with one water molecule , which occupied the γ-phosphate binding position to Mg2+ . Water molecules were described with the TIP3P model [78] . In each system , the protein was solvated in a cubic box ( edges at least 1 . 3 nm from the protein ) , and the system was neutralized using a physiological ion concentration ( 140 mM ) of K+ and Cl- ions . After energy minimization , the system preparation was done in four stages to obtain properly equilibrated and different initial structures for replica simulations ( see S1 Table for details ) . The simulations were performed with periodic boundary conditions in the NpT ensemble . The V-rescale and Parrinello-Rahman methods were used for temperature ( 310 K ) and pressure ( 1 atm ) coupling , respectively [79 , 80] . The default 2 fs time step was used for integration of equations of motion . To preserve the length of all bonds , the LINCS algorithm was used [81] . For Lennard-Jones interactions and the real-space part of the particle mesh Ewald electrostatics , a cutoff of 1 . 0 nm was used [82] . Each system was simulated for 2 μs with 5–10 independent replicas , such that the individual system simulation time was 10–20 μs and the total simulation time was 170 μs . The RMSF were calculated using GROMACS tools [69] . Principal component analysis ( PCA ) was conducted for the backbone atoms by the GROMACS covariance analysis tools . For PCA , we discarded the first 300 ns and used only the last 1 . 7 μs of each simulated system , to remove the potential bias of the starting GDP crystal structure from the results . To reduce noise from the flexible terminal regions , we excluded from the analysis the first three residues from the N-terminal and the last five residues from the C-terminal . The PCA structures ( Fig 2 , S5–S9 Figs ) were obtained utilizing the GROMACS tool gmx_anaeig , and visualized with PyMOL 1 . 8 [83] using a pymol-script Modevectors [84] . The analysis of KRAS interaction networks ( the hydrophobic clusters and the salt-bridges ) was conducted with PyInteraph [44] and visualized with PyMOL [83] . This analysis was conducted for full trajectories and all residues were included in the analysis . Markov state model generation was conducted with PyEMMA 2 , following the general recommendations [46] . As an input , we used distances between the residues 12 , 32 , 34 , 36 , 48 , 59 , 62 , 64 , 67 , 105 , 122 , 126 , and 138 from the simulation trajectories . We selected these residues based on their functional importance in KRAS ( location in the interaction surface with the effector proteins ) , the results of PCA ( dynamical importance ) , or both . Furthermore , a slow linear subspace from this input was estimated by TICA [85] , as TICA highlights the slowest motions from simulations and is highly suitable for generation of a MSM [86] , using 40 ns as a lag-time , and two dimensions . Furthermore , the output of TICA was clustered using the k-means clustering , and the discretized trajectories from the clustering analysis were used to generate the BayesianMSM . The number of clusters in k-means were set as √N , as recommended in [46] . The microstates were grouped in seven metastable states by the Perron-cluster cluster analysis ( PCCA++ ) method [87] , based on the spectral analysis ( S13C Fig ) . The generated models were validated by two methods . First , we calculated the resulting timescales and found that the timescales were constant in the used 40 ns lag-time ( S13–S15 Figs ) . Furthermore , we conducted the Chapman-Kolmogorov test , which displayed that the model followed the expected estimates . The occupations of individual mutants in each metastable state ( Fig 5 ) were computed from their individual Markov models . The KRAS G12X mutation data was collected from COSMIC database v . 79 ( http://cancer . sanger . ac . uk/cosmic/ ) [2] . In our assessment of the KRAS G12X mutation probability , we included only single nucleotide substitutions . This choice was made based on the fact that more complex mutations and their probabilities ( e . g . , adjacent double substitutions ) are not predictable with the existing knowledge . Fisher’s exact test was used to analyze the specific differences in mutation frequencies . | The oncogene KRAS is frequently mutated in various cancers . When the amino acid glycine 12 is mutated , KRAS protein acquires oncogenic properties that result in tumor cell-growth and cancer progression . These mutations prevail especially in the pancreatic ductal adenocarcinoma , which is a cancer with an exceptionally dismal prognosis . To date , there is a limited understanding of the different mutations at the position 12 , also regarding whether the different mutations would have different consequences . These discrepancies could have major implications for the future drug therapies targeting KRAS mutant harboring tumors . In this study , we made a critical assessment of the observed frequency of KRAS G12X mutations and the underlying causes for these frequencies . We also assessed KRAS G12X mutant discrepancies on an atomistic level by utilizing state-of-the-art molecular dynamics simulations . We found that the dynamics of the mutants does not only differ from the wild-type protein , but there is also a profound difference among the different mutants . These results emphasize that the different KRAS G12X mutations are not equal , and thereby they suggest that the future research related to mutant KRAS biology should account for these observations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"crystal",
"structure",
"markov",
"models",
"condensed",
"matter",
"physics",
"multivariate",
"analysis",
"mutation",
"mathematics",
"statistics",
"(mathematics)",
"mutation",
"databases",
"crystallography",
"research",
"and",
"analysis",
"methods",
"solid",
"state",
"physics",
"mathematical",
"and",
"statistical",
"techniques",
"principal",
"component",
"analysis",
"biological",
"databases",
"salt",
"bridges",
"chemistry",
"exocrine",
"glands",
"probability",
"theory",
"physics",
"biochemistry",
"biochemical",
"simulations",
"point",
"mutation",
"electrochemistry",
"anatomy",
"database",
"and",
"informatics",
"methods",
"pancreas",
"genetics",
"endocrine",
"system",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"statistical",
"methods"
]
| 2018 | Assessment of mutation probabilities of KRAS G12 missense mutants and their long-timescale dynamics by atomistic molecular simulations and Markov state modeling |
[PSI+] is an amyloid-based prion of Sup35p , a subunit of the translation termination factor . Prion “strains” or “variants” are amyloids with different conformations of a single protein sequence , conferring different phenotypes , but each relatively faithfully propagated . Wild Saccharomyces cerevisiae isolates have SUP35 alleles that fall into three groups , called reference , Δ19 , and E9 , with limited transmissibility of [PSI+] between cells expressing these different polymorphs . Here we show that prion transmission pattern between different Sup35 polymorphs is prion variant-dependent . Passage of one prion variant from one Sup35 polymorph to another need not change the prion variant . Surprisingly , simple mitotic growth of a [PSI+] strain results in a spectrum of variant transmission properties among the progeny clones . Even cells that have grown for >150 generations continue to vary in transmission properties , suggesting that simple variant segregation is insufficient to explain the results . Rather , there appears to be continuous generation of a cloud of prion variants , with one or another becoming stochastically dominant , only to be succeeded by a different mixture . We find that among the rare wild isolates containing [PSI+] , all indistinguishably “weak” [PSI+] , are several different variants based on their transmission efficiencies to other Sup35 alleles . Most show some limitation of transmission , indicating that the evolved wild Sup35 alleles are effective in limiting the spread of [PSI+] . Notably , a “strong [PSI+]” can have any of several different transmission efficiency patterns , showing that “strong” versus “weak” is insufficient to indicate prion variant uniformity .
Prions in yeast are a new form of gene , composed of proteins instead of nucleic acids [1] . As such , their inheritance , mutation and segregation are not expected to follow the same rules as the majority DNA/RNA genes . The [PSI+] prion was first recognized as a non-chromosomal genetic element enhancing the read-thru of the premature termination codon in ade2-1 [2] . Its unusual genetic properties led to its identification as a prion of Sup35p [1] , a subunit of the translation termination factor [3] , [4] , specifically an amyloid form ( β-sheet-rich filamentous polymer of protein subunits ) of the normally soluble Sup35p [5]–[9] . In the amyloid form , the protein is largely inactive , resulting in increased read-through of termination codons . Yeast prions are important models for mammalian prion diseases , and for amyloid diseases in general . Sup35p consists of C , an essential C-terminal domain ( residues 254–685 ) , responsible for the translation termination function [3] , [4] , [10]; N , an N-terminal domain necessary for prion propagation ( residues 1–123 ) [10] that normally functions in the general mRNA turnover process [11]–[15] and functionally interacts with Sla1p [16]; and M ( residues 124–253 ) , a middle charged region that is also implicated in prion propagation [17]–[20] . In the infectious amyloid form , the N domain , and probably part of the M domain , is in an in-register parallel β-sheet form , with folds in the sheet along the long axis of the filament [21] , [22] . Prions can often be transmitted between species , as was first recognized by infectivity of sheep scrapie brain extracts for goats [23] . However , cross-species transmission is inefficient ( or completely blocked ) as a result of sequence differences between the donor and recipient prion proteins [24] . This phenomenon is called the species barrier , and has also been observed in yeast prions [19] , [25]–[31] . Wild isolates of S . cerevisiae also show considerable sequence variation in Sup35p sequence [20] , [32] , and these sequence differences produce barriers to transmission of [PSI+] [20] , presumably evolved to protect cells from the detrimental , even lethal , effects of this prion [33] , [34] . A single prion protein can propagate any of a number of prion variants ( called ‘prion strains’ in mammals ) , with biological differences due to different self-propagating conformations of the amyloid [9] , [35] , [36] . Although there is evidence for conformational differences between prion variants , the nature of those differences is not yet known . In yeast , prion variants differ in intensity of the prion phenotype , stability of prion propagation , interactions with other prions , response of the prion to overproduction or deficiency of various chaperones , and ability to cross species barriers [30] , [31] , [37]–[41] . Different variants arise during prion generation as a result of some stochastic events occurring in the initial formation of the prion amyloid . Generally , prion variant properties are rather stable , even during propagation in a species different from that in which the prion arose ( e . g . [42] ) . In a previous report , we demonstrated transmission barriers between Sup35 alleles from wild strains of S . cerevisiae , an ‘intraspecies barrier’ . These intraspecies barriers are of particular interest since they must operate in nature , when S . cerevisiae strains mate among themselves . Interspecies matings are less efficient than intraspecies matings ( e . g . , [43] ) , and diploids formed produce almost no viable meiotic spores [44] , [45] . In most cases , the intraspecies barriers were incomplete , with occasional transmission between strains with different Sup35 sequences . Were the prions transmitted the same variant as the original , or were they prion ‘mutants’ , heritably changed in their properties ? Under selective conditions , prion variant properties may change , a phenomenon first demonstrated in mice [46] and also known in yeast [30] , [47] . Selection in the presence of a different prion protein sequence , or a drug interacting with amyloid could induce a new prion by inaccurate cross-seeding , and reflect generation of a new prion , rather than propagation of one of several sub-variants already present . Here , we examined variation in prion properties under non-selective conditions , finding evidence for the existence of a ‘cloud’ of variants with stochastic fluctuation .
Wild SUP35 alleles fall into three groups: the ‘reference’ sequence is essentially that of laboratory strains; Δ19 has a 19 residue ( 66–84 ) deletion in the prion domain; E9 is representative of a group with N109S and several polymorphisms in the M domain [20] . Three independent prion variants of the E9 Sup35p ( E9A , E9F , E9G ) were selected in strain 4828 ( Table S1 ) . We tested the transmission of these variants by cytoduction to strain 4830 expressing E9 itself , Δ19 or reference Sup35 . None of these variants were transmitted well into the strain containing the Δ19 Sup35 polymorph . However , two variants ( A , G ) propagated very poorly with reference Sup35 sequence , while the other variant ( F ) was able to efficiently transmit the prion to the reference sequence ( Table 1 , p<10−10 ) . This indicates that intraspecies transmission barriers are variant-specific . Two of the few E9G→Δ19 cytoductants that were [PSI+] ( Table 1 ) were tested for transmission to strains with different Sup35 polymorphs ( Table 2 ) . Each had lost the transmission specificity and were now able to transmit the prion more efficiently into all sequences ( Table 2 , p<10−10 ) , unlike two [PSI+] isolates initially selected in cells expressing Δ19 , which propagated poorly to E9 or reference [20] ( p values between 10−10 and . 002 ) . This again shows the prion variant specificity of transmission barriers . Note that these two E9G→Δ19 cytoductants differ in that [PSI+E9G]Δ19A was white ( a strong [PSI+] ) while [PSI+E9G]Δ19B was pink ( a weak [PSI+] ) . This indicates that either the original E9G was a mixture of two prions or that new prion variants were selected by the difficulty of transmission into Δ19 . An E9G→ref cytoductant from Table 1 , similarly analyzed , showed ready propagation into reference ( 100% , p<10−10 ) and the original E9 sequence from which the prion originated ( 69% ) , but only poor transmission to the Δ19 sequence ( Table 2 ) . This result differs from a [PSI+ref]ref ( originating and propagating in the ref sequence ) which propagates poorly into E9 ( 19% , p<10−8 ) [20] , again showing prion variant dependence of prion transmission . As expected the E9G prion transmitted to another yeast strain with the E9 Sup35 had similar propagation characteristics to the original [PSI+E9G] ( compare Table 1 and Table 2 ) . The [PSI+ref]ref in strain 779-6A was transmitted to cells with the other Sup35p polymorphs and , as expected , transmission was limited ( Table 3 ) . When [PSI+] cytoductants were examined for stability on extensive further mitotic growth , we found that the [PSI+ref]ref cytoductants were fully stable , while the [PSI+ref]Δ19 were significantly less stable and [PSI+ref]E9 cytoductants even less so . Nonetheless , stability was sufficient that [PSI+ref]Δ19→Δ19 and [PSI+ref]E9→E9 cytoductions showed >90% transmission ( Table 3 ) . The variant-dependence of transmissibility was again evident in cytoduction of [PSI+ref]ref in strain 779-6A [48] to cells with the other Sup35p polymorphs ( Table 3 ) . This variant originated in the reference sequence , but when transferred to Sup35Δ19 , is then transferred well to either the reference or the Δ19 Sup35s , but very poorly to E9 ( Table 3 ) . In contrast , either of two E9-originating prions in a Δ19 host ( [PSI+E9G]Δ19 ) , transfer well to all polymorphs ( Table 2 , p<10−10 ) . The [PSI+ref]E9 transfers well to both reference and E9 sequences ( Table 3 ) , like [PSI+E9F] , but unlike two other prions originating in E9 ( Table 1 , p<10−10 ) . As expected , the prion originating in E9 and transmitted to E9 , or that originating in the reference sequence and transmitted to the reference sequence , each maintain their original properties . Having transferred [PSI+ref] to each of the Sup35 polymorphs , we transferred them back to the original host ( cured of [PSI+] ) and re-examined their transmission properties to see if they had changed as a result of their experience ( Table 4 ) . The original [PSI+ref] transmitted poorly to either Δ19 or E9 hosts , but the ‘experienced’ prions all transmitted better to E9 than the original , indicating selection of a ‘mutant’ prion ( Table 4 , p< . 002 , 10−6 , 10−10 ) . Moreover , the prion that passed through Δ19 could transmit 91% to another Δ19 ( Table 3 ) , but when passed back to the reference sequence , only transmitted 20% to Δ19 ( Table 4 ) . Similarly , the prion passed through E9 , and able to transmit to another E9 host at 92% ( Table 3 ) , once passed back to the reference host could only transmit 46% to E9 ( Table 4 ) . These results indicate that the predominant variant has changed . But is this change due to mistemplating as the prion passes from Sup35 molecules with one sequence to those with a different sequence , or is there an ensemble of variants present within the population that can be selected based on the specific selection pressure , to be visible with a specific transmission phenotype ? If the population contains an array of prion variants from which one or another can be selected , one might expect these to segregate during mitotic growth , much as differently marked plasmids sharing the same replicon or mitochondrial genomes will segregate mitotically , even without exposure to a selective condition . In contrast , if the changes in prion variant are due solely to mistemplating when a prion crosses a transmission barrier to a different sequence , then the transmission pattern should not change substantially even after extensive propagation in the original strain . We designed this experiment to separate the mitotic segregation phase , in which there was no change of Sup35p polymorph , from the transmission phase , in which the test of prion variant is then made by cytoduction to the three Sup35 polymorphs . We subcloned single colonies of the 779-6A [PSI+ref] yeast strain ( reference Sup35p ) without selection on ½ YPD plates for at least 75 generations . Table 5 illustrates our surprising result , that many subclones had transmission profiles considerably different from the parent strain 779-6A . This indicates that there is an ensemble of variants or a prion cloud that has different transmission profiles . We have classified these variants as being type A if they transmit well into reference sequence but poorly into Δ19 and E9 sequences . Type B transmits well into reference and E9 sequences , but poorly into the Δ19 sequence . Type C transmits well into Δ19 and reference sequence , but poorly into the E9 sequence and type D transmits well into all sequences . From this subcloning we now had yeast strains that were carrying prion variants of type B ( Y1 ) , type C ( Y2 ) and of type D ( Y5 ) . These strains repeatedly display these propagation patterns even after many months in frozen stocks . We then wanted to determine if we had now isolated single variants within the original ensemble so each of three clones , of transmission types B , C and D , were subcloned an additional 75 generations on ½ YPD plates with ten clones of each tested as before . To our surprise these extensively grown subclones of each of the three types still produced clones with an ensemble of prion variants ( Table 6 ) . Even the Y1 strain , which did not initially propagate into the Δ19 sequence , produced subclones with a variety of transmission profiles . To determine if the appearance of different predominant variants was due to some unrecognized selective pressure on these strains while propagating on ½ YPD plates , the subcloning was performed in liquid YPD media maintaining the culture in exponential growth phase throughout . Once cell density reached 0 . 3 absorbance units at 600 nm the cultures were diluted , transferring only 1000 cells to a fresh culture , a process continued for at least 84 generations . Even under exponential growth phase ( Table S2 ) , an array of transmission profiles was observed similar to that in Table 6 . The presence of changed transmission patterns in a majority of the clones without any selection having been applied made it clear that the changes were not due to a chromosomal mutation . Nonetheless , we tested for such a chromosomal change by curing [PSI+] from Y5 by growth on guanidine , and cytoducing cytoplasm from Y1 , Y2 or Y5 into strain 4830 and then 8 cytoductants from each were cytoduced into a rho° derivative of the cured Y5 ( Table S3 ) . These cytoductants were then cytoduced into recipients each carrying one of the three SUP35 polymorphs . In each case the transmission pattern followed that of the original Y1 , Y2 , or Y5 donor of cytoplasm , rather than the Y5 pattern of the recipient ( Table S3 ) , confirming that the change was due to a new variant of [PSI+] and not a chromosomal change . The frequency with which the transmission pattern changed without selection or protein over expression is orders of magnitude higher than for the generation of any new prion , and the fact that the change is one of changing the specificity of transmission to different Sup35p polymorphs proves that it is indeed a change of [PSI+] , and not the generation of some other prion . To further test the presence of an ensemble of prion variants , one subclone of Y1 , which had the same profile as the parent , not being able to transmit into the Δ19 sequence , was subcloned for an additional 75 generations . As shown in Table S4 , subclones were obtained with various profiles some with very good transmission into the Δ19 sequence containing strain . These results indicate that a single variant had not been selected and that an ensemble or cloud of prion variants must exist with a dynamic propagation pattern under non-selective conditions . Each isolate has a specific transmission pattern , even after frozen storage for many months ( Table S5 ) . We infer that during growth , events must allow for a stochastic shift of the ensemble to allow for isolation of variants with specific reproducible transmission patterns . [PSI+] is rare in wild strains [33] , but was found in 9 of 690 wild isolates [49] , each expressing the reference Sup35 ( ref . [49] and Amy Kelly , personal communication ) . How do these wild [PSI+] variants respond to the intraspecies barriers we previously reported [20] ? We used both reference sequence and E9 sequence Sup35 fused to GFP and could see dots in the reported wild [PSI+] strains 5672 , UCD#885 , UCD#978 and UCD#2534 , though infrequently , but not in strains UCD#521 , 587 , 779 , 824 , 939 ( Figure S1 ) . To test these strains genetically for nonsense suppression , we crossed the wild strains with strain 4972 ( Table S1 ) , carrying the [PSI+]-suppressible ade1-14 marker , and tested dissected tetrads to determine if ade1-14 is suppressed . We found that for seven of the wild strains , ade1-14 was weakly suppressed in the segregants , and this suppression could be cured by growth in the presence of guanidine , which is known to cure the [PSI+] prion . We could not obtain tetrads from diploids formed with strain UCD# 978 and strain 5672 gave poor spore germination . The transmission of the wild [PSI+] isolates into cells expressing the Sup35 polymorphs in strains 4828 and 4830 by cytoduction is shown in Table 7 . The wild [PSI+] strains transmit well into the reference sequence , but most showed poor transmission to one or both of the Δ19 or E9 sequences ( Table 7 ) . All four transmission patterns were observed ( Table 7 ) , but all of the isolates were ‘weak’ [PSI+] ( Figure 1B ) . Thus , each of the strains tested transmitted [PSI+] even though several did not show dots with Sup35NM-GFP . Of course , their presumed independent origin means that these wild isolates are not derived from one prion cloud . Variants of [PSI+] may be weak or strong in phenotype , stable or unstable in propagation , and have various responses to deficiency or over expression of chaperones or other cellular components , have different patterns of ability to cross species barriers , and , as shown here , to cross intraspecies transmission barriers . To what extent these various parameters are correlated is largely unknown . We tested the several prion variants derived from the [PSI+] in strain 779-6A with different transmission patterns for their ‘strong’ vs ‘weak’ character ( Figure 1A ) . We note that , with identical chromosomal genotype , they are indistinguishable in the ‘strength’ parameter in spite of having substantially different transmission properties . As noted above , the wild [PSI+] variants are indistinguishably ‘weak’ , but have different transmission patterns to the Sup35 polymorphs .
Yeast prion variants are distinguishable based on intensity of the prion phenotype , stability or instability of prion propagation , sensitivity of prion stability to overproduction or deficiency of several chaperones and other cellular components and ability to overcome barriers to transmission between species [30] , [31] , [37]–[41] – or even within species , the last documented here for transmission across the barriers found in wild strains of S . cerevisiae . Yeast prion amyloids are all folded parallel in-register β-sheet structures [21] , [50] , [51] , but within this architectural restraint , different prion variant structures are proposed to vary in the extent of the β-sheet structure ( how much of the N and M domains are in β-sheet ) , the locations of the folds in the sheets and the association of protofilaments to form fibers . We find that separation of prion variants based on sensitivity to intra-species barriers cuts across separation based on ‘strong’ vs ‘weak’ assessment of strength of prion phenotype . The four transmission variant types derived from the [PSI+] in strain 779-6A were all strong [PSI+] , like the parent prion . Interestingly , the prions in wild strains were all weak [PSI+] , presumed to arise independently and thus not part of the same ‘prion cloud’ , but fell into the same four transmission variant types . Likewise , two similarly ‘weak’ [PSI+] variants showed different transmission across a barrier set up by deletions in the prion domain [52] . These results show that prion variant uniformity is not demonstrated by showing uniformity of a single property ( for example , colony color ) . It is unlikely that the variation in transmission barriers observed are due to a prion other than [PSI+] because the sequences of Sup35p are involved , and no yeast prion is known to arise at a frequency high enough to explain our results . After crossing an intraspecies barrier , we find that the [PSI+ref] examined is unstable in its new host , emphasizing the effectiveness of these barriers . We also find that the rare [PSI+] prions found in wild strains are , in most cases , sensitive to the intraspecies barriers , suggesting that these barriers have evolved to protect yeast from the detrimental effects of this prion . The [PSI+] in strain 779-6A , with the reference Sup35p sequence , showed a reproducible strong preference for the reference sequence , transferring only very inefficiently to the Δ19 or E9 Sup35 backgrounds . However , simple mitotic growth of this strain resulted in the mitotic segregation of at least four variants distinguished by their abilities to cross intraspecies barriers . These variants were stable and reproducible with limited expansion of the corresponding clones , but following many generations of growth , each of those tested gave rise again to the same four general classes of subclones . Prion mutation is well documented in mammals and in yeast under selective conditions [30] , [46] , [47] , [53] , and Weissmann's group has suggested that prions resistant to a drug can arise during prion propagation in tissue culture cells in the absence of the drug [54] , [57] . We observe changes in the predominant prion variant under non-selective conditions in vivo . Selection only happens during the test , when cytoplasm is passed by cytoduction from the subclones to be tested to the recipient expressing one of the three Sup35p polymorphs . A new prion variant , recently described by Sharma and Liebman [55] , may represent a phenomenon similar to that described here . Certain induced [PSI+] clones continually gave off subclones that were a mixture of strong and weak variants , what the authors called “unspecified [PSI+]” . Although multiple de novo prion generation events in forming amyloid in vitro result in multiple prion variants on transfection into yeast , even a [PIN+] cell generates [PSI+] clones too rarely to explain our results as de novo prion generation . Rather , mis-templating must be the mechanism of generation of variant diversity that we are observing . Our results imply that there must be a finite rate of amyloid mis-templating that is not due to a mismatch of two prion protein sequences . In spite of extensive purification by mitotic growth and subcloning , we were unable to obtain a prion variant that was completely stable in its transmission pattern to polymorphs . These results are consistent with the ‘prion cloud’ hypothesis [56] , [57] , in which it is supposed that even a prion variant purified by end-point titration consists of a major variant as well as an array of minor variants . This production of new prion variants during non-selective growth is analogous to the generation of RNA virus mutants during viral replication ( reviewed in ref . [58] ) , in which a cloud of sequence variants accumulate because of the error-prone nature of RNA-dependent RNA polymerases . The segregation of a mixed prion population could be considered analogous to the segregation of differently marked plasmids with the same replicon . The latter situation has been carefully examined by Novick and Hoppenstadt [59] , who find that the fraction of cells remaining with a mixture of plasmids is H = H0 [ ( N−1 ) ( 2N+1 ) / ( 2N−1 ) ( N+1 ) ]n , where H0 is the starting fraction of mixed cells , N is the copy number of the plasmid , and n is the number of generations [59] . Random replication of plasmids and equal partition at mitosis is assumed . One result of this treatment is that after N generations , H≈0 . 36 H0 . The copy number in the case of yeast prions might be taken as the ‘seed number’ determined by the methods developed by Cox et al . [60] , found to be ∼20–120 for the strains examined . The assumption of equipartition is probably not accurate here , since yeast daughter cells are smaller than mother cells [60] . Moreover , the sticky nature of amyloids might suggest that progeny filaments might stick to parent filaments exaggerating this effect . We have propagated our [PSI+] strains for a number of generations comparable to the presumed copy number , so segregation of different prions is not surprising . However , we find that even when we have apparently purified a variant , further non-selective growth and subcloning leads to further appearance of the full range of variants among the progeny ( Figure 2 ) . This indicates that we are not only observing segregation , but also the ( repeated ) generation of variants during growth . While varying with respect to transmission , they remain ‘strong’ variants , suggesting that the structural differences responsible for this transmission barrier differ from those involved in the strong vs . weak differences . King has shown that residues 1–61 are sufficient to propagate strong vs weak prion strains [8] , [61] , but the sequence differences among the Sup35 polymorphs are outside this area , and transmission variants may thus largely differ in the region C-terminal to the 1–61 area , perhaps a region with more variable structure . Other studies have indicated effects of this region on propagation of some prion variants [52] , and β-sheet structure of Sup35NM amyloid extends throughout N and even into M [21] , [22] .
We refer to the standard laboratory yeast sequence [62]–[64] as the ‘reference sequence’ . Two common sequence polymorphs found within the wild population were used . The first , with deletion of 19 amino acids from residues 66 to 84 and the G162D change , is referred to as Δ19 , and the other includes N109S , G162D , D169E , P186A , T206K , H225D and is denoted E9 [20] . A prion originating with the Sup35p sequence of strain E9 , for example , but being propagated in a strain expressing only the reference sequence will be designated [PSI+E9]ref , in analogy with similar nomenclature for [URE3] [31] . Cytoductants ( see below ) generated with strain A as donor and strain B as recipient are denoted A→B . They have the nuclear genotype of strain B and the cytoplasmic genotype of both A and B . In an abuse of language , we often use “[PSI+E9]ref was transferred to Sup35 Δ19” to mean “[PSI+E9]ref was transferred to cells expressing Sup35 Δ19” . Sup35p is a subunit of the translation termination complex , and the incorporation of a large proportion of Sup35p into the prion amyloid filaments makes it inactive , resulting in increased read-through of termination codons . This is measured by read-through of ade2-1 , with an ochre termination codon in the middle of the ADE2 gene . In addition to ade2-1 , strains carry the SUQ5 weak suppressor mutation , which leaves cells Ade- unless the [PSI+] prion is also present [2] . The strains used are listed in Table S1 . Plasmids used containing reference , Δ19 or E9 sequences were generated as described [20] . All yeast media and plates contained 20 µM copper sulfate unless noted . Rich and minimal media ( YPAD and SD ) are as described [65] . Only nutrients required by the strains used in a given experiment were added to minimal plates . Cytoplasm may be transferred from one strain to another utilizing the kar1-1 mutation [66] , defective for nuclear fusion . Cells fuse , but the nuclei do not fuse , and nuclei separate at the next cell division . However , cytoplasmic mixing has occurred , and so a genetic element ( prion or mitochondrial DNA ) present in one strain ( identified by its nuclear genotype ) will be transferred to the other . We use transfer of mitochondrial DNA as a marker of cytoplasmic transfer , and score prion transfer . Reference , Δ19 or E9 sequence plasmids were transformed into both laboratory strains 4828 and 4830 , loss of p1215 ( URA3 SUP35C ) was selected by growth on 5-fluoroorotic acid media and Ade- transformants were made rho° by growth on YPAD containing 1 mg/ml ethidium bromide . Donor and recipient strains at high density were mixed in water at a ratio of about 5∶1 , and the mixture was spotted onto a YPAD plate . After 18 hours at room temperature , the mating mix was streaked for single colonies on media selective against growth of the donor strain . Clones are shown to be cytoductants by their growth on glycerol and failure to grow on media selective for diploids . As further tests of a sample confirm , Ade+ cytoductants are judged to have received and propagated [PSI+] . [PSI+] Strain 779-6A [48] was streaked to single colonies on ½ YPD media and twelve colonies were selected , named Y1-Y12 . These isolates were streaked to single colonies three additional times , each time selecting just one colony for further propagation . From the third plate a single colony was selected and expanded on ½ YPD , and cells from this plate were used for cytoduction . From dilution tests there are approximately 2×107 cells per colony , indicating a total of at least 75 generations of growth of clones Y1-Y12 before cytoduction . Additional subclones were handled in the same manner with only ten colonies selected from the initial ½ YPD plate . In experiments to rule out selection during stationary growth phase , subclones of Y1 and Y2 were grown in a 125 ml Erlenmyer flask containing 25 ml of liquid YPD medium . When A600 reached 0 . 3 , the culture was diluted , transferring 1000 cells of each to a fresh flask . These subclones were propagated in exponential phase for 84 generations and were then streaked for single colonies on ½ YPD plates . After one day of growth on ½ YPD , 10 subclones were selected for each of Y1 and Y2 , expanded and tested for transmission via cytoduction . Strains reported to be [PSI+] [49] were obtained from the UC Davis Department of Viticulture and Enology culture collection . The cultures were first tested to determine if dots were visible using either reference sequence Sup35NM-GFP pDB65 or E9 sequence Sup35NM-GFP pDB81 [20] . Images were obtained with a Nikon Eclipse TE2000-U spinning disc confocal microscope with 100× NA 1 . 4 Nikon oil lens with 1 . 5× magnifier and captured with a Hamamatsu EM-CCD ImagEM digital camera with IPLab version 4 . 08 . Wild strains were sporulated and spores were crossed on rich medium with strain 4972 selecting G418-resistant prototrophs . The diploids formed were again sporulated and tetrads were dissected for each wild strain except for strain 978 , whose diploid with 4972 would not sporulate . Ade positive segregants were tested for guanidine curing using two successive streaks on YPAD with 5 mM guanidine . MATα strains were cytoduced into strain 4830 carrying pRS316 ( URA3 ) for selection . Lys2 mutants of MATa strains were selected on plates with DL-α-aminoadipic acid as a nitrogen source [67] . Selected strains were retested for Ade positive growth and curing and cytoduced into strain 4828 . The cytoduction data follows the binomial distribution , because each data point expresses two alternative results , transmission of [PSI+] or failure of its transmission . However , because of the large number of observations , the results should be approximately normally distributed . We want to calculate the probability that two sets of data are due to chance . Let p1 and p2 be the observed proportions of transmission in cytoductions 1 and 2 , and ni the number of cytoductants tested in each experiment . Let p = ( p1n1+p2n2 ) / ( n1+n2 ) be the average of the proportion of transmission in the two experiments . The estimated standard error of the difference between the two proportions isThe null hypothesis is that cytoductions 1 and 2 are samples from the same population with transmission efficiency p and standard error S . Then the expected proportions are expected to be the same and their difference is expected to be zero . [ ( p1−p2 ) −0]/S = z = the number of standard deviations that the observed difference in proportions differs from the expected difference ( 0 ) . The frequency of “z” being greater or equal to the observed value ( assuming the null hypothesis ) is obtained from a table of the normal distribution . The calculated “p values” are shown in the tables and at appropriate points in the text . Cytoductants examined have been treated as independent since the chance that they represent sister cells is close to zero . This is because cytoductant mixtures were incubated at 20C where the cells divide slowly and because only about 30 cells were examined from several million in the zygote mixture on each plate . | The [PSI+] prion ( infectious protein ) of yeast is a self-propagating amyloid ( filamentous protein polymer ) of the Sup35 protein , a subunit of the translation termination factor . A single protein can form many biologically distinct prions , called prion variants . Wild yeast strains have three groups of Sup35 sequences ( polymorphs ) , which partially block transmission of the [PSI+] prion from cell to cell . We find that [PSI+] variants ( including the rare [PSI+] from wild yeasts ) show different transmission patterns from one Sup35 sequence to another . Moreover , we find segregation of different prion variants on mitotic growth and evidence for generation of new variants with growth under non-selective conditions . This data supports the “prion cloud” model , that prions are not uniform structures but have an array of related self-propagating amyloid structures . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"genetics",
"epigenetics",
"biology",
"genetics",
"of",
"disease",
"genetics",
"and",
"genomics"
]
| 2013 | The [PSI+] Prion Exists as a Dynamic Cloud of Variants |
Eukaryotic chromosomes initiate DNA synthesis from multiple replication origins . The machinery that initiates DNA synthesis is highly conserved , but the sites where the replication initiation proteins bind have diverged significantly . Functional comparative genomics is an obvious approach to study the evolution of replication origins . However , to date , the Saccharomyces cerevisiae replication origin map is the only genome map available . Using an iterative approach that combines computational prediction and functional validation , we have generated a high-resolution genome-wide map of DNA replication origins in Kluyveromyces lactis . Unlike other yeasts or metazoans , K . lactis autonomously replicating sequences ( KlARSs ) contain a 50 bp consensus motif suggestive of a dimeric structure . This motif is necessary and largely sufficient for initiation and was used to dependably identify 145 of the up to 156 non-repetitive intergenic ARSs projected for the K . lactis genome . Though similar in genome sizes , K . lactis has half as many ARSs as its distant relative S . cerevisiae . Comparative genomic analysis shows that ARSs in K . lactis and S . cerevisiae preferentially localize to non-syntenic intergenic regions , linking ARSs with loci of accelerated evolutionary change .
The first step in the initiation of DNA synthesis in eukaryotes is the binding of the Origin Recognition Complex ( ORC ) to the replication origin [1] . However , replication origins have very different characteristics depending on the organism and sometimes the locale within the genome of an organism [2] , [3] . In the budding yeast Saccharomyces cerevisiae , replication origins are short , modular , cis-acting sequences that support the autonomous replication of extrachromosomal plasmids [4]–[6] . The S . cerevisiae ORC binds to a T-rich 17 bp motif termed the ARS Consensus Sequence ( ACS ) that is sometimes supported by interaction with nearby B-elements [1] , [7] . In the fission yeast Schizosaccharomyces pombe , replication origins are degenerate A/T-rich sequences which ORC binds in a stochastic manner [8]–[11] . Intergenic sequences 500 bp in length that contain >75% A-T are potential initiation sites for DNA replication . A genome-wide map based on high resolution tiling microarray indicates that there are 401 strong origins and 503 putative weak origins spaced at an average length of ∼14 kb in clusters throughout the S . pombe genome [12]–[14] . In this respect , the genomic landscape for replication initiation in S . pombe resembles that of insect and amphibian embryonic cells [15] , [16] . In mammals , some replication origins consist of physically dispersed combinatorial cis elements that serve as defined ORC binding sites while others are degenerate sequences that ORC recognizes stochastically [17] . Because of the defined nature of its replication origins , S . cerevisiae is the only eukaryote whose entire repertoire of replication origins have been identified , but not without extensive efforts [5] , [14] , [18]–[23] that took almost three decades . As it is , the S . cerevisiae replication origins genome map is less than 100% complete as new ARSs are still being identified with each new study ( this study and [24] ) . Recent work has also made strides towards constructing a genome map of replication origins in another pre-whole-genome duplication ( pre-WGD ) budding yeast Lachanacea kluyveri [25] . However , the great divergence between these pre-WGD species inhibits meaningful comparative genomic analyses . Studies of additional yeast species that bridge these wide divides would greatly enhance informative evolutionary comparisons . In this study , we have devised an approach that combines computational prediction and experimental validation to study the genome-wide location of replication origins of Kluyveromyces lactis . The motivation for this study is to investigate the evolutionary changes that lead to the divergence of replication origins in eukaryotes . To facilitate our goal , we developed new tools to map and study replication origins at a genome-wide scale with greater efficacy . We chose K . lactis as our model because it is distant from S . cerevisiae – they diverged prior to WGD in the evolution of yeast species about 100 million years ago [26] , [27] , and its emergence is dated between that of S . pombe and L . kluyveri in the fungal tree of life . Another advantage of working with this species is that previous studies of three K . lactis autonomously replicating sequences ( ARSs ) showed that they are defined sequences , but with features that are significantly different from S . cerevisiae and S . pombe ARSs [28]–[30] . These factors make the K . lactis genome an excellent model for our origin mapping study since it encompasses both similarities and differences with other established models . This study focuses on further characterizing the DNA replication initiation sites of K . lactis at a genomic scale .
We began our study of the genome-wide locations of K . lactis ARSs by constructing genomic libraries of K . lactis in a non-replicating URA3 vector and screening these libraries for autonomous replication function in a uraA ( the K . lactis equivalent of ura3 ) auxotrophic strain of K . lactis [31] ( Figure 1A ) . Efficient plasmid replication was required for colony growth on medium lacking uracil , therefore only ARS bearing plasmids yielded colonies . The plasmids were isolated from yeast colonies , sequenced , and used to re-transform yeast to confirm ARS function . Sixty-nine unique ARS-containing fragments were identified from this non-saturating screen . Due to the random nature of the screen , the ratio of novel to rediscovered ARSs rapidly declined after several rounds of screening . To overcome the diminishing returns of our screen we adopted a more targeted , iterative , approach to identify additional ARSs . Each cycle of the procedure consisted of compiling a list of ∼20 new candidate ARSs whose functionality would then be manually verified ( Figure 1B ) . In a manner somewhat reminiscent of PSI-BLAST [32] , the prediction step of each cycle was made based on the information gained from all previous cycles including the initial set of 69 K . lactis screened ARSs . Specifically , applying the de-novo motif finder GIMSAN [33] to the set of all hitherto verified K . lactis ARSs , we defined a new ACS Position Weight Matrix ( PWM ) . Visual inspection of the KlACS motif found by GIMSAN applied to the initial set of screened KlARSs strongly suggested that the correct motif length was 50 bp ( Figure 1C , top panel ) . Importantly , this agreed with the predicted length from previous work on two genomic KlARSs [29] , [30] . We therefore consistently instructed GIMSAN to search for a motif of length 50 bp in all the iterations . The newly GIMSAN-defined ACS PWM was then used by the SADMAMA application [34] to scan the K . lactis genome for new high scoring matches . Due to the discrete nature of SADMAMA's score , in some cases , the new list of top candidates could either have slightly less or slightly more than 20 candidates . We consistently chose the latter . Each batch of the approximately 20 best matches were then used to identify candidate ARSs by cloning a genomic fragment encompassing the ACS and 200 bp of flanking DNA from both sides of the predicted sequence into the non-replicating vector and assaying for ARS function ( Figure 1B ) . Our initial screen indicated that , as in S . cerevisiae , the vast majority ( 66 out of 69 ) of K . lactis ARSs resided in intergenic regions . We therefore largely restricted our predictions to the cases where the predicted ACS falls in K . lactis intergenic regions . Starting from the set of 69 initially screened ARSs in 5 predict-and-verify cycles we define a total of 123 candidate intergenic K . lactis ARSs , ( Dataset S1 ) . Of these , five candidates resisted multiple attempts at PCR and could not be verified , 75 were verified to have K . lactis ARS functionality , whereas 43 turned out to be non-functional . In addition , using a combination of other techniques to computationally predict ARSs ( see Materials and Methods ) , we verified another 4 KlARSs at a considerably higher cost of 32 negatively verified sequences ( Dataset S2 ) . Thus , we verify a total of 148 K . lactis ARSs ( 69 from our original screen and 79 verified predictions ) , 145 of which lie in annotated intergenic regions ( Figure 2 , Table S1 ) . How complete is our coverage of K . lactis ARSs , or , how many ARSs does K . lactis have ? To address these two virtually equivalent questions more effectively , we set the following conditions . Due to the paucity of ARSs in coding regions , which account for 70% of the genome , we limited our estimation to the number of ARSs in intergenic regions of K . lactis as in doing so we minimized the variability in our estimate . In addition , due to the repetitive structure of the ribosomal DNA , telomeric and sub-telomeric regions , we only considered unique representative ARSs in these regions . We estimated the total number of non-repetitive intergenic ARSs in the K . lactis genome using a technique known as “leave-one-out” ( LOO ) to account for the missing or unseen ARSs . One of the original 69 screened ARSs was systematically left out from the existing sample as the new training set to test how well the specific PWM predicts the ACS locations in the held out data [35] . Using this general approach we devised a new method to find a confidence interval for the number of intergenic KlARSs and conclude that [145 , 156] is the 95% confidence interval for the number of non-repetitive intergenic KlARSs [36] . Briefly , the process by which we estimate the total number of ARSs is independent of the predict-and-verify process used to actually find the ARSs . This estimation is based only on PWMs derived from the original set of screened ARSs and hence it is not biased by any computational predictions . To estimate the size of a finite set of binding sites S , we need a well-defined binding site ( such as the K . lactis ACS ) , an initial random sample of a subset of binding sites from S ( which in our case is the initial set of screened ARSs ) and a verification procedure that validates a site if and only if it belongs to our set S ( a K . lactis ARS in our example ) . To verify that our PWM model does not oversimplify the real intricacy of the presumed ORC binding site we tested our estimation procedure on S . cerevisiae [36] . We showed that starting from a randomly sampled subset of the set S of confirmed S . cerevisiae ARSs our 95% confidence interval for the total number of confirmed S . cerevisiae ARSs underestimates the true number less than 5% of the time . This result is exactly what one would expect from a 95% confidence interval and it demonstrates that our model ( as described in details in [36] ) correctly captures the nature of the binding sites . The estimate of [145 , 156] interval for the total set of K . lactis ARSs means that we are 95% confident that our coverage of K . lactis ARSs is essentially exhaustive . The credibility of this claim is further strengthened by noting that the 16 lowest scored candidates in our 5th and last iterative list include only one verified ARS compared with 13 candidates that were verified to be non-functional while another 2 could not be cloned ( Dataset S1 ) . From the 148 ARSs we verified , a refined 50 bp ACS was derived ( Figure 1C , bottom panel and Dataset S3 ) . Visually , there is no significant difference between the initial ( top ) and refined ( bottom ) PWM logos in Figure 1C . However , there is a significant difference in terms of the predictive power of the underlying numerical PWMs . Specifically , the top 100 intergenic ACS matches to these two matrices have a negative ( non-functional ARS ) to positive ( functional ARS ) ratio of 0 . 49 for the initial PWM but only 0 . 28 for the refined PWM . The most notable difference between the KlACS and the ScACS is the length ( compare Figure 1C and 1D ) . While the “canonical” ScACS consists of a 17 bp asymmetric A-T sequence ( Figure 1D , nucleotides 15–31 ) [23] , the KlACS is 50bp in length with short patches of A or T in positions suggestive of an inverted dimeric motif with similar spacing of T patches on each half ( Figure 1C , bottom panel ) . As expected , KlARSs have an overwhelming preference for intergenic regions . Only three of the 69 KlARSs isolated in the initial screen are located in coding sequences and one of these three also happens to be in a repetitive telomeric region with conflicting annotations . Moreover , all of the 11 high scoring candidates we verified outside of intergenic regions ( not part of our predict-and-verify procedure ) turned out to be non-functional . Nieduszynski et al . [37] observed that ScARSs show a clear bias for intergenes between convergently transcribed over intergenes between divergently transcribed genes . Indeed , among the 296 verified ScARSs for which we could find flanking pairs of genes ( f = forward , r = reverse ) we found 108 fr ( both genes transcribing toward ARS ) , 50 rf ( both genes transcribing away from ARS ) , 140 ff + rr ( one gene transcribing towards ARS ) , or more than a 2∶1 ratio between converging and diverging flanking gene pairs ( Figure 3 ) . This distinct bias completely disappears in KlARSs where the strand statistics of flanking genes are: 37 ( fr ) , 37 ( rf ) , 71 ( ff + rr ) . Assuming that the total number of KlARSs is equal to the projected upper bound of 156 , the average spacing of ARSs in K . lactis is 69 kb compared to the 37 kb estimated for S . cerevisiae based on the total number of 323 verified ScARSs [37] . Although a similar estimate for the number of ARSs has not been carried out for S . cerevisiae , in testing our predict-and-verify approach we were able to isolate 3 new ScARSs in this study ( see Text S1 ) . The composite genome map and coordinates of the 148 KlARSs and predicted ACSs are shown in Figure 2 and Table S1 . The definition of a defined replication origin , as opposed to stochastic initiation sites , is first , that it is an ARS and second , that it is “mutable” or contains a specific sequence motif that when mutated would destroy ARS activity . To investigate whether the KlACS is necessary for function , and to define it functionally , we mutated tri-nucleotides in KlARS515 across the span of the KlACS and flanking DNA . We constructed and tested a total of 21 mutants ( Figure 4A ) . From this set , 9 mutations either destroyed or weakened KlARS function ( “weakened” ARS mutants produced very small colonies , which did not show growth if re-streaked or inoculated into liquid medium ) , while 12 mutants did not have a noticeable effect on KlARS activity ( Figure 4B ) . Nucleotides essential for KlARS function clustered in the central region of the KlACS , but mutating nucleotides in the distal regions of the KlACS also had a strong deleterious effect on KlARS function . The mutagenesis analysis showed that functionally essential nucleotides were interspersed with non-essential ones , in a pattern corresponding with conserved and non-conserved positions in the PWM that form the putative inverted dimeric motif . To test whether mutating the KlACS would destroy function in other KlARSs , we made trinucleotide substitutions in 5 additional KlARSs that should destroy the KlACS motif . In all 5 cases ( Figure 4C , only 3 shown ) , these mutations destroyed the ARS function of the target sequence confirming that the KlACS motif is necessary for KlARS function . To investigate to what extent this motif is sufficient for KlARS function , we systematically shortened DNA flanking the KlACS and then tested for ARS function ( Figure 4D ) . We truncated 10 ARSs leaving 100 bp flanking either side of the KlACS and all 10 retained function ( Figure 4D , top row ) . Further truncating these and 10 others to leave only 50 bp flanks did not destroy ARS function except for two ( Figure 4D , middle row , Table S1 ) . From the remaining 18 ARSs , we further truncated 10 to short 10 bp flanks . Eight of these shortest constructs retained ARS function ( Figure 4D , bottom row , Table S1 ) . Interestingly , the two non-functional shortest ARS constructs had the lowest and 4th lowest scored KlACS match among these 10 ARSs . These findings suggest that the KlACS is largely sufficient for function and that KlARSs rely almost exclusively on the ACS motif and not on flanking sequences . This motif dependence is profoundly different from S . cerevisiae , where the ACS is not sufficient for function and from S . pombe and metazoans , where no consensus motif can be identified [3] . While the ARS assay has been commonly used as a surrogate for replication origin function , it is important to verify that plasmid-borne ARSs are origins of replication in their native genomic loci . Ten randomly chosen KlARS regions and 3 non-ARS control regions were analyzed for initiation activity by 2-D gel electrophoresis analysis [38] ( Figure 5 ) . All 10 ARS-containing regions showed an initiation “bubble” arc indicating that initiation of DNA synthesis occurred within a specific site within the regions analyzed . In contrast , none of the non-ARS controls showed origin firing activity within the probed regions ( Figure 5 ) . These results suggest that the KlARSs are also active replication origins in K . lactis . Eukaryote genomes contain a much larger repertoire of replication origins than the number that is actually used in every cell division [3] . This apparent redundancy suggests that any one particular ARS should not be under strong selection . Thus if indeed a great redundancy is built into the replication initiation process we should see no significant evolutionary conservation in the locations of individual ARSs . To test this hypothesis , we compared the K . lactis ARSs characterized in this study and the cumulative S . cerevisiae ARSs analyzed to date . Of the 38 S . cerevisiae intergenes that are syntenic ( with respect to the adjacent gene pairs ) to ARS-containing K . lactis intergenes , only 2 ( 5% ) contain a verified ScARS . Incidentally , one of the only two syntenically conserved ARSs is KARS101 ( KlARS406 ) which was previously identified by Fabiani et al . [28] and included in our initial set of 69 K . lactis screened ARSs ( Table S1 ) . When contrasted with the fact that altogether 51 ( 2 . 6% ) of the verified ScARSs fall in the 1992 intergenes syntenic between S . cerevisiae and K . lactis , there is no statistically significant syntenic conservation of ARSs between these two species ( insignificant two-sided hypergeometric p-value of 0 . 51 ) ( Dataset S4 and Text S1 ) . We therefore examined the conservation of KlARSs at genetic loci where ARSs are known to perform specific functions in S . cerevisiae . The binding of ORC to ACS at HMLa and HMRa has been shown to be responsible for the silencing of these loci in S . cerevisiae . These silencing sites can also function as ARSs . We examined whether any of our K . lactis ARSs resides proximal to the orthologous silent mating type loci of K . lactis . The closest ARSs are 6 kb and 20 kb away suggesting such dual-function sites are unlikely to exist in K . lactis . An interpretation of this observation is that unlike S . cerevisiae [39] , [40] , ORC binding may not be involved in the repression of the silencing of these cryptic cassettes . K . lactis may deploy some other means to repress the silent mating type loci as exemplified by S . pombe [41] , [42] . Ribosomal DNA repeat units are organized in large tandem blocks in eukaryotes such that complete replication of these large chromosomal segments must rely on internal initiation sites . The initiation sites of DNA replication are localized at the non-transcribed region ( NTS ) within rDNA repeats in every eukaryote analyzed to date including S . cerevisiae [43] , S . pombe [44] , mouse [45] and human [46] . We found no exception in K . lactis . KlARS418 is located at the NTS of rDNA repeats on chromosome D in synteny with ScARS1200 ( SGD notation ) shielded from the traffic of actively transcribing rRNA genes . Recently published analyses showed that S . cerevisiae ARSs co-localize with putative evolutionary breakpoints [47] , [48] . Does this observation apply to K . lactis ARSs ? We found that ARSs are located preferentially in intergenic regions that are syntenically non-conserved between S . cerevisiae and K . lactis . We found 34 . 7% ( 1992/5736 ) of S . cerevisiae intergenic regions are flanked by genes that are syntenic with K . lactis , but only 17 . 3% ( 51/294 ) of intergenic ScARSs lie in syntenic regions . Using a hypergeometric distribution analysis we conclude that S . cerevisiae origins prefer nonsyntenic intergenes relative to K . lactis ( p = 2 . 0e−11 ) and that there is no evolutionary conservation in the location of ARSs . The analysis above is based on a null model where ARSs are randomly assigned to intergenic regions ( see Text S1 ) . There is however a possible flaw with this null model . Namely , it does not take into account the fact that the average length of the intergenic regions that are syntenically conserved is 398 bp while the average length of the intergenic regions that are not syntenically conserved is 655 bp . Adjusting this analysis accordingly gives a more conservative but still significant p-value of 0 . 003 ( Dataset S4 ) . Similarly , only 26% ( 37/145 ) of KlARSs localize to syntenic regions compared with 38% ( 1988/5165 ) of all K . lactis intergenic regions that are syntenically conserved with S . cerevisiae . These statistics suggest a preference for non-syntenic intergenic spaces ( p = 0 . 001; adjusted for intergene length , p = 0 . 04 ) ( Dataset S4 ) . One possible explanation for the surprisingly small number of ARSs between syntenically conserved genes is that ARSs , particularly in S . cerevisiae , are associated with hot-spots for chromosomal instabilities [47] . tRNAs are also believed to be associated with chromosomal instability [49] , [50] . In light of the co-localization of ARSs with non-syntenic intergenic regions it is therefore obvious to ask whether they are also associated with tRNAs . This association was previously addressed by Wyrick et al . [19] and by Gordon et al . [47] . Both studies affirmed a positive association using essentially the same null model . However , previous work did not take into account the fact that S . cerevisiae intergenic regions that include tRNAs are considerably longer than intergenic regions that do not include any tRNA: average lengths of 1709 bp vs . 495 bp . Indeed , our own analysis suggests that using the “random intergenic” model ( similar to our first model above as well as to that used by Gordon et al . ) one finds a very significant association between S . cerevisiae tRNAs and ARSs ( 2-sided p-value of 3e-10 see Text S1 ) . However , after adjusting for the different lengths of regions that do and do not contain a tRNA , we find that this association is statistically insignificant ( 2-sided p-value of 0 . 5 ) . While both models have validity , our analysis points out the stark difference between the results of these two plausible models . However , unlike in S . cerevisiae , we do find statistically significant co-localization between K . lactis ARSs and tRNAs ( 2-sided p-value of 2e-6 ) even after adjusting for intergene length differences ( The average length of a K . lactis intergenic region with/without a tRNA is 1114 bp vs 600 bp , respectively , see Text S1 ) . This suggests that the association between ARSs and tRNAs is statistically stronger in K . lactis than in S . cerevisiae , however the functional relationship between these two elements is not yet understood . How much does the effectiveness of our iterative approach in discovering K . lactis ARSs owe to the unusually long 50 bp ACS motif ? How well can it predict S . cerevisiae ARSs where the ACS is significantly shorter ? In particular , how does it compare with Oriscan [22] , another computational approach to discovering S . cerevisiae ARSs ? Oriscan relies on a fairly complicated computation model that surveys a much larger region around the ACS . Our method , on the other hand , simply models the ACS , and relies on experimental feedback to help it successively improve the ACS PWM . To ensure a fair comparison between the two approaches we adopted the same training set of 26 S . cerevisiae ARSs that was used in the Oriscan study . We also used the same initial 17 bp PWM that the authors compiled from all experimentally verified ACS matches in this training set . We next applied our iterative predict-and-verify approach to locating the S . cerevisiae ARSs , using OriDB [37] as a surrogate for the experimental verification process . OriDB has three categories of ARSs: “confirmed” , “likely” and “dubious” . Based on various criteria and independent lines of verification ( see Text S1 ) , we classified the 284 “confirmed” and 216 “likely” ARSs as functional ARSs and the “dubious” , as well as any other locations for which we have no data , as nonfunctional ARSs . At the end of 5 iterations of our method , our top 100 distinct predictions ( Dataset S5 ) include 91 verified ARSs , 8 nonfunctional sequences , and 1 “dubious” ARS . According to our definition , this gives a 9% failure rate . In comparison , the top 100 distinct predictions of Oriscan ( Dataset S6 ) , which includes 84 functional ARSs , 14 nonfunctional sequences , and 2 “dubious” ones , give a 16% failure rate . Thus our iterative method , which relies only on predicting the ACS , represents more than 44% improvement over Oriscan's top 100 predictions ( see Text S1 for details ) . We also evaluated the efficacy of our iterative approach by comparing the first 100 predictions from our iterative approach with the top 100 sites predicted from a PWM deduced only from the initial set of 69 KlARSs ( no iterations ) . The latter , non-iterative , predictions include 63 functional ARSs , 31 candidates that were verified to be non-functional and 6 candidates that could not be or were not verified . Of note is that 4 of the 6 unverified candidates were ranked at the bottom of the list where the ratio of functional to non-functional predictions was 4 to 8 . The top 100 predictions from the iterative approach yields 70 functional ARSs , 28 candidates that were verified to be non-functional and 2 candidates that could not be verified . Comparing the ratio of negative to positive prediction we have 0 . 4 for the iterative method and a higher 0 . 49 for the non-iterative method . In summary , the iterative approach is superior to other approaches for identifying ARSs of a naive genome both in efficacy , accuracy and scope . KARS12 ( alias KlARS503 here ) is one of the two genomic K lactis ARSs that were known prior to this work and its 50 bp ACS was previously predicted based on the essential element determined by linker-scanning analysis of KARS101 ( alias KlARS406 ) [28] , [29] . A PWM offers a more refined motif representation than a consensus sequence and it is therefore not surprising that it can better predict putative ACS sites . Specifically , in the case of KARS12 the previously reported ACS ( Figure 6A , highlighted green ) scored significantly lower than another overlapping site ( Figure 6A , highlighted blue ) ( 4 . 23 vs . 6 . 8 SADMAMA scores ) . Mutational analysis directed to 8 bp that affected only the originally annotated KARS12 ACS had no effect on the activity of KARS12 ( Figure 6A , black lower case letters ) while mutations that replaced 8 bp from the higher scoring predicted ACS destroyed its activity ( Figure 6A , red lower case letters ) . A closer look at KARS12 reveals an interesting feature that is unique to this ARS . If we delete the ( GG ) nucleotides in positions 22–23 we get a significantly better match to the ACS ( score 17 . 6 ) . To see whether this prediction is borne out in an experiment we created another mutant of KARS12 where we deleted this ( GG ) dinucleotide ( Figure 6A ) . This mutant was indeed functional , supporting the conjectured dimeric nature of the ACS and suggests further that the KlORC might have some flexibility in the spacing between the two halves . To further test this hypothesis , we scanned the K . lactis genome for other sites whose ACS score is high if we allow a deletion of 1–3 bases at this same position . None of the 11 high scoring such candidates we tested showed ARS activity ( see Text S1 ) . To further test whether there is room for flexibility between the two putative parts of the KlACS we used MAFFT engine in the JALView multiple alignment GUI [51] , [52] to realign the extended ACS motif ( best match to ACS plus 5 flanking bp on each side ) . The idea is that if there is a flexible gap between the two parts of the motif the alignment could be improved by inserting the correct gap length . The results were almost identical to aligning the original gapless ACS . In other words , there was no support for a flexible gap in the dimeric structure using the above criteria for flexibility . The 50-bp ACS of KARS101 was previously verified by linker-scanning mutagenesis [28] and it coincides with the ACS ( KlARS406 ) that we predicted computationally , confirming the accuracy of our predictions . However , because of the size of the linkers ( 8-mer ) and their positions , we are unable to decipher the relative importance of the conserved nucleotides within the ACS motif ( Figure 6B ) .
We have developed a novel iterative approach that uses computational predictions to guide experimental verification in mapping the genome-wide locations of replication origins in a naive genome , K . lactis . We found that K . lactis , similarly to S . cerevisiae , uses defined sequences for the initiation of DNA replication . A unique feature of K . lactis ARSs is an unusually long ACS of 50bp , which is both necessary and largely sufficient for ARS activity . Compared to the 17 bp ACS of ScARSs that requires additional auxiliary B elements for function [53] , [54] , [55] , [56] , [57] , the KlACS may have incorporated the necessary features of the auxiliary B elements of S . cerevisiae in a spatially more restricted fashion . The entire repertoire of non-repetitive intergenic K . lactis ARSs is projected to be no more than 156 , which is significantly smaller than the 326 verified S . cerevisiae ARSs . Considering that these two yeast species have similar genome sizes ( 12Mb for S . cerevisiae and 10 . 6Mb for K . lactis ) , the reduced number of origins corresponds with an almost two-fold difference in their average replicon size . Given that only a fraction of all origins is used in any cell division cycle in S . cerevisiae [58] , [59] and S . pombe [60] , a decreased number of potential replication origins in K . lactis should coincide with either an increase in origin efficiency [61] , a significant increase in replication fork rate , or a longer S phase . A recent study using DNA combing showed that the actual number of ARSs in a given genome has little bearing on the average replicon size of the replicating genome [62] . Rather , ARSs are potential initiation sites that are activated in a stochastic fashion in individual cells . The number of sites being activated appears to be responsive to the replication stress presented by different growth conditions . A comparison of the replicon sizes of S . cerevisiae and K . lactis under the same growth condition may provide information about the relative efficiency of origin usage in these two species . The conjectured dimeric configuration of the KlACS may also shed some light on the efficiency of usage of the ARSs in these two species . If two , instead of one copy of ORC binds to each KlACS , it may improve the probability of activation of ARSs in K . lactis and a smaller repertoire of ARSs in K . lactis would perform just as well as a repertoire twice its size in S . cerevisiae . In support , a recent study in S . pombe showed that limiting initiation factors regulate the efficiency of replication origin usage [60] . Replication origins that are more accessible or have higher affinity for the limiting initiation factors are preferentially initiated . Conforming to replication origins of other eukaryotes , K . lactis ARSs have an extreme bias for intergenic regions but without any bias with respect to the orientation of transcription of the flanking genes . This independence from nearby transcription is consistent with a more robust initiation mechanism [24] . Another possibility for the different densities of replication origins in the two species is that this phenomenon is a remnant of the WGD event . S . cerevisiae may have retained the duplicated origins which are also sites of active evolutionary changes , and therefore the lack of synteny of ARSs between these two strains . Future comparative analysis of additional non-WGD strains such as S . kluyveri [25] and WGD strains could address this possibility . Based on the newly derived genome map of K . lactis ARSs and the cumulative genome map of S . cerevisiae ARSs , we found that there is no evolutionary conservation in the location of nonrepetitive ARSs . Rather , there is a statistically significant preference for ARSs to be located at nonsyntenic intergenic spaces of these two species as previously noted for S . cerevisiae [47] . It is unclear whether there are evolutionary forces that exert a positive selection for ARSs to be located at rearranged chromosomal breakpoints , or replication origins are the hotspots for generating chromosomal rearrangements . It is equally intriguing whether some of the syntenically conserved ARSs are positively selected for to serve a specific function as suggested by the invariant position of rDNA initiation sites in eukaryotes . These questions could be addressed by expanding the comparative analysis to include additional genomic replication origin maps of yeast species of various evolutionary distances .
Genomic DNA from the uraA auxotrophic MW98-8C strain of K . lactis [31] was isolated using standard methods . Cells were broken using glass bead lysis and the genomic DNA was separated from mitochondrial DNA using a standard CsCl gradient protocol . Following the gradient , DNA was precipitated with 75% ethanol and resuspended in water . Prior to library construction , the DNA was digested with MboI and treated with Antarctic Phosphatase ( New England Biolabs ) to prevent the multimerization of genomic fragments . The digested insert DNA was ligated into the unique BamHI site in the pIL07 vector ( see below ) using T4 DNA Ligase ( New England Biolabs ) . The ligation reaction was purified using a PCR Purification Kit column ( QIAgen ) and redigested with BamHI to linearize all empty pIL07 molecules . The resulting reaction was used to transform chemically competent E . coli cells using standard methodology . Colonies from the transformed ligation reaction were pooled and plasmid DNA was extracted using the Wizard Plus SV Miniprep Kit ( Promega ) . The pIL07 plasmid was made from the pUC19 subcloning vector for the purpose of isolating ARSs . S . cerevisiae LEU2 gene was PCR amplified with primers containing XbaI sites , digested and ligated into the XbaI site in pUC19 . URA3 and S . cerevisiae CEN5 were cloned similarly into the EcoRI and HindIII sites respectively . The BamHI site used for screening ARS fragments is located between the divergently transcribed URA3 and LEU2 genes . Full sequence of this vector is available upon request . The libraries were used to transform the MW98-8C strain using a standard Lithium Acetate protocol [31] and plated on medium lacking uracil to select for ARS function . Transformants were individually restreaked onto fresh plates and subsequently grown in culture medium lacking uracil to enrich for the ARS bearing plasmid . The plasmids were isolated from yeast using a modified DNA extraction protocol . 2 mL of culture was pelleted and resuspended in 500 µL of a buffer consisting of 1 M sorbitol , 0 . 1M EDTA , 1 µL/mL β-mercaptoethanol , and 0 . 5 mg/mL Yeast Lytic Enzyme ( VWR , catalog # IC360951 ) . The suspension was incubated at 37 degrees for 1 hour . The treated cells were pelleted and processed using the Wizard Plus SV Miniprep kit ( Promega ) . 5 µL of the resulting eluate was used to transform E . coli and transformants were miniprepped to isolate the ARS-bearing plasmid . A small sample of each plasmid was used to re-transformed MW98-8C to confirm ARS function . Confirmed ARS plasmids were sequenced using primers IL325 ( 5′-GCCAAACAACCAATTACTTGTTGAGA-3′ ) and IL326 ( 5′-TTCGTTGCTTGTCTTCCCTAGTTTC-3′ ) from both ends of the ARS fragment . To clone and test predicted regions of DNA for ARS activity , primers containing BamHI and BglII cloning sites were designed to anneal to the relevant region . PCR amplified DNA was cloned into pIL07 and confirmed by sequencing . Several clones of each predicted region were used to transform MW98-8C to test for ARS function . ARS fragments were truncated by amplifying and cloning smaller fragments of the ARS region . Site-directed mutagenesis was performed using a fusion-PCR mutagenesis method [63] . The DNA region to be mutagenized was PCR amplified in two separate fragments , which overlap by 50–60 base pairs . The overlap primers contained the desired mutation . After the initial PCR , the two fragments were purified separately and used together in another PCR reaction without any template DNA . The overlapping regions in the two DNA fragments acted as primers for each other and PCR produced a final molecule which contained the entire DNA fragment including the mutation of interest . This fragment was then cloned into the vector and sequenced as above . Two-dimensional DNA gel electrophoresis was performed according to the neutral–neutral method [64] . Purified DNA was digested to completion with NcoI . To enrich the sample for replicating DNA , digested DNA was passed through BND cellulose ( Sigma-Aldrich ) columns as described in [65] . ARS probes were made by amplifying 1 . 5 kb region centered at the ARS by PCR and radiolabeled with [α-32P] dATP using the Prime-It II Random primer labeling kit from Stratagene . The gibbsMarkov part of GIMSAN was applied to the set of previously verified ARSs using a training file consisting of all K . lactis intergenic regions with the parameters: ( - gibbsamp -l 50 -p 0 . 05 -best_ent -em 0 -t 200 -L 200 -ds -markov 4 ) . In addition to the PWM reported by GIMSAN , SADMAMA was given the same training file as GIMSAN , which was also used as the input target set for scanning for matches to the PWM . Matches at repetitive telomeric regions were ignored . Other parameters were: ( -printScoresGTT first_set -siteThresholdLearnedFrom 1e-4 both_strands nullTrainFile -tests – -pwmPC 0 . 01 -m 4 both_strands -siteNullScore avg_strands ) . The K . lactis genome downloaded from NCBI is ( ACCESSION NC_006037 VERSION NC_006037 . 1 GI:50313006 ) . Intergenic sequences were generated from the GenBank files of the genome by filtering out all sequences with feature type ‘gene’ , ‘LTR’ , or ‘gap’ . The fifth iterative list of K . lactis candidates was compiled from the set of ARSs that were verified in the previous 4 steps with one exception . A telomeric ARS ( C - 44929 ) was removed , as its similarity to the originally screened telomeric ARS at the end of chromosome C would have biased the estimated PWM . As it was designed to be the last iteration this 5th list contained 41 candidates . To complement our main iterative approach to predicting K . lactis ARSs we tried several other prediction techniques . Excluding the overlap with the iterative procedure the net result of these predictions was 4 verified ARSs and 32 candidates that were verified to be non-functional . Briefly , the following techniques were used: Our conservation analysis hinges on the analysis of syntenic pairs of S . cerevisiae and K . lactis protein coding genes . The latter are defined as follows: a pair of lactis-cerevisiae genes is declared “homologous” if there exists a pairwise alignment of the two with BLAST E-value <1e-4 . Two pairs of adjacent K . lactis and S . cerevisiae genes are declared syntenically conserved if the two pairs of genes are pairwise homologous and if their orientations are syntenically conserved . Note that a pair of K . lactis genes can thus be syntenically conserved with more than one pair of S . cerevisiae genes and vice versa . Fortunately , these cases do not occur often: the 2018 syntenic pairs are made of 1992 distinct S . cerevisiae pairs of genes and 1988 distinct K . lactis pairs . By extension , we consider S . cerevisiae and K . lactis intergenic regions are syntenically conserved if their corresponding flanking genes are . Our list of intergenic KlARSs consists of 145 sequences identified as described above ( Table S1 ) . There is however some freedom in defining the set of cerevisiae intergenic ARSs . In particular , OriDB lists “Confirmed” , “Likely” , and “Dubious” ARSs that are subject to judicious classification . To minimize the potential effects of this “judicious classification” on our analysis we performed the analysis twice: first with essentially only the confirmed OriDB ARSs and then again when we added high scoring likely OriDB ARSs . Reassuringly , the results were essentially the same ( see more details in Text S1 ) . To determine the strand statistics of the pairs of genes flanking the ARSs in both species we used the following criteria . For K . lactis we positioned the 145 intergenic ARSs based on the center position of their best ACS match . The strand of the flanking pair of genes was determined from the annotations in the K . lactis GenBank files mentioned above . For S . cerevisiae we made use of a list of 298 verified intergenic ARSs for which we identified an intergenic best ACS match ( all ARSs with an ARS type different than ‘Likely’ in Dataset S7 , see section on ARS positional conservation analysis in Text S1 ) . List of adjacent genes and their strands was compiled from SGD's latest version of the S . cerevisiae protein coding sequences as of Jun 6th 2008 . These genes were then mapped back to the October 2003 genome ( NC_001133 . 3 GI:37362608 - NC_001148 . 2 GI:37362698 ) based on shared systematic name . | DNA replication is an evolutionarily conserved , cell cycle–regulated , spatially and temporally coordinated mechanism in eukaryotes . It is initiated by the binding of the Origin Recognition Complex ( ORC ) to multiple replication origins . While the ORC is highly conserved , its DNA binding specificity and the primary sequences of replication origins are not . Comparative functional genomics is an obvious approach to addressing questions about the positional conservation and chromosomal determinants of replication origins . However , to date , Saccharomyces cerevisiae is the only eukaryote with a complete genome-wide replication origin map , one which took three decades to compile . We have devised an iterative approach , combining computational prediction and functional validation by direct cloning of replication origins that efficiently identifies a high resolution , near complete repertoire of replication origins in Kluyveromyces lactis . Comparing these two yeast genome maps provides a wealth of information about the DNA elements and positional conservation of replication origins in these two distantly related yeast species . This approach is generally applicable to the construction of high-resolution genome maps of evolutionarily conserved sequences associated with assayable biological functions . Rapid generation of comprehensive functional maps of uncharacterized genomes is critical to whole genome studies of all biological functions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"genetics",
"and",
"genomics/genomics",
"genetics",
"and",
"genomics/microbial",
"evolution",
"and",
"genomics",
"genetics",
"and",
"genomics/comparative",
"genomics",
"genetics",
"and",
"genomics/functional",
"genomics",
"genetics",
"and",
"genomics/chromosome",
"biology",
"genetics",
"and",
"genomics",
"genetics",
"and",
"genomics/bioinformatics"
]
| 2010 | A Comprehensive Genome-Wide Map of Autonomously Replicating Sequences in a Naive Genome |
Merkel cell carcinoma ( MCC ) frequently contains integrated copies of Merkel cell polyomavirus DNA that express a truncated form of Large T antigen ( LT ) and an intact Small T antigen ( ST ) . While LT binds RB and inactivates its tumor suppressor function , it is less clear how ST contributes to MCC tumorigenesis . Here we show that ST binds specifically to the MYC homolog MYCL ( L-MYC ) and recruits it to the 15-component EP400 histone acetyltransferase and chromatin remodeling complex . We performed a large-scale immunoprecipitation for ST and identified co-precipitating proteins by mass spectrometry . In addition to protein phosphatase 2A ( PP2A ) subunits , we identified MYCL and its heterodimeric partner MAX plus the EP400 complex . Immunoprecipitation for MAX and EP400 complex components confirmed their association with ST . We determined that the ST-MYCL-EP400 complex binds together to specific gene promoters and activates their expression by integrating chromatin immunoprecipitation with sequencing ( ChIP-seq ) and RNA-seq . MYCL and EP400 were required for maintenance of cell viability and cooperated with ST to promote gene expression in MCC cell lines . A genome-wide CRISPR-Cas9 screen confirmed the requirement for MYCL and EP400 in MCPyV-positive MCC cell lines . We demonstrate that ST can activate gene expression in a EP400 and MYCL dependent manner and this activity contributes to cellular transformation and generation of induced pluripotent stem cells .
Merkel cell carcinoma ( MCC ) is an aggressive skin cancer with a high rate of mortality . Risk factors for developing MCC include immunosuppression and UV-induced DNA damage from excessive exposure to sunlight [1] . Recognition of the immunosuppressive risk for MCC prompted a search to identify pathogens and led to the discovery of Merkel cell polyomavirus ( MCPyV ) [2] . MCPyV-positive MCC tumors contain clonally integrated copies of viral DNA and express small T antigen ( ST ) and a truncated form of large T antigen ( LT ) . Genome sequencing of virus-negative MCC revealed an extremely high number of single nucleotide polymorphisms containing the C>T transition consistent with UV damage [3 , 4] . In contrast , MCPyV positive tumors contain very few somatic mutations suggesting that MCPyV ST and LT contribute the major oncogenic activity to MCC development . In all virus-positive MCC cases reported to date , LT has undergone truncations that disrupt viral replication activities but leave the LXCXE , RB-binding , motif intact [5] . While LT can bind and inactivate RB , it is less clear how ST contributes to MCC tumorigenesis [6 , 7] . MCPyV ST has oncogenic activity and can contribute to tumor development in mouse models [8 , 9] . Similar to ST from other polyomaviruses , MCPyV ST can bind to the A and C subunits of protein phosphatase 2A ( PP2A ) [10] . However , ST binding to PP2A may not be necessary for its oncogenic activity . A unique domain of ST known as the LT stabilizing domain ( LSD ) has been reported to bind FBXW7 and CDC20 , functions to increase levels of LT , and contributes to its transforming function [11 , 12] . ST can activate MTOR signaling and promote the phosphorylation of EIF4EBP1 ( 4EBP1 ) [13] . ST can also perturb NFκB and inflammatory signaling [14 , 15] . When expressed in fibroblasts , ST led to increased levels of glycolytic genes including the lactate transporter SLC16A1 ( MCT1 ) and induction of aerobic glycolysis also known as the Warburg effect [16] . However , it is not clear how ST contributes to so many distinct activities . The EP400 histone acetyltransferase complex is involved in multiple biological events including transcription , stem cell maintenance and DNA damage response . The mammalian EP400 complex contains at least 15 distinct components including the large subunits EP400 ( also known as p400 ) and TRRAP plus ACTL6A , BRD8 , DMAP1 , EPC1 ( and its homologue EPC2 ) , ING3 , KAT5 ( also known as Tip60 ) , MBTD1 , MEAF6 , MORF4L1 ( and MORFL2 ) , MRGBP , RUVBL1 ( and RUVBL2 ) , VPS72 and YEATS4 [17–20] . The EP400 complex contains several intrinsic enzymatic activities including EP400 chaperone activity for histone variants H3 . 3 and H2AZ , KAT5 mediated acetylation of histones H2A and H4 , and the DNA helicase activity of RUVBL1 and RUVBL2 . TRRAP can bind directly to MYC and has been reported to bind equally well to the homologue MYCN and poorly to MYCL ( L-MYC ) [21 , 22] . Here we demonstrate that MCPyV ST recruits MYCL to the EP400 complex to activate specific gene expression , promote cellular transformation and contribute to its oncogenic potential .
To understand how MCPyV T antigens contribute to MCC oncogenesis , we performed a large-scale immunoprecipitation with a monoclonal antibody ( Ab5 ) specific for the shared N-terminal region of LT and ST to identify associated cellular proteins from lysates of virus-positive MCC cell lines MKL-1 and WaGa ( Fig 1A ) [23] . Identification of the immunoprecipitated proteins by multi-dimensional protein identification technology ( MudPIT ) revealed MCPyV LT and ST ( Fig 1B and S1 Table ) [24] . RB1 and VPS39 were detected as expected given their previously reported association with LT [5 , 25] . Both homologues of the PP2A scaffold ( PPP2R1A , PPP2R1B ) and catalytic ( PPP2CA , PPP2CB ) subunits were also detected , likely due to association with ST [10 , 14] . Unexpectedly , Ab5 also co-precipitated MYCL and MAX as well as all known subunits of the EP400 complex listed above including ACTL6B , a homologue of ACTL6A , and the recently reported MBTD1 [20] . In contrast , MudPIT using Ab3 , specific for LT only , identified LT , RB1 and VPS39 and none of the components of the EP400 complex . To validate the interactions with endogenous proteins , MKL-1 cell lysates were immunoprecipitated with antibodies to MAX , EP400 , ACTL6A , EPC1 and VPS72 . Each of these antibodies co-precipitated ST , PPP2CA and MYCL as well as several components of the EP400 complex ( Fig 1C ) . MudPIT with antibodies for EP400 identified all 15 subunits including homologs of the EP400 complex plus MYCL , MAX , ST and PP2A ( Fig 1B and S1 Table ) . MudPIT with antibodies for MAX enriched for MYCL , ST , PP2A , all components of the EP400 complex plus several MAD and MAD-associated proteins [26 , 27] . MudPIT with an IgG control antibody detected small amounts of RUVBL1 , RUVBL2 , MEAF6 and ACTL6B but none of the other EP400 complex components . Therefore , antibodies for MAX , EP400 and MCPyV ST each specifically co-precipitated MYCL , the EP400 complex , ST and PP2A ( Fig 1B and S1A Fig ) . To determine if ST could form a single complex with MYCL and the EP400 complex , we performed gel filtration of MKL-1 cell lysates [28] . Fractions #5–7 contained protein complexes of ≥ 2 MDa with ST , MYCL , MAX , several EP400 complex components and EP400 itself ( Fig 1D ) . EP400 was only detected in the large complex fraction while other subunits of the complex including TRRAP , KAT5 , RUVBL2 , DMAP1 and ING3 were present in the large complex and in fractions with smaller sized complexes . Of note , MYCL1 isoform 1 ( i1 ) was present in the ST-containing fractions #5–7 whereas the larger MYCL i3 was detected in intermediate sized fractions and the shortest form ( i2 ) in smaller size fractions ( Fig 1D and 1E , S1B and S1C Fig ) . An immunoprecipitation for MAX with lysates from fraction #5 co-precipitated EP400 , TRRAP and ST ( Fig 1F ) . In contrast , MAX co-precipitated TRRAP and ST but not EP400 from fraction #13 and neither TRRAP or EP400 from fraction #21 . This indicates that a specific fraction of MAX binds to EP400 , a key component of the ST-MYCL complex [29] . To determine the contribution of MCPyV ST binding to MYCL and the EP400 complex in MCC , we transduced MKL-1 cells with lentiviral shRNAs targeting both LT and ST ( shPanT ) or ST only ( shST ) [13 , 30] . Expression of either shRNA but not scrambled shRNA ( shScr ) led to reduced levels of ST and MYCL ( Fig 2A , S2A Fig ) . Reduced levels of ST led to decreased ability of MAX to co-precipitate EP400 , TRRAP , DMAP1 and YEATS4 and reduced the ability of EP400 to co-precipitate MYCL and MAX . Of note , when ST levels were reduced , EP400 retained the ability to bind to other components of the EP400 complex including TRRAP , DMAP1 and YEATS4 . To determine if MCPyV ST could increase the ability of MAX to bind to the EP400 complex , we introduced ST or C-terminal HA-tagged ST into HCT116 cells and a virus-negative MCC cell line UISO . Immunoprecipitation for MAX from parental HCT116 and UISO cell lysates readily co-precipitated MYC but not any EP400 complex components ( Fig 2B ) . However , in the presence of MCPyV ST , MAX efficiently co-precipitated TRRAP , EP400 , DMAP1 and KAT5 . Of note , MYCL levels increased in HCT116 and UISO cells when ST was expressed ( Fig 2B , Input ) . Stable expression of ST in primary human foreskin fibroblasts ( HFF ) also increased levels of MYCL ( S2B Fig ) . To determine if ST interaction with the EP400 complex was specific , we generated a series of ST mutants and stably expressed them in HCT116 cells . We focused the mutagenesis on a region of MCPyV ST within the unique domain that is not well conserved with ST from other human polyomaviruses and includes the LSD motif ( residues 91–95 ) ( Fig 2C , S2C Fig ) . Immunoprecipitation for ST containing alanine substitutions of residues 83 to 88 ( 83-88A ) showed decreased binding to EP400 complex components while retaining strong binding to PP2A ( Fig 2D , S2D Fig ) . Within this region , substitution of E86 and E87 with serine ( E86S , E87S referred to as 2M ) led to reduced EP400 complex binding yet retained some PP2A binding . In contrast , alanine substitution of residues 83 to 95 ( 83-88A , 90-95A , 93-95A ) or 102 to 105 ( 102-105A ) led to increased levels of ST relative to WT ST ( Fig 2D , S2D Fig ) . Within this region , substitution of K92 and D93 with serine ( K92S , D93S , referred to as 3M ) led to increased levels of MCPyV ST ( Fig 2D , S2D Fig ) . Combining the 2M and 3M mutants to create 4M ( E86S , E87S , K92S , D93S ) resulted in a ST construct that expressed at levels higher than WT ST and retained PP2A binding , but was unable to co-precipitate the EP400 complex . To test the ability of these ST constructs to promote MAX binding to the EP400 complex , we performed an IP with MAX antibodies . WT , 102-105A , and 3M ST led to increased ability of MAX to co-precipitate the EP400 complex and PP2A , while the 2M and 4M mutants were not co-precipitated by MAX and did not enable MAX to co-precipitate the EP400 complex or PP2A ( Fig 2D ) . We observed that MCPyV ST binds specifically to MYCL and the EP400 complex . However , it was not clear if any of these factors were required for proliferation . To identify essential genes in MKL-1 cells , we performed a CRISPR-Cas9 screen of 18 , 493 genes using two pooled sgRNA libraries H1 and H2 , each containing 5 unique sgRNAs for each gene . Using the MAGeCK-VISPR analysis pipeline , Gene Set Enrichment Analysis ( GSEA ) of known human housekeeping genes revealed that these genes were significantly negatively correlated with the results of the CRISPR screen [31] ( S3A Fig ) . After accounting for copy number variations in MKL-1 cells , we identified 481 genes that were negatively selected in the CRISPR-Cas9 screen with a false discovery rate ( FDR ) < 0 . 05 [32] , of which 276 have been classified as housekeeping genes ( S3B and S3C Fig , S2 Table ) [33] . Among the 205 genes not classified as housekeeping genes , 79 genes were identified with FDR < 0 . 01 and the remaining 126 genes were identified with FDR > 0 . 01 but < 0 . 05 . MYCL , EP400 and RUVBL2 were identified as essential ( FDR < 0 . 05 , Fig 3A ) . Additional components of the EP400 complex were identified in the CRISPR-Cas9 negative selection screen with p-values < 0 . 05 but with higher FDR values and included KAT5 , TRRAP , DMAP1 , ING3 and YEATS4 . Given the requirement of MYCL for viability of MKL-1 cells in the CRISPR-Cas9 screen and the presence of MYCL in the ST-EP400 complex in both MKL-1 and WaGa cell lines ( Fig 1B and S1 Table ) , we examined the levels of the three MYC family members in MCC cell lines . Six different virus-positive MCC cell lines that expressed ST and LT also expressed MYCL while some had low levels of MYCN and none expressed full length MYC ( Fig 3B ) . For controls , we tested HCT116 cells that predominantly expressed MYC and Kelly neuroblastoma cells that expressed MYCN . The virus-negative MCC cell line UISO did not have detectable levels of MYCL until a C-terminal epitope tagged ST ( ST-CT ) was introduced ( Fig 3B ) . To determine if MYCL was required for ST to binding to the EP400 complex , we generated MKL-1 cells that contained doxycycline ( Dox ) inducible shRNA ( shMYCL ) or miRNA ( mirMYCL ) that specifically targeted MYCL . Expression of shMYCL or mirMYCL led to reduced levels of MYCL and decreased MAX co-precipitation of EP400 , TRRAP , KAT5 and ST ( Fig 3C ) . Notably , depletion of MYCL reduced the ability of ST to co-precipitate the EP400 complex and reduced EP400 binding to ST ( Fig 3C ) . Omomyc is a modified fragment of MYC that can bind to MAX and disrupt endogenous MYC-MAX heterodimers [34] . To test if MYCL-MAX heterodimers were necessary for ST interaction with the EP400 complex , we introduced a Dox-inducible , HA-tagged , Omomyc construct into MKL-1 cells . When expressed , HA-Omomyc co-precipitated MAX as expected but not MYCL , ST or subunits of the EP400 complex and led to decreased levels of both MAX and MYCL ( Fig 3D ) . However , ST retained the ability to co-precipitate components of the EP400 complex but not MYCL or MAX when Omomyc was expressed , indicating that ST can bind to the EP400 complex independent of the MYCL/MAX heterodimer . The viability of MKL-1 cells was decreased when MYCL levels were depleted by shMYCL or mirMYCL and when the MYCL-MAX heterodimer was disrupted by Omomyc ( Fig 3E ) . To determine regions of MYCL that contributed to ST and EP400 binding , a series of HA tagged C-terminal constructs of MYCL was stably expressed in HCT116 cells in the presence or absence of ST . When ST was present , MYCL robustly co-precipitated TRRAP , EP400 , YEATS4 , DMAP1 as well as ST ( S3D Fig ) . In the absence of ST , WT MYCL co-precipitated MAX but only weakly bound to EP400 complex components . In contrast , the N-terminal 165 residues of MYCL , unable to bind MAX , could co-precipitate ST and the EP400 complex in the presence or absence of ST ( S3D Fig ) . The N-terminal 165 residues of MYCL contains several highly conserved MYC homology boxes ( MB ) that function to bind MYC modifying proteins [35] . MB1 binds to FBXW7 and MB2 contributes to TRRAP binding ( S1 Fig ) [21 , 36] . We generated HCT116 cells that stably expressed HA-tagged MYCL full length constructs with small in-frame deletions of MB1 or MB2 . An HA IP for ΔMB2 MYCL co-precipitated ST and MAX but not the EP400 complex , while the ΔMB1 MYCL co-precipitated MAX but neither ST or the EP400 complex ( S3D Fig ) . These data indicate that MB1 and MB2 of MYCL contribute to ST and EP400 complex binding . To test the requirement for EP400 in virus-positive MCC , we generated MKL-1 cell lines containing three different dox-inducible shRNAs targeting EP400 ( Fig 4A ) . In the presence of dox , levels of EP400 were reduced and an immunoprecipitation for EP400 was unable to co-precipitate DMAP1 or MAX ( Fig 4A ) . Of note , knockdown with shEP400-1 led to decreased levels of ST and MYCL in addition to lower levels of EP400 ( Fig 4B ) . In contrast , shEP400-2 and shEP400-3 reduced EP400 levels but did not affect ST and MYCL levels . When EP400 levels were reduced by shEP400-2 or -3 , MAX and ST retained the ability to co-precipitate each other , as well as MYCL and TRRAP but not DMAP1 or YEATS 4 ( Fig 4B ) . The viability of MKL-1 cells decreased significantly upon depletion of EP400 with each of the 3-inducible shRNA ( Fig 4C ) . To determine if the effect of EP400 depletion was specific to virus-positive MCC cell lines , we introduced the 3 inducible shEP400 constructs into the virus-negative MCC line UISO . We confirmed that knockdown of EP400 led to reduced levels of EP400 in UISO cells ( Fig 4D ) . In contrast to the MCPyV-positive MCC cell line MKL-1 , the viability of UISO cells was unaffected by EP400 knock-down ( Fig 4E ) . Given that the Kelly neuroblastoma cell line is dependent on continued expression of MYCN , we tested its sensitivity to EP400 depletion [37] . Depletion of EP400 in Kelly cells by shEP400-1 led to reduced binding of MAX to DMAP1 and YEATS4 while retaining binding to MYCN and TRRAP ( Fig 4F ) . Of note , shEP400-1 did not affect MYCN levels in Kelly cells . As shown in Fig 4G , Kelly cells had reduced viability when EP400 levels were reduced . Expression of MYC or MYCL together with OCT4 , SOX2 and KLF4 ( OSK ) can generate induced pluripotent stem ( iPS ) cells from a variety of somatic cell types [38 , 39] . Furthermore , MYC interaction with the EP400 complex has been implicated in the generation and maintenance of embryonic stem ( ES ) and iPS cells [40 , 41] . Given that MCPyV ST can bind to MYCL and the EP400 complex , we tested its ability to contribute to iPS cell generation . Since keratinocytes have higher reprogramming efficiency compared to other cell types due to lower p53 and p21 protein levels [42] , we generated hTERT-immortalized human keratinocytes with an inducible OSK expression vector and stably introduced MYCL , ST or ST mutants . Expression of OSK in the presence of MYCL , ST or 3M led to the appearance of flat human ES cell-like colonies with defined borders that could be stained by alkaline phosphatase and ES cell surface markers TRA-1-60 and TRA-1-81 ( Fig 5A–5C ) [43] . In contrast , the ST EP400-binding defective 2M and 4M mutants were unable to generate iPS cells . These results indicate that ST binding to MYCL and the EP400 complex was able to cooperate with OSK to promote the generation of iPS cells . We examined if ST mediated transformation was dependent on interaction with MYCL and the EP400 complex . When a MCC tumor-derived MCPyV early region ( E ) that encoded truncated LT and wild type ST was expressed in IMR90 human diploid fibroblasts , we observed a senescent phenotype with elevated levels of p53 and p21 [23] . To suppress this phenotype , we stably expressed a dominant negative form of p53 ( p53DD abbreviated as P ) and hTERT ( H ) in IMR90 cells to generate PH cells [44] . The PH cells tolerated MCPyV early region with wild type ST ( PHE ) , 3M ST ( PH3 ) or 4M ( PH4 ) mutant ST , and exogenous MYCL ( PHL ) without undergoing senescence . Immunoprecipitation of ST with Ab5 from PHE cell lysates revealed a weak interaction with DMAP1 , a component of the EP400 complex ( Fig 5D ) . However , when MYCL was co-expressed with wild type ST in PHEL cells , ST and MAX readily co-precipitated the EP400 complex . The 3M ST mutant could efficiently co-precipitate the EP400 complex even without exogenous MYCL expression ( PH3 ) . In contrast , the 4M ST mutant ( PH4 ) was unable to co-precipitate the EP400 complex . We tested the ability of these fibroblasts to grow in an anchorage-independent manner when cultured in soft agar . IMR90 cells expressing p53DD and hTERT ( PH ) alone or with MYCL ( PHL ) were unable to form colonies . Cells expressing the MCPyV early region ( PHE ) formed a few colonies , while co-expression of MYCL ( PHEL ) led to an increased number of soft agar colonies ( Fig 5E and 5F ) . The highly expressed 3M ST mutant ( PH3 ) could induce anchorage-independent growth while the 4M mutant ( PH4 ) failed to form soft agar colonies . The number of colonies formed for each cell type reflected the relative binding of ST and MAX to the EP400 complex in the presence of the various ST constructs ( compare Fig 5D and 5F ) . Given the known chromatin binding activities of MYCL and the EP400 complex , we tested if MCPyV ST could bind specifically to DNA . We performed chromatin immunoprecipitation followed by sequencing ( ChIP-seq ) with the mass spectrometry-validated antibodies to EP400 and ST . Since no antibody suitable for immunoprecipitation or ChIP was available for MYCL , we performed ChIP-seq with the MAX antibody . Although it has been reported that MCPyV truncated LT does not bind to chromatin , we considered the possibility that ChIP with Ab5 might also enrich for chromatin-bound LT [6] . To account for this possibility , we generated a MCC cell line ( MKL-1 ) derivative that stably expressed MCPyV ST with a C-terminal HA epitope tag and performed ChIP with an HA antibody . Replicas of MAX and EP400 ChIP-seq identified many peaks that were also identified by anti-ST ( Ab5 ) and anti-HA ChIP-seq . Common gene targets were identified by assigning peaks to the nearest genes ( Fig 6A and S4A Fig ) . De novo DNA motif analysis identified the MYC target E-box sequence CACGTG as the most frequently observed motif with Z-scores -42 . 1726 , -20 . 0773 , -23 . 9634 , -19 . 137 for MAX , EP400 , ST-HA and Ab5 antibodies respectively ( Fig 6B ) . Peaks were highly enriched for promoters and 5’UTR sequences ( Fig 6C and S4B Fig ) . ChIP for MAX , EP400 or ST followed by re-ChIP for these three factors indicated that they could bind simultaneously to DNA ( S4C Fig ) . Given the strong enrichment for promoters , we performed ChIP-seq with antibodies to histone H3 modified by lysine 4 trimethylation ( H3K4me3 ) , a histone mark enriched at actively transcribed gene promoters [45] . H3K4me3 ChIP-seq identified 20 , 222 peaks with MAX , EP400 and ST centered on the same peaks ( Fig 6D and 6E ) . These results indicate that MAX , EP400 and MCPyV ST bind as a complex specifically to E boxes near the transcription start sites ( TSS ) of actively expressed genes . We validated the ChIP-seq experiments on several promoters . We prepared chromatin from MKL-1 cells after transduction with vectors expressing shMYCL , mirMYCL or controls and performed ChIP with Ab5 . As shown in S5A–S5C Fig . we observed ST binding to the MYCL gene as well as three additional gene promoters that were significantly reduced by MYCL depletion . We also prepared chromatin from MKL-1 cells containing the inducible shEP400-1 before and after dox addition . We observed strong MAX binding to several gene promoters that was reduced upon EP400 depletion ( S5D Fig ) . To identify genes and associated biological functions that are controlled by the ST-MYCL-EP400 complex , we performed RNA-seq of MKL-1 cells containing inducible shMYCL , shEP400-2 , shEP400-3 and shScr with RNA isolated from cells treated with dox for 5 days . The differentially expressed genes ( DEG ) list consists of 2157 genes that passed the cutoff Padj < 0 . 001 in all three comparisons ( shEP400-2 , shEP400-3 and shMYCL vector , relative to shScr control ) . To create heatmaps , counts were normalized separately for the two experiments ( shEP400 and shMYCL ) and then corrected for batch effect using ComBat ( S6 Fig ) [46] . These genes were first grouped into 62 clusters using model-based clustering [47] . The average expression profiles of each cluster were then merged into four general patterns of behavior using hierarchical clustering ( Fig 7A and S3 Table; see Methods for more details ) . The genes in each of the four merged clusters were evaluated for statistical enrichment in Gene Ontology ( GO ) biological process terms . Cluster membership and all results of the GO term analysis are presented in S3 Table . We observed that genes upregulated by shEP400 and shMYCL fell into the cluster DEG-CL2 and were enriched in neurogenesis , skin development and hair cycle . DEG-CL4 contained genes downregulated by EP400 and MYCL and were enriched in cellular component biogenesis , RNA processing and amide biosynthetic process . Two smaller clusters represent genes that behaved differently under shEP400 and shMYCL conditions . DEG-CL1 genes were decreased by shEP400 , increased by shMYCL and enriched for actin cytoskeleton and regulation of signaling . DEG-CL3 exhibited the opposite pattern of expression and was enriched in nerve development and liposaccharide biosynthesis . These results show that both MYCL and EP400 support cell growth by upregulating bulk synthesis of biomolecules including ribosomes and proteins while simultaneously repressing cell adhesion and developmental programs in neurogenesis and skin . To integrate expression profiling with the aforementioned ChIP-seq experiments ( Fig 6 ) , we performed Binding and Expression Target Analysis ( BETA ) that links the proximity of the ChIP-seq binding peaks to the TSS with expression level changes in the corresponding genes to predict activating and repressive activities of transcription factors [48] ( Fig 7B ) . We performed BETA analysis for MAX , EP400 and ST ChIP-seq studies with RNA-seq analysis for shEP400-2 , shEP400-3 and shMYCL . We observed that the genes whose levels decreased ( downregulate ) upon EP400 or MYCL depletion were significantly enriched for MAX , EP400 and ST chromatin binding ( S7A Fig , S4 Table ) . In contrast , genes whose levels increased ( upregulate ) with EP400 depletion were not significantly associated with the MAX , EP400 and ST ChIP-peaks . This indicates that the ST , MYCL/MAX and EP400 complex binding contributes to specific gene activation . We compared the target genes identified for each ChIP-seq analysis with the RNA-seq analysis for shEP400-2 , shEP400-3 and shMYCL and identified 951 shared target genes of MAX , EP400 and ST whose levels went down upon EP400 or MYCL depletion and had significant evidence for direct ChIP binding by BETA analysis ( BETA3 , Fig 7C & S7B Fig ) . When the RNA-seq data for shEP400-1 was also included in the analysis , a total of 379 target genes were identified ( BETA4 , S7A–S7C Fig ) . We then examined 951 genes identified in BETA3 that were downregulated by shEP400 and shMYCL with evidence of direct binding according to BETA analysis of the ChIP-seq data . We note that these genes exhibited a wide range of fold changes upon depletion of EP400 and MYCL , with 136 of the 951 genes showing greater than 2-fold downregulation due to shEP400 ( shEP400 inverse fold change ) , and 62 out of 951 genes showing greater than 2-fold downregulation due to shMYCL ( S8 Fig and S5 Table ) . To find global patterns of expression that reflected functional regulation , we centered and scaled their expression profiles and created model-based clusters and merged clusters , using the same procedure as in the analysis of the DEG list . The final merged clusters were then evaluated for GO term enrichment . Cluster membership of the BETA3 genes and full GO term enrichment results are listed in S3 Table . We found that these genes naturally divide into two groups: genes that were more strongly affected by shEP400 ( BETA3-CL1 and 2 ) and genes that were more strongly affected by shMYCL ( BETA3-CL3 and 4 ) ( Fig 7D ) . The shEP400 clusters are enriched for nucleobase-containing compound metabolic process and translation initiation and elongation whereas the shMYCL clusters are involved in RNA processing and peptide metabolic processes . Among the target genes identified in the shEP400-2 , -3 and shMYCL depletion analyses was the translational control factor 4EBP1 that has been reported to be upregulated by ST [13] . To test this effect , lysates were generated from MKL-1 cells before or after depletion of EP400 and MYCL were blotted for 4EBP1 . As expected , levels of MYCL were depleted by shMYCL and EP400 by shEP400-2 and -3 ( Fig 7E ) . Of note , levels of 4EBP1 and the phosphorylated serine residue 65 form ( pS65-4EBP1 ) were reduced upon EP400 or MYCL knockdown . In addition , we have recently reported that levels of the lactate transporter MCT1 ( SLC16A1 ) increase upon expression of ST [16] . Levels of MCT1 were also decreased upon depletion of EP400 or MYCL ( Fig 7E ) . We compared the effect of depleting EP400 or MYCL in the virus-positive MKL-1 cells to the effect of expressing ST in normal cells . We examined RNA-seq profiles from IMR90 human fibroblasts with inducible expression of GFP or MCPyV ST over the course of 4 days [16] . As shown in the heatmap in S9 Fig , the genes that were downregulated by shEP400 and shMYCL in MKL-1 cells tend to be upregulated by ST in IMR90 cells consistent with the model that ST activates functional interactions with EP400 and MYCL and their transcriptional targets .
Here , we demonstrate that MCPyV ST specifically recruits the MYCL and MAX heterodimer to the 15-component EP400 complex . These interactions are essential for the transforming function of MCPyV ST , the viability of virus-positive MCC cells and likely to be a major contributor to the oncogenic potential of MCPyV in MCC . Consistent with this model , a genome-wide CRISPR-Cas9 screen revealed that MYCL and several components of the EP400 complex were essential for viability of the virus-positive MCC cell line MKL-1 . The interaction of MCPyV ST with MYCL and the EP400 complex is unique to the family of polyomaviruses . To date , no other polyomavirus ST has been reported to bind the EP400 complex or a MYC homolog [15] . Several other viruses have been reported to engage the EP400 complex . Perhaps most similar to the results reported here , the adenovirus E1A oncoprotein binding to MYC and the EP400 complex contributes to its transforming potential [49 , 50] . We observed a striking relationship between MYCL and MCPyV ST . MYCL was expressed in all 6 virus-positive MCC cell lines tested . Furthermore , ST appeared to regulate MYCL levels . For example , introduction of ST into several naïve cell lines led to increased levels of MYCL . Conversely , depletion of ST from MKL-1 cells led to decreased levels of MYCL . ST together with EP400 and MAX could bind to the MYCL promoter . In addition , the virus-positive MKL-1 cell line was sensitive to Omomyc expression indicating that the MYCL-MAX heterodimer was required for viability as well as ST interaction . These results are consistent with a positive feedback loop where ST binding to the MYCL promoter contributes to transcriptional activation of MYCL leading to increased levels of MYCL that in turn binds to ST and the EP400 complex . MYCL has been the forgotten MYC species , since it is not required for normal mouse development and expression is highly restricted in adult tissues [51] . Despite its relative obscurity , MYCL can function as a bona fide oncogene [52 , 53] . For example , amplification of MYCL or the presence of the RLF-MYCL1 fusion gene are mutually exclusive of amplifications of MYC or MYCN in small cell lung cancer . Better still , amplification of the 1p34 locus containing the MYCL gene has been reported in MCC suggesting that Merkel cell carcinogenesis is dependent upon excessive MYCL function [54] . Consistent with this report , we estimate that the copy number of MYCL was 3 . 5 copies in MKL-1 cells based on analyzing ChIP-seq input DNA ( S3B Fig , S2 Table ) . In keeping with this notion , MCPyV ST shows a strong preference for recruiting MYCL to the EP400 complex . It is possible that , despite the widespread presence of MCPyV in healthy individuals , only cells capable of expressing MYCL are susceptible to the oncogenic potential of MCPyV [55 , 56] . MYC and MYCL can cooperate with the OSK reprogramming factors to induce a pluripotent state in somatic cells [39 , 57] . Comparison of the contributions of MYC to transformation and iPS cell generation show significant overlap with the interaction with the EP400 complex as a key component [41] . Here we find that MCPyV ST can substitute for MYCL in iPS cell generation and that this activity is strictly dependent upon ST interaction with the EP400 complex . Our data indicate that , at least in part , MCPyV ST functions similarly to MYC by binding to the EP400 complex , recruiting it to specific promoters to transactivate gene expression and thereby promoting the generation of iPS cells . These functions may also prove to be critical in establishing and maintaining the oncogenic state of MCC . Our data reveal that the ST-MYCL-EP400 complex functions , at least in part , to activate specific gene expression . Depletion of MYCL and EP400 led to significant changes in gene expression and cell viability . Those genes whose levels decreased upon MYCL and EP400 depletion were significantly associated with ST , MAX and EP400 binding to their promoters and include classic MYC targets involved in RNA processing , ribosome biogenesis , nitrogen compound and peptide metabolic processes . Additional target genes are involved in cell morphogenesis and signaling in the TNF , WNT , NFκB and DNA damage pathways . Importantly , a large number of metabolic genes were activated by the ST-MYCL-EP4000 complex including a number of transporters including SLC16A1 and SLC7A5 and the MYC-metabolism genes MLX and MLXIP ( Mondo ) [16 , 58] . Factors that promote transcription elongation were also highly enriched including EIF4E , EIF4EBP1 , EIF5A [13] . Interestingly , genes whose levels increased upon MYCL or EP400 depletion were involved in neurogenesis , axon guidance , wound healing and cell-cell adhesion . These results can be interpreted to indicate that ST-MYCL-EP400 complex serves to repress differentiation markers and induce a more primitive , progenitor or embryonic state , consistent with its ability to generate iPS cells . The MYC family functions to activate gene expression at least in part by interaction with a variety of chromatin factors . In addition to the EP400 complex , MYC can bind to the TRRAP-containing STAGA ( SPT3-TAF9-GCN5 acetylase ) complex that in turn interacts with Mediator [59] . MYC binds to BRD4 and the pTEFb complex to facilitate transcriptional elongation by release of paused RNA polymerase II [60 , 61] . The conserved Myc Boxes contribute to transformation with the Myc Box 3b ( MB3b ) binding to WDR5 and Myc Box 4 ( MB4 ) binding to HFCF1 ( S1 Fig ) [62 , 63] . MB3a , or simply referred to as MB3 , found only in MYC and MYCN and not MYCL , is required for tumorigenic activity of MYC in vitro and in vivo [64] and contributes to transcriptional repression by recruiting HDAC3 [65] . At oncogenic expression levels , MYC interacts with MIZ-1 ( ZBTB17 ) to repress transcription , which can be disrupted by mutating valine 394 ( V394 ) in the helix-loop-helix ( HLH ) domain [66] . We only detected the EP400 complex and did not detect any of these other MYC binding factors in any of the ST complexes . Both MB1 and MB2 of MYCL contribute to ST and MYCL binding . Of note , it appears that ST and MYCL bind directly to TRRAP as evidenced by co-precipitation of TRRAP with ST and MAX antibodies after EP400 depletion ( Fig 4B ) . In contrast to MCPyV ST , transformation by SV40 ST is strictly dependent on its interaction with PP2A . SV40 ST binding to PP2A perturbs its ability to de-phosphorylate certain substrates including MYC that in turn leads to higher levels of MYC . While it has been reported that MCPyV ST interaction with PP2A is not required for its transforming function , it is possible that PP2A contributes to activity of the MYCL-EP400 complex . These results highlight an important mechanism for MCPyV ST mediated transformation . The ST-MYCL-EP400 complex functions as a powerful engine to transactivate gene expression and promote oncogenesis . Important questions to be pursued include whether any of the specific downstream transcriptional targets of the ST-MYCL-EP400 complex contribute to MCC oncogenesis and if any of these target genes provide a therapeutic opportunity for virus-positive MCC . In addition , the ST-MYCL-EP400 complex itself may provide a therapeutic opportunity to disrupt interactions between ST , MYCL and the EP400 complex or with any of its components to DNA . In addition , these results may help to explain why the oncogenic activity of MCPyV ST is limited to MCC because of its dependency on MYCL or if this virus is capable of inducing MYCL and cancers in other tissue types .
Human Subject Research performed in this study complied with all relevant federal guidelines and institutional policies . The Dana-Farber Cancer Institute Institutional Review Board ( IRB ) approved the study . All subjects were adults and provided informed written consent . MCC cell lines MKL-1 , MKL-2 and MS-1 were gifts from Masa Shuda ( University of Pittsburgh , PA ) ; MCC cell lines WaGa and UISO from Jürgen Becker ( Medical University Graz , Austria ) ; MCC cell lines PeTa and BroLi from Roland Houben ( University of Wuerzburg , Germany ) . Kelly neuroblastoma cell line was a gift from Rani George ( Dana-Farber Cancer Institute , MA ) . 293T , HCT116 and IMR90 cells were obtained from ATCC . HFK-hTERT cells were a gift from Karl Münger ( Tufts University , MA ) . MCPyV early region was PCR amplified from DNA extracted from a Merkel cell carcinoma sample [23] . The cDNA for ST was modified to eliminate the LT splice donor by introducing silent mutations ( GAG|GTCAGT to GAa|GTCtcc ) . Additional ST mutants were generated using QuikChange Lightning Site-Directed Mutagenesis Kit ( Agilent ) . The EP400 , MYCL shRNA target sequence was designed using Block-iT RNAi Designer ( Life Technologies ) and annealed forward and reverse oligos of hairpin sequence were cloned between AgeI/EcoRI sites of the doxycycline inducible shRNA vector Tet-pLKO-puro ( a gift from Dmitri Wiederschain , Addgene #21915 ) [67] . The MYCL miRNA target sequence was designed using Block-iT RNAi Designer and cloned into pcDNA 6 . 2-GW/EmGFP-miR vector ( Life Technologies ) and the pre-miRNA expression cassette targeting MYCL was transferred to pLIX_402 Dox-inducible expression vector via consecutive BP and LR recombination reactions to generate pLIX-mirMYCL plasmid . shRNAs constitutively expressed from lentiviral PLKO vector targeting MCPyV LT/ST ( shPanT ) , ST ( shST ) or scramble ( shScr ) have been published before [13 , 30 , 68] . pMXs-Hu-L-Myc was a gift from Shinya Yamanaka ( Addgene # 26022 ) [39] . MYCL was PCR amplified with C-terminal 3xHA tag or with original stop codon and cloned into pLenti-CMV gateway vector . Omomyc was a kind gift from Sergio Nasi ( Sapienza University of Rome , Italy ) , modified by PCR amplification to include C-terminal HA tag and cloned into pLIX_402 . The OCT4-2A-SOX2-2A-KLF4 polycistronic coding sequence was PCR amplified from pKP332 Lenti-OSK1 ( Addgene #21627 ) [69] and cloned into pLIX_402 . Expression vectors include pLenti-CMV ( a gift from Eric Campeau , Addgene #17451 ) [70] , doxycycline inducible lentiviral gateway expression vector pLIX_402 ( a gift from David Root , Addgene #41394 ) . Lentiviral packaging plasmid psPAX2 and envelope plasmid pMD2 . G were gifts from Didier Trono ( Addgene #12260 , #12259 ) . Retroviral packaging plasmid pUMVC3 was a gift from Robert Weinberg ( Addgene # 8449 ) [71] and envelope plasmid pHCMV-AmphoEnv from Miguel Sena-Esteves ( Addgene # 15799 ) [72] . Retroviral plasmids pBabe-neo-p53DD and pBabe-hygro-hTERT were previously described [44] . Packaging and envelope plasmids were co-transfected with lentiviral or retroviral expression vectors into 293T cells using Lipofectamine 2000 ( Life Technologies ) . Two days after transfection , 293T cell supernatant was purified with 0 . 45 μm filter and supplemented with 4 μg/ml polybrene before transducing recipient cells . Stable cell lines were generated after selection with 1–2 μg/ml puromycin , 5–10 μg/ml blasticidin , 500 μg/mL neomycin , and 100 μg/mL hygromycin as required by each vector . CellTiter-Glo Luminescent Cell Viability Assay was performed according to the protocol from Promega . Basically , 3000 MKL-1 parental or dox-inducible cells were plated in 96 well plate . Fresh medium was supplemented every two days with or without doxycycline . The number of days that cells had been treated with doxycycline was labelled on X-axis . At the end of time course , CellTiter-Glo reagents were added to lyse cells . For each cell line , doxycycline treated samples were normalized to untreated samples . Anchorage independent growth was performed as described [23] using 6-well dishes with SeaPlaque Agarose ( Lonza ) at concentrations of 0 . 3% top and 0 . 6% bottom layers . Agarose was diluted with 2X MEM ( Gibco ) supplemented with 2X Gluta-max ( Gibco ) , 2X pen-strep ( Gibco ) , and 30% FBS . IMR90 cells ( 105 ) were seeded in triplicate in the top agarose layer . Wells were fed with top agarose twice per week . After 4 weeks , cells were stained with 0 . 005% crystal violet ( Sigma ) in PBS and colonies were counted . Statistical significance was determined by ordinary one-way ANOVA for multiple comparisons with p < 0 . 05 . HFK-hTERT cells were transduced with pLIX-OSK and selected with puromycin to establish the parental cell line ( P ) followed by transduction with MYCL or ST in pLenti-CMV vector and selection with blasticidin . 200 , 000 cells were seeded in Matrigel ( BD Biosciences ) coated 6-well plate in triplicate on day 0 in Keratinocyte-SFM medium ( Gibco ) supplemented with 0 . 5 μg/ml doxycycline . On day 3 , medium was changed to mTeSR1 ( Stemcell Technologies ) supplemented with doxycycline . iPS colonies were visible under microscope after 3 weeks and stained with StainAlive TRA-1-60 or TRA-1-81 antibodies ( Stemgent ) and Alkaline Phosphatase Detection Kit ( Millipore ) . The following antibodies were used: Ab5 and Ab3 [23 , 73]; HA ( Abcam ) ; EP400 , RUVBL2 ( Bethyl ) ; MAX , KAT5 , DMAP1 , MNT ( Santa Cruz ) ; MYCL ( R&D Systems ) ; ING3 ( Sigma ) ; PPP2CA ( BD Biosciences ) ; and H3K4me3 ( Millipore; 07–473 ) . Cell lysates were prepared in EBC Lysis buffer ( 50 mM Tris pH 8 . 0 , 150 mM NaCl , 0 . 5% NP-40 , 0 . 5 mM EDTA , 1 mM β-Mercaptoethanol and freshly added protease inhibitor and phosphatase inhibitor cocktail ) . Immunoprecipitations were performed with protein G Dynabeads ( Life Technologies ) mixed with immunoprecipitation antibodies or anti-HA magnetic beads ( Pierce Biotechnology ) . After overnight incubation on a rotating apparatus at 4°C , magnetic beads were washed with high salt wash buffer ( 50 mM Tris pH 7 . 4 , 300 mM NaCl , 0 . 5% NP-40 , 0 . 5 mM EDTA ) five times . Bound proteins were eluted from magnetic beads with 2x Laemmli sample buffer ( Bio-Rad ) . After electrophoresis , the separated proteins were transferred to PVDF membrane and blotted . Immunoblots were developed using Clarity Western ECL substrate ( Bio-Rad ) and imaged with G:BOX Chemi system ( Syngene ) . MudPIT was performed with MKL-1 or WaGa suspension cells ( 30 x 15-cm diameter plates ) harvested in 30 ml EBC lysis buffer . Clarified cell extract ( 100–300 mg ) was incubated overnight at 4°C with 30 μg antibodies crosslinked to 30 mg protein G agarose beads by dimethyl pimelimidate ( DMP ) . Beads were washed with high salt wash buffer five times , then eluted with 0 . 2 M glycine pH 3 and neutralized with 1 M Tris pH 8 . 0 . Proteins were precipitated with 1/5 TCA overnight at 4°C and washed with cold acetone twice and analyzed by MudPIT as described [74] . The triple-phase fused-silica microcapillary column was packed with 8–9 cm of 5-μm C18 Reverse Phase ( Aqua , Phenomenex ) , followed by 3 to 4 cm of 5-μm Strong Cation Exchange material ( Partisphere SCX , Whatman ) and 2 to 3 cm of C18 RP and equilibrated with Buffer A ( 5% ACN , 0 . 1% Formic Acid ) . A10-step chromatography run was performed with the last two chromatography steps consisting of a high salt wash with 100% Buffer C ( 500mM Ammonium Acetate , 5% ACN , 0 . 1% Formic Acid ) followed by an acetonitrile gradient to 100% Buffer B ( 80% ACN , 0 . 1% Formic Acid ) . 2 . 5 kV voltage was applied distally to electrospray the eluting peptides . Full MS spectra were recorded on the peptides over a 400 to 1 , 600 m/z range , followed by five tandem mass ( MS/MS ) events sequentially generated in a data-dependent manner on the first to fifth most intense ions selected from the full MS spectrum ( at 35% collision energy ) . A frozen pellet of MKL-1 cells was resuspended in mammalian cell lysis buffer ( MCLB; 50mM Tris pH 7 . 8 , 150 mM NaCl , 0 . 5% NP40 ) in the presence of protease and phosphatase inhibitors ( Roche Complete , EDTA-free Protease Inhibitor Cocktail and 25 mM sodium fluoride , 1 mM sodium orthovanadate , 5 mM β-glycerophosphate ) . The lysate was incubated on ice for 15 minutes then clarified by centrifugation in a refrigerated microfuge for 10 minutes at top speed . The supernatant was further clarified using 0 . 45 μM Durapore PVDF spin filters ( Millipore ) . Approximately 7 mg of total cellular protein was applied to a Superose 6 10/300 GL column run in an AKTA pure FPLC ( GE Healthcare ) with MCLB as the running buffer . The injection volume was 500 μl , the flow rate was 0 . 5 ml/minute , and 0 . 5 mL fractions were collected from 0 . 2 column volumes to 1 . 5 column volumes . The molecular weights were estimated by loading 1 mg of individual protein standards from the Gel Filtration Markers Kit for Protein Molecular Weights 29 , 000–700 , 000 Da ( Sigma-Aldrich ) . MCV T antigens were knocked down in MKL-1 cells using shRNAs previously published [13 , 30] . The shRNAs were cloned into pLKO . Puro vectors , lentivirus was generated in 293T cells using psPax2 and pVSV . G vectors , and MKL-1 cells were infected using spinoculation ( centrifugation at 800g for 30 mins with viral supernatants ) followed by infection overnight in the presence of 1 μg/ml Polybrene . 24 hours post infection , MKL1 cells were spun down and resuspended in medium containing puromycin ( 1 μg/ml ) . Cells were harvested after 72 hours and processed for immunoblotting and immunoprecipitation . CRISPR lentiviral libraries H1 and H2 each contain 92 , 817 pooled sgRNAs targeting 18 , 493 human genes . CRISPR screen was performed by following a previous protocol ( https://doi . org/10 . 1101/106534 ) . Briefly , 2x108 MKL-1 cells were transduced with H1 and H2 CRISPR libraries separately at MOI 0 . 3 to ensure single sgRNA incorporation per cell . After 6 days of 1 μg/ml puromycin selection , surviving cells from each sgRNA library transduction were split in half , 3x107 cells were saved as initial state controls , the rest were cultured for a month with at least 3x107 cells maintained and used as final state samples . Genomic DNA was extracted and 200 μg from each sample were used to PCR amplify integrated sgRNAs and to generate 4 libraries for next generation sequencing . 50 million reads were obtained for each sequencing library . To filter out false positive targets due to strong correlation between decreased cell viability and increased gene copy number in CRISPR/cas9 screens [32] , copy numbers of every 50-kb segment of MKL-1 genome were called from the input of ChIP-seq experiments using QDNAseq software . Segmented copy numbers were converted to copy numbers per gene based on gene coordinates . MAGeCK-VISPR pipeline was used to assess data quality , correct copy number variation effect and identify statistically significant targets [75] . MKL-1 cells or a derivative stably expressing MCPyV ST with a C-terminal 3xHA tag were used for ChIP . For MAX , EP400 , Ab5 and HA antibodies , ChIP was performed as described [76] with the modification that cells were dual cross-linked with 2 mM disuccinimidyl glutarate ( DSG ) and 1% formaldehyde [77] and sonicated at 4°C with a Branson Sonifier 250 at 20% duty cycle for 1 minute with 1 minute rest in between for 15 cycles . ChIP- reChIP was performed using the Re-ChIP-IT kit ( Active Motif ) . For ChIP-seq , 10 ng of DNA from ChIP experiments or input DNA were prepared for sequencing with NEBNext ChIP-seq Library Prep Reagent Set for Illumina ( New England BioLabs ) . Amplified libraries were cleaned up using AMPure XP beads ( Beckman Coulter ) and checked on a Bioanalyzer ( Agilent ) to confirm a narrow distribution with a peak size around 275 bp . Diluted libraries were used for 50 cycles single-end sequencing on HiSeq 2000 system ( Illumina ) at the Center for Cancer Computational Biology ( CCCB ) at Dana-Farber Cancer Institute following the manufacturer’s protocol . H3K4me3 ChIP-seq was performed as described with minor changes [78] . 0 . 5xE06 MKL-1 cells were cross-linked with 1 . 1% formaldehyde and sonicated at 4°C with a Bioruptor ( Diagenode ) . Samples were sonicated on the high setting for 30 seconds with 30 seconds rest in between . Libraries for Illumina sequencing were prepared using the ThruPlex FD DNA-seq kit ( Rubicon Genomics ) . Amplified libraries size-selected using a 2% gel cassette in the Pippin Prep system ( Sage Science ) to capture fragments between 200–700 basepairs . Libraries were run in Illumina Nextseq . ChIP-seq mapping was performed using Bowtie ( version 0 . 12 . 7 ) against human genome version hg19 allowing only uniquely mapping reads . Peak calling was done using MACS2 ( version 2 . 1 . 0 . 20140616 ) on either single replicate mapped files or replicates merged as mapped bam files using the samtools ( version 0 . 1 . 18-dev ( r982:313 ) ) merge function . Top ranking peaks ( 5000 most significant , p-val as reported by MACS2 ) and the Macs2 generated tag pileup output was visualized using the Meta Gene signal distribution function of the CEAS ( version 0 . 9 . 9 . 7 ) analysis package [79] . Significantly enriched transcription factor binding element motifs where found using the Cistrome SeqPos tool using the 1000 most significant ( Macs2 p-value ) peaks [80] . To calculate genome-wide overlap , all enriched H3K4me3 peaks were extended 5kb in each direction , divided into 250 bins and the read density was calculated in each bin . Density was normalized to the largest value observed in each experiment genome-wide and plotted either as an average of all regions ( meta plot ) or as a heat map . MKL-1 cells containing tet-PLKO-shEP400 , tet-PLKO-shMYCL and tet-PLKO-shScramble were used to perform RNA-seq . Cells ( 107 ) were collected before and 5 days after dox addition . Total RNA was purified using RNeasy Plus Mini Kit ( Qiagen ) . mRNA was isolated with NEBNext Poly ( A ) mRNA Magnetic Isolation Module ( New England BioLabs ) . Sequencing libraries were prepared with NEBNext mRNA library Prep Master Mix Set for Illumina ( New England BioLabs ) and passed Qubit , Bioanalyzer and qPCR QC analyses . 50 cycles single-end sequencing was performed on HiSeq 2000 system . Reads were mapped to the Hg19 genome by TOPHAT . HTSeq was used to create a count file containing gene names [81] . The R package DESeq2 was used to normalize counts and calculate total reads per million ( TPM ) , and determine differential gene expression . QC was performed to generate a MA plot to display differentially expressed genes . To create the heatmaps , counts were normalized separately for the two experiments ( shEP400 and shMYCL ) using “voom” from the R/Bioconductor package limma . Data were then corrected for batch effect using ComBat from the R/Bioconductor package sva . In ComBat , the normalized gene expression data were fit to a linear model capturing the effects of basal expression , sample conditions , batch variation , and noise [46] . In our case , the sample conditions corresponded to shScr , shEP400 , and shMYCL , and the batch variable corresponded to the experiment in which the data were measured . In the final step of ComBat , the best-fit parameters from the linear model were used to subtract only the effect of the batch variable from the data . Principal Components Analysis ( PCA ) plots of the data before and after ComBat show that shScr samples from the two different batches cluster together after ComBat ( S7 Fig ) . We then subsetted the batch-adjusted data using the “BETA3” list , defined as previously described in the section on BETA analysis of the ChIP-Seq data , or a “DEG” list of genes that were differentially expressed with Padj < 0 . 001 in all three comparisons: shP400-2 vs . shScramble , shEP400-3 vs . shScramble , and shMYCL vs . shScramble . Differential expression was determined using DESeq2 . Note that the DEG list includes both up- and down-regulated genes , whereas the BETA3 list includes only down-regulated genes . For each gene list , the batch-adjusted expression values were first standardized across all 15 samples by mean-centering and scaling so that standard deviations are all set to 1 . Genes were then clustered using model-based clustering as implemented in the R package mclust . An average profile was created for each gene cluster by taking the mean over the standardized expression values for all the genes in the cluster . Next , the average profiles were merged using complete linkage hierarchical clustering with a Euclidean distance metric . By cutting the tree at a height of 3 . 5 ( for the BETA3 list ) or 5 ( for the DEG list ) , we merged the model-based clusters into larger patterns of gene expression . Gene Ontology ( GO ) term enrichment was run on the final merged clusters using the R/Bioconductor package GOstats with the following parameters: the background set consisted of all the genes from the original RNA-seq alignment , the Benjamini-Hochberg method was applied for multiple testing correction , and the conditional hypergeometric test was used to take into account relationships between GO terms . Heatmaps depict the average standardized expression profiles and were created using the “heatmap . 2” function from the R package gplots . The IMR90 ST and GFP RNA-seq data is available from the Gene Expression Omnibus ( GEO ) with accession number GSE79968 . The IMR90 data were processed using Tophat and Bowtie , and the log-transformed FPKM values were used for all analysis , as described [16] . The genes in the DEG list that also had non-zero expression values across all IMR90 expression profiles were used to create the final heatmap . To visualize both datasets in the same setting , the IMR90 profiles were each subtracted by a corresponding control , which was defined as the average expression level in the IMR90 GFP cell line at the same time point . The MKL-1 shEP400 profiles were subtracted by the average expression level in the shScr samples from the shEP400 batch . Likewise , the shMYCL profiles were subtracted by the average expression level in the shScr samples from the shMYCL batch . Finally , for each gene , all its expression values across both IMR90 and MKL-1 datasets were centered and scaled to the same standard deviation to create the final heatmap . Complete linkage hierarchical clustering with Euclidean distance was used to create the row dendrogram . MAX , EP400 , ST ChIP-seq data were integrated with individual differential expression data from shEP400–1 , -2 , -3 and shMYCL RNA-seq using Binding and Expression Target Analysis ( BETA ) software package , which infers activating or repressive function of MAX , EP400 , ST and predict the target genes based on rank product of binding potential and differential expression [48] . Shared targets of all three factors were termed shEP400-1 BETA , shEP400-2 BETA , shEP400-3 BETA and shMYCL BETA respectively . Common targets of all four aforementioned datasets were termed BETA4 , or BETA3 if shEP400-1 BETA was excluded . | Merkel cell carcinoma ( MCC ) is a highly aggressive , neuroendocrine cancer of the skin . MCC frequently contains integrated copies of Merkel cell polyomavirus DNA and expresses two viral transcripts including a truncated form of Large T antigen ( LT ) and an intact Small T antigen ( ST ) . While LT binds the Retinoblastoma protein and inactivates its tumor suppressor function , it is less clear how ST contributes to MCC tumorigenesis . Here we show that ST specifically recruits the MYC homolog MYCL ( L-MYC ) to the 15-component EP400 histone acetyltransferase and chromatin remodeling complex . The ST-MYCL-EP400 complex binds to specific gene promoters to activate their expression . Both MYCL and EP400 are required for maintenance of MCC cell line viability and can cooperate with ST to promote gene expression . We demonstrate that ST enhances the interaction between MYCL and the EP400 complex interaction and this activity contributes to transcriptional activation , oncogenesis and reprogramming of MCC . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"biotechnology",
"genome",
"engineering",
"medicine",
"and",
"health",
"sciences",
"engineering",
"and",
"technology",
"synthetic",
"biology",
"crispr",
"synthetic",
"bioengineering",
"oncology",
"immunoprecipitation",
"stem",
"cells",
"induced",
"pluripotent",
"stem",
"cells",
"genome",
"analysis",
"epigenetics",
"molecular",
"biology",
"techniques",
"chromatin",
"synthetic",
"genomics",
"research",
"and",
"analysis",
"methods",
"bioengineering",
"synthetic",
"genome",
"editing",
"genomics",
"chromosome",
"biology",
"animal",
"cells",
"gene",
"expression",
"molecular",
"biology",
"gene",
"ontologies",
"precipitation",
"techniques",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"carcinogenesis",
"cell",
"biology",
"library",
"screening",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"computational",
"biology"
]
| 2017 | Merkel cell polyomavirus recruits MYCL to the EP400 complex to promote oncogenesis |
Plants are continuously exposed to a myriad of abiotic and biotic stresses . However , the molecular mechanisms by which these stress signals are perceived and transduced are poorly understood . To begin to identify primary stress signal transduction components , we have focused on genes that respond rapidly ( within 5 min ) to stress signals . Because it has been hypothesized that detection of physical stress is a mechanism common to mounting a response against a broad range of environmental stresses , we have utilized mechanical wounding as the stress stimulus and performed whole genome microarray analysis of Arabidopsis thaliana leaf tissue . This led to the identification of a number of rapid wound responsive ( RWR ) genes . Comparison of RWR genes with published abiotic and biotic stress microarray datasets demonstrates a large overlap across a wide range of environmental stresses . Interestingly , RWR genes also exhibit a striking level and pattern of circadian regulation , with induced and repressed genes displaying antiphasic rhythms . Using bioinformatic analysis , we identified a novel motif overrepresented in the promoters of RWR genes , herein designated as the Rapid Stress Response Element ( RSRE ) . We demonstrate in transgenic plants that multimerized RSREs are sufficient to confer a rapid response to both biotic and abiotic stresses in vivo , thereby establishing the functional involvement of this motif in primary transcriptional stress responses . Collectively , our data provide evidence for a novel cis-element that is distributed across the promoters of an array of diverse stress-responsive genes , poised to respond immediately and coordinately to stress signals . This structure suggests that plants may have a transcriptional network resembling the general stress signaling pathway in yeast and that the RSRE element may provide the key to this coordinate regulation .
Plants are persistently challenged with numerous biotic and abiotic environmental stresses . To cope with environmental stresses plants have evolved phytohormones such as jasmonic acid , salicylic acid , ethylene , and abscisic acid , which are utilized to regulate plant responses to both abiotic and biotic stresses with considerable signaling crosstalk [1 , 2] . While these phytohormone pathways have been well studied , knowledge of stress perception and initial signaling events , aside from plant pathogen interactions , are less defined . It is known that application of insect oral secretions containing protein fragments of chloroplastic ATP synthase or application of purified oligouronides ( OGAs ) derived from the plant cell wall are capable of inducing plant defense responses , although a receptor has not yet been identified [3–5] . Additionally , a cellulose synthase ( CESA3 ) mutant cev1 shows enhanced resistance to powdery mildew as a result of constitutive increase in jasmonic acid levels in these plants [6] . This has led to the hypothesis that mechanical disruption of the cell wall may result in stress signaling [3 , 7] . The perception of cold stress has been hypothesized to be mediated through the detection of changes in membrane fluidity and protein conformation [8–10] . Finally , secondary messengers such as Ca2+ , reactive oxygen species ( ROS ) , and phosphatidic acid have been implicated in initial signaling cascades in response to both abiotic and biotic stresses [11–17] . One mechanism of response to stress that has been studied extensively in yeast and animals is the general stress response ( GSR ) ( also referred to as the cellular stress response ) [18] . The GSR acts in a transient manner in response to a diverse array of stresses . The GSR is initiated in response to strain imposed by environmental forces on macromolecules such as membrane lipids , proteins , and/or DNA . A critical aspect of the GSR , downstream of perception of macromolecular damage , is generation of ROS [19] . Furthermore , key molecular components of the GSR are evolutionarily conserved in all organisms [18] . To better understand plant stress responses , transcript profiling experiments have been successfully employed for many different abiotic and biotic stresses [1 , 20 , 21] . One common emerging theme from these experiments is that abiotic and biotic stresses regulate different but overlapping sets of genes [1] . For example , cDNA–amplified fragment length polymorphism analysis of the Avr9- and Cf-9-mediated defense response in tobacco cell culture revealed overlap between race-specific resistance and response to wounding [22 , 23] . Additionally , partial genome microarray analysis of the Arabidopsis wound response revealed that a number of wound-responsive genes encode proteins known to be involved in pathogen defense [24] . Examination of the AtGenExpress abiotic datasets demonstrates that the initial transcriptional abiotic stress response may comprise a core set of multi-stress-responsive genes . The abiotic stress response then becomes stress specific at later time points [25–27] . Finally , recent analysis of the AtGenExpress abiotic and biotic datasets has uncovered ∼200 genes that are expressed in response to a broad range of stresses , which may represent the GSR of Arabidopsis [28] . Recently , a shift in stress tolerance engineering has been proposed that transfers the focus from pathway endpoints to factors governing upstream reactions . Focusing on upstream signaling components may enable the engineering of multi-stress tolerance [20 , 29] . Identification of cis-regulatory elements for use in synthetic promoters to confer stress tolerance has also recently been proposed [30 , 31] . Towards this aim , we have utilized mechanical wounding , as it uniquely confers an instantaneous and synchronous stimulus , to identify primary stress-responsive transcripts . Comparison of the 5 min rapid wound response ( RWR ) genes we identified with published transcript profiles demonstrated a large overlap with previously identified abiotic and biotic stress-responsive genes . Notably , RWR genes also exhibit a striking level and pattern of circadian regulation . Further investigation via real-time quantitative RT-PCR ( RT-qPCR ) of a wounding time course revealed genes that are expressed rapidly and transiently as well as rapidly and stably . Two rapidly and transiently expressed genes , ETHYLENE RESPONSE FACTOR #018 ( ERF#018; AT1G74930 ) and CCR4-ASSOCIATED FACTOR 1 ( CAF1-like; AT3G44260 ) , were confirmed as wound and biotic stress inducible in vivo using stable transgenic lines expressing transcriptional luciferase fusions . Detailed analysis of the RWR promoters identified a novel cis-regulatory element we term the rapid stress response element ( RSRE ) , which is sufficient to confer reporter gene induction in response to abiotic and biotic stress . RWR genes identified in this study may represent initial components of the GSR and be useful in engineering multi-stress tolerance .
To identify primary stress-responsive transcripts we utilized Agilent microarrays to monitor gene expression changes 5 min after mechanical wounding of Arabidopsis rosette leaves . Because of the short duration of our stress treatment we hypothesized that expression changes would be low . In order to accurately detect these changes we utilized three biological replicates of pooled plants per treatment . In addition , two technical replicates , with dye swap of each technical replicate , were performed on each biological replicate . Using this approach , we found that the expression of 162 genes was upregulated and the expression of 44 genes was downregulated at least 2-fold and had a p-value ≤ 0 . 01 five min after mechanical wounding ( Table S1 ) . The expression level of selected RWR genes representing a range of high-to-low-fold change was then validated using RT-qPCR . The expression changes determined by RT-qPCR data are in good agreement with the fold change observed by microarray with a spearman rank order correlation coefficient of 0 . 927 ( p-value = 0 . 000 ) ( Figure 1A ) . RWR genes were then classified according to gene ontology ( GO ) terms in order to provide insight into their biological function [32] . The two largest defined classes of GO terms involve response to stress or abiotic/biotic stimuli ( Figure 1B ) . It is also of interest to note that genes classified for an involvement in signal transduction were observed in upregulated but not downregulated RWR genes . These data indicate that 5 min of mechanical wounding was sufficient for induction of known stress-responsive genes as well as unknown genes that may play a role in multi-stress responses . We next created stable transgenic lines expressing transcriptional fusions of the ERF#018 and CAF1-like promoters to luciferase to validate in vivo RWR genes and to investigate their temporal expression pattern . For each construct , three independent T2 lines were imaged to control for positional effects of the transgene insertion site . Luciferase activity was then monitored following the wounding of a single leaf per plant to enable the observation of whether the induced activity occurred only locally or also systemically . The wound-induced expression of PERF18:LUC occurs rapidly and peaks ∼1 h 45 min after the wound stimulus ( Figure 2A and 2B ) . Additionally , expression of PERF18:LUC was observed in the petiole and shoot apex . The expression of PCAF1-like:LUC was detected rapidly and peaked ∼1 h 25 min surrounding the wound site ( Figure 2C and 2D ) . These data provide in vivo confirmation that RWR genes do respond rapidly to mechanical wounding . To gain further insight into how the RWR genes may be acting , we performed a RT-qPCR time-course on selected genes . We classified genes as rapidly and stably expressed if 60 min post wounding they remained greater than 2-fold induced . In contrast , we classified genes that had decreased in expression to less than half of maximal expression by 60 min post wounding as rapidly and transiently expressed . Among the rapidly and stably expressed transcripts are genes with either a known or predicted role in stress signal transduction events ( Figure 3 ) . CML38 is a calmodulin-like gene that is predicted to be a sensor of Ca2+ , a known secondary messenger of stress responses [33] . MPKK9 , a MAPK signal transduction component , was also identified as rapidly and stably expressed . Additionally , the transcription factor WRKY40 , which is known to be involved in pathogen defense , was identified [34] . Finally , BAP1 , a negative regulator of defense responses whose binding of phospholipids is enhanced by calcium , was shown to respond rapidly and stably to wounding [35] . The upregulation of RWR genes in a rapid and stable manner may indicate that these genes play a more prolonged role in response to stress . We also uncovered rapidly and transiently expressed RWR genes with a wide range of functions ( Figure 4 ) . One such example is the chromatin remodeling ATPase SPLAYED ( SYD ) , which peaks 15 min post wounding . Because of the large changes in gene expression following stress it has been hypothesized that chromatin remodeling may be required to allow for stress-induced transcription to occur [36] . Examples of stress-induced changes in histone acetylation state in plants have been described [37 , 38] . Rapid and transient upregulation of SYD suggests that ATP-dependent chromatin remodeling may also take place in order to facilitate downstream stress-induced transcriptional changes . Genes involved in signal transduction via reversible phosphorylation were also upregulated rapidly and transiently following wounding ( Figure 4 ) . One kinase identified was MPK3 , a MAPK signal transduction component , which has been shown to function in innate immunity and stomatal development [39 , 40] . AP2C1 , a PP2C-type phosphatase with a MAPK interaction motif , was also shown to exhibit a rapid and transient expression pattern resulting from wounding [41] . These results indicate that both phosphorylation and dephosphorylation of MAPK signaling components is involved in transduction of initial stress signaling events . A third process implicated by genes identified in this study is that of mRNA turnover ( Figure 4 ) . Specifically , this process is demonstrated by the expression pattern of CAF1-like . In yeast , CAF1 has been shown to be a component of the major cytoplasmic deadenylase , which functions to remove the poly ( A ) tail , thereby initiating mRNA turnover [42] . Additionally , the mouse CAF1 ortholog has been demonstrated to function as a 3′-5′-RNase with a preference for poly ( A ) substrates [43] . In Arabidopsis , RNA processing appears to play a role in response to cold stress [44] . It is also of interest to note the difference in promoter and transcript expression patterns in response to wounding for ERF#018 and CAF1-like . When promoter activity was monitored using transcriptional luciferase fusions , activity peaked ∼1 h 30 min after wounding ( Figure 2 ) . In contrast , transcript abundance measured via RT-qPCR peaked 15–30 min after wounding ( Figure 4 ) . This discrepancy may be an artifact due to measuring promoter activity via luciferase protein activity , while RT-qPCR assayed transcript levels . An alternative explanation of these results is that mRNA levels of rapidly and transiently expressed genes such as ERF#018 and CAF1-like are controlled post-transcriptionally , possibly via mRNA turnover . In support of this hypothesis , CAF1-like mRNA has a half-life of 38 min and has been classified as a gene with an unstable transcript [45] . Furthermore , the mRNA of a second rapid and transient gene , MPK3 , has a half-life of 43 min and is classified as a gene with an unstable transcript [45] . Examination of the RWR genes reveals a large number of known genes involved in abiotic and biotic stress responses ( Table S1 ) . Among the upregulated RWR genes were genes involved in ethylene signaling including ACC synthase 6 ( ACS6 ) as well as 15 of the 122 ethylene response factors ( ERFs ) in Arabidopsis , which have been shown to be involved in the response to both biotic and abiotic stresses [46] . The transcriptional activators CBF1 ( DREB1B ) , CBF2 ( DREB1C ) , and CBF3 ( DREB1A ) , which confer tolerance to cold and drought , were also among the RWR genes [47–50] . Additional RWR genes known to confer tolerance to a range of abiotic stresses include STZ ( ZAT10 ) and ZAT12 [51 , 52] . RWR genes also include genes with a known function in response to biotic stress . Examples include BAP1 , a negative regulator of defense responses [35] . MPK3 and FLS2 which function in response to pathogen-associated molecular patterns in the Arabidopsis innate immune response [39 , 53] . The transcription factor TGA3 which regulates pathogenesis-related ( PR ) genes and is required for basal pathogen resistance was also among the RWR genes [54] . Finally , ERD15 regulates not only cold and drought tolerance but also resistance to the bacterial necrotroph Erwinia carotovora subsp . carotovora [55] . The abundance of RWR genes with known abiotic and biotic stress tolerance functions led us to examine the role of wounding as a general stress perception mechanism on a global level . For this analysis , we compared the overlap in gene lists between the RWR genes and published transcript profiles for a number of stress conditions . The statistical significance of the observed overlap in transcript profiles was then analyzed using empirical permutation tests [56] . We first compared RWR genes with published abiotic microarrays and found a strong overlap ( unpublished data ) , which is in agreement with work recently published by Kilian et al . [25] . For example , 49% of upregulated RWR genes have been previously shown to be upregulated upon cold treatment [57 , 58] . Additionally , four of the nine genes ( At1g27730 , At5g51190 , At5g47230 , and At5g04340 ) found by Kilian et al . [25] to be upregulated by 30 min of cold , drought , UV-B , salt , osmotic stress treatment , and wounding we discovered to be upregulated within 5 min of wounding . We next compared the RWR transcript profile with published transcript profiles of plants challenged with different biotic stresses [59–61] . For upregulated datasets there was a statistically significant overrepresentation of RWR genes in the transcript profile of all biotic stresses tested ( Figure 5 ) . Furthermore , the overrepresentation of RWR genes occurred only at early time points of P . rapae and OGA stressed plants . These data indicate that perception of mechanical stress may play a central role in the perception and initial response to a wide range of environmental stresses . Various biological compounds are known to elicit stress-signaling networks . Due to the overlap between RWR genes and biotic stresses we tested whether the RWR genes ERF#018 and CAF1-like respond to the biological elicitors OGA and insect regurgitant ( IR ) as well as cabbage looper ( Trichoplusia ni ) feeding . We first tested whether PERF18:LUC or PCAF1-like:LUC activity was induced by cabbage looper feeding . Indeed , cabbage looper feeding did result in enhanced luciferase activity , which verified that biological stress does induce ERF#018 and CAF1-like ( Videos S1 and S2 ) . We therefore proceeded to test induction resulting from OGA and IR treatment . When OGA , IR , or H2O were added to a nonwounded ( NW ) leaf , no induction of PERF18:LUC or PCAF1-like:LUC activity was observed ( Figure 6A–6F ) . The lack of induction is likely due to the application method we used ( single droplet per plant ) rather than an actual lack of response to the elicitors . When transcript profiling was performed on liquid cultured 10-d-old plants incubated with 50 μg ml−1 OGAs , CAF1-like was shown to be induced 1 h after addition of OGA to the media [61] . In PERF18:LUC-expressing plants , addition of OGA and IR to the wound site resulted in a significantly greater ( p < 0 . 05 ) induction of luciferase activity than addition of H2O in both local and systemic tissue ( Figure 6B and 6C , respectively ) . In contrast , addition of OGA or IR to the wound site did not result in a significant difference ( p > 0 . 05 ) in PCAF1-like:LUC activity compared to addition of H2O to the wound site in local tissue ( Figure 6E ) . However , a significantly greater induction , compared to H2O , in PCAF1-like:LUC activity resulted from addition of OGA or IR to the wound site in systemic tissue ( Figure 6F ) . There are a number of common second messengers downstream of mechanical wounding , cabbage looper feeding , and OGA treatment that may signal for the observed induction of RWR genes . One such secondary messenger is Ca2+ , which increases in intracellular concentration rapidly following wounding as well as OGA treatment [3 , 16 , 62 , 63] . ROS are another secondary messenger that have been shown to increase in response to chewing insects , wounding , and OGA treatment [3 , 12] . Furthermore , while OGAs do not move systemically , ROS do accumulate systemically following wounding . This increase in ROS is likely through OGAs released by systemically induced polygalacturonase [14 , 16 , 64–66] . Finally , OGA , chewing insects , and wounding may all have a common mechanism of perception resulting in similarly induced secondary messengers . Both chewing insects and wounding have a physical effect on the plasma membrane . The perception of OGA has also been hypothesized to be a result of its physical effect on the plasma membrane , rather than through an actual receptor [3 , 67] . The circadian clock has been shown to regulate a number of environmentally regulated genes [68] . Additionally , cold-induced expression of RWR genes ZAT12 , CBF1 , CBF2 , and CBF3 was recently reported to be gated by the circadian clock [69] . These findings led us to examine globally whether RWR genes are under circadian regulation . Towards this aim , we compared the RWR genes with genes recently identified as circadian regulated [56] . Surprisingly , not only were RWR genes rhythmically expressed but the upregulated and downregulated genes also showed unexpected phase distributions ( Figure 7 ) . Forty-two percent of RWR upregulated genes are expressed at subjective dusk while 81% of downregulated RWR genes peak at subjective dawn ( p ≤ 0 . 0001 ) . The circadian regulation of RWR genes may provide a mechanism to anticipate stresses caused by daily environmental changes . To begin dissecting the molecular mechanism underpinning the rapid stress response , we examined the promoters of the RWR genes for novel cis-regulatory elements . We identified the six-nucleotide repeat , CGCGTT , which we are terming the Rapid Stress Response Element ( RSRE ) , as significantly overrepresented ( 58 hits in 47 of the 162 upregulated promoters ) in the promoters of upregulated RWR genes . To determine whether the RSRE is sufficient alone to confer stress-responsive transcription , we used luciferase reporter constructs . Four tandem repeats of the RSRE and its consensus flanking sequence were separated by six nucleotides and cloned upstream of the minimal promoter region of the nopaline synthase ( NOS ) gene and modified luciferase coding region ( 4xRSRE:LUC ) . Additionally , to verify that the RSRE was the region conferring stress responsiveness , we mutated three of the six nucleotides in the RSRE ( 4xmtRSRE:LUC ) . The wound-induced expression of these constructs was then tested in 24 independent T1 plants to control for differences in expression resulting from the site of transgene insertion . All 24 4xRSRE:LUC transgenic plants exhibited wound-induced luciferase expression ( Figure 8 ) . Furthermore , luciferase activity increased immediately following wounding and peaked ∼80 min post wounding . Conversely , no 4xmtRSRE:LUC transgenic plant exhibited luciferase activity before or after wounding . These data demonstrate that the RSRE is sufficient to confer a rapid response to stress . Both abiotic and biotic stresses appear to share common signaling components with the RWR . We were therefore interested in whether the RSRE confers a rapid response to a range of stresses . To enable accurate quantification of the stress response we used a homozygous T3 4xRSRE:LUC line . Because RWR genes respond to both OGAs and IR ( Figures 3 and 7 ) , we tested the expression of 4xRSRE:LUC following treatment with these biotic elicitors . In the local leaf where the wound site was treated with OGA or IR , no induction over that of H2O treatment was observed ( Figure 9A and 9B ) . Notably , in systemic tissues , a statistically significant ( p < 0 . 05 ) synergistic enhancement of luciferase activity was detected in OGA- and IR-treated plants when compared to H2O treatment ( Figure 9A and 9C ) . To further demonstrate that the RSRE responds to biotic stress , 4xRSRE:LUC plants were challenged with cabbage loopers and B . cinerea . When 4xRSRE:LUC plants were exposed to both forms of biotic stress , luciferase activity was induced , whereas no activity was observed in vector control lines ( Videos S3 and S4 ) . Additionally , in B . cinerea–infected plants , a low level of transient luciferase activity was first observed in the inoculated leaf . Luciferase activity was then observed at a greater level in systemic tissues ( Video S4 ) . These data clearly demonstrate that the RSRE responds to biotic stress . While the RSRE responds to the abiotic stress of mechanical wounding , we wished to further demonstrate the role of the RSRE in response to abiotic stress . Towards this aim , we exposed 4xRSRE:LUC expressing plants to 5 °C . Plants were then removed from cold treatment at the indicated time for imaging . Additionally , control 4xRSRE:LUC plants were also kept at 22 °C in equivalent light conditions and moved similarly to cold-treated plants to ensure that transfer to the imaging chamber did not result in induced luciferase activity . Induction of luciferase activity was observed after ∼2 h of cold treatment ( Figure 10A and 10B ) . Furthermore , luciferase activity peaked after 5 h of cold treatment and then decreased towards basal expression levels . Notably , 4xRSRE:LUC expression was also observed in the roots of cold-treated plants . To ensure that the 4xRSRE:LUC plants were still competent to express luciferase , they were mechanically wounded after 120 h of cold treatment . Both cold- and 22 °C-treated 4xRSRE:LUC plants exhibited luciferase induction following wounding ( unpublished data ) . These data demonstrate that the RSRE is cold responsive and that initial signaling events leading to a cold response dampen even in continuous exposure to cold . The rapid and transient response of the RSRE to multiple stress conditions is reminiscent of the yeast GSR promoter element STRE ( stress response element; AGGGG ) [18] . The STRE is responsible for rapid induction following various treatments such as heat , nitrogen starvation , low external pH , osmotic , and oxidative stress [70–73] . Furthermore , even in the presence of continuous stress exposure , STRE-mediated gene induction dampens over time . An increase in unsaturated fatty acids upon stress appears to be responsible for the transient nature of STRE-mediated induction [74 , 75] . When plants are exposed to abiotic and biotic stresses ( cold and P . syringae , respectively ) , there is an increase in unsaturated fatty acids [9 , 76] . Upon cold treatment , acyl-lipid desaturases are the enzymes that most efficiently introduce double bonds in membrane lipids , which results in the increased level of unsaturation [9] . Similar to cold treatment , two of the RWR upregulated genes are acyl-lipid desaturases ( ADS1 and ADS2 ) , which may increase the unsaturation of membranes upon wounding . It is therefore tempting to speculate that , as with the STRE in yeast , the increase in unsaturated membrane lipids resulting from both abiotic and biotic stresses may mediate the transient induction of RSRE-driven reporter gene expression . We have shown that 5 min of mechanical stress is a sufficient amount of time for the plant to perceive the stress and mount a robust transcriptional response . The rapid transcriptional response to mechanical wounding shares a large overlap with both abiotic and biotic stresses and may therefore represent the initial GSR of Arabidopsis . In support of this view , the RWR upregulated genes comprised 25% of the genes identified as potential GSR genes via analysis of the AtGenExpress abiotic and biotic stress datasets [28] . Additionally , in mammalian cells , physical stress to membranes during osmotic and UV radiation stress result in the nonspecific clustering of growth factor receptor tyrosine kinases and cytokine receptors [18 , 77] . A similar nonspecific clustering of receptors during mechanical wounding and other environmental stresses may underlie the GSR of plants . We also show that the RWR genes are circadian regulated with consolidated phases of peak expression . Circadian regulation of RWR genes , which likely encompass initial components of the GSR , may enable plants to anticipate daily environmental changes and mount a general defense against these changes . Finally , we identified a cis-regulatory element ( RSRE ) overrepresented in the promoters of RWR genes . The RSRE confers a rapid and transient response similar to the yeast GSR promoter element ( STRE ) and is a novel GSR cis-regulatory element in plants . Since the RWR genes likely represent initial components of the GSR , they provide a valuable resource of candidate genes for engineering of multi-stress resistance . Similarly , the RSRE , which responds rapidly and transiently to abiotic and biotic stresses , may prove useful as a synthetic element for engineering of multi-stress tolerance . Finally , use of the RSRE in yeast one-hybrid and the 4xRSRE:LUC line for mutant screens should help elucidate the upstream mechanisms of stress perception and initial signal transduction .
Arabidopsis ( Arabidopsis thaliana ) ecotype Columbia-0 plants were grown in a 16 h light/8 h dark photoperiod at 22 °C . All experiments were conducted on 3-wk-old soil-grown plants unless otherwise noted . All rosette leaves ( unless otherwise noted ) were mechanically wounded one to two times with a hemostat ( resulting in ∼20% of the leaf being damaged ) . Mechanical wounding was performed 4–6 h after dawn . For cloning of PERF18:LUC , the 1 . 4 kb upstream of the translation start site of ERF#018 ( At1g74930 ) was PCR amplified using primers listed in Table S2 . For cloning of PCAF1-like:LUC , the 2-kb upstream region of the translation start site of CAF1-like ( AT3G44260 ) was PCR amplified using primers listed in Table S2 . The PCR amplicons were cloned into the pENTR/D-TOPO vector and subcloned into the Gateway destination vector pBGWL7 [78] by an LR reaction ( Invitrogen , Carlsbad , CA ) . 4xRSRE:LUC ( 5′-cataaCGCGTTtttagatatcataaCGCGTTtttaggatccataaCGCGTTtttatctagaataaCGCGTTtttac-3′ ) and 4xmtRSRE:LUC ( 5′-cataaCATGCTtttagatatcataaCATGCTtttaggatccataaCATGCTtttatctagaataaCATGCTtttac-3′ ) constructs were created by cloning into the SacI/XhoI sites of pATM-Nos [79] . Transformations were performed into Columbia-0 plants by floral dip using Agrobacterium tumefaciens strain GV3101 [80] . The Arabidopsis 2 . 0 oliogoarray chip containing 60-mer oligos and representing a total of 21 , 500 probes ( TAIR ATH1 v4 . 0 ) was obtained from Agilent ( G4137A; Wilmington , Delaware ) . Total RNA from leaf tissue was isolated using TRIzol reagent ( Invitrogen ) following the manufacturer's suggested protocol . Prior to hybridizations , the quality and quantity of the total RNA sample was confirmed by running 10-ng samples on an Agilent bioanalyzer ( RNA chip ) , and by using a spectrophotometer . The oligoarray hybridization experiments utilized three biological replicates of pooled plants ( ∼40 plants per pool ) . Each biological replicate was comprised of two technical replicates with dye reversal . Total RNA ( 500 ng ) was used as a template for cRNA production and cyanine dyes were incorporated using the Agilent low RNA input linear amp kit ( Agilent ) . Normal yields from 500 ng of total RNA input using a 4-h in vitro transcription were 15 μg cRNA ( 15 pmol cyanine dye incorporated/μg cRNA ) . One microgram of labeled cRNA ( cy3- and cy5-labeled sample ) was diluted to 175 μl and defragmented at 60 °C for 30 min following the Agilent hybridization protocols ( Agilent ) . Defragmented samples were diluted to 500 μl ( 30% formamide final concentration ) and hybridized for 20 h at 40 °C . Arrays were washed and dried and scanned on an Agilent G2565BA microarray scanner [81] . The raw TIFF images were analyzed using the Agilent Feature Extraction software v . 8 . 1 using the recommended default settings . The intensities of Cy3- and Cy5-labeled probes were normalized by comparing signal intensities of housekeeping genes ( positive controls ) for both dyes and using the determined ratio as a correction factor for differences in labeling efficiencies ( Agilent Feature Extraction v . 8 . 1 software ) . The genes that had valid signal in all six replicates were exported to Rosetta Resolver software and analyzed according to the manufacturer's instructions ( Seattle , Washington ) . The normalized values were used to calculate the ratio of channel intensities ( Cy5/Cy3 ) , which were then log10 transformed . The transformed ratios were plotted in a scatter plot ( cy5/cy3 ) A ±1 . 7-fold increase or decrease in signal intensity ( p-chance value , 0 . 01 ) “Signature” Features ( >1 . 5-fold deviation from the median , p-chance value 0 . 01 ) was exported into Microsoft Excel for further analysis . In Excel , data were sorted and genes with a fold-change ≥ 2 . 0 were selected as differentially regulated . Additionally , differentially expressed genes were binned based on signal intensity . RT-qPCR analysis was conducted based on Yamagishi et al . [82] Total RNA was isolated by TRIzol extraction ( Life Technologies , Grand Island , NY ) and treated with DNase , MseI , and DdeI to control for DNA contamination . One microgram of RNA was reverse transcribed using Superscript III ( Invitrogen , Carlsbad , California ) . PCRs were conducted using 12 . 5 μl of SYBR Green mix ( 40 mM Tris HCL pH 8 . 4 , 100 mM KCl , 6 mM MgCl2 , 8 % glycerol , 20 nM fluorescein , 0 . 4× SYBR Green [Molecular Probes , Eugene , Oregon] , 100-fold dilution of BSA [New England Biolabs , Beverly , MA] , and 1 . 6 mM dNTP ) , 2 μl of a 30-fold dilution of the RT reaction , 0 . 6 U iTaq DNA Polymerase ( Bio-Rad Laboratories , Hercules , CA ) , 9 . 38 μl H2O , and 0 . 24 μM of each primer . Reactions were carried out on a Bio-Rad iCycler iQ multicolor real-time detection system ( Bio-Rad Laboratories ) using a two-step reaction condition ( extension temperature was primer specific but was typically ∼60 °C ) , followed by a melt curve encompassing 80 steps of 0 . 5 °C from 60 °C to 100 °C . Gene-specific primers were designed using Beacon Designer software ( Premier Biosoft Palo Alto , CA ) and are listed in Table S2 . Primary data analysis was performed with Bio-Rad iCycler iQ software . The Bio-Rad gene expression macro version 1 . 1 software ( Bio-Rad Laboratories ) was used to calculate relative RNA levels normalized to an internal control [83–85] . The 60S ribosomal protein L14 ( At4g27090 ) described in [82] or the TIP41-like gene ( At4g34270 ) described in [86] was used as an internal control . For this analysis , we compared the overlap in gene lists between the RWR genes and published transcript profiles determining circadian-regulated genes as well as abiotic and biotic stress-responsive genes . The statistical significance of the observed overlap in transcript profiles was then analyzed using a recently described empirical permutation test [56] based on sampled randomization testing [87] with a p-value cutoff of p < 0 . 0001 . For luciferase imaging , 10–14-d-old plants grown on plates containing 1× Murashige and Skoog basal salt mixture ( Sigma ) were utilized . Plants were sprayed with 2 . 5 mM luciferin ( Promega , Madison , WI ) in 0 . 001% Triton X-100 ∼16–20 h prior to treatment . Mechanical wounding was performed on a single leaf per plant . For elicitor application , 5 μl of 30 mg/ml OGA , 1μl of cabbage looper ( Trichoplusia ni ) regurgitant , or 5 μl of sterile ddH2O was applied to a single leaf . For cold treatment , plants were placed in a 5 °C chamber under low light . Control plants were kept at 22 °C in equivalent light conditions and handled similarly to cold-treated plants to ensure that transfer to the imaging chamber did not result in induced luciferase activity . Plants were then removed from cold treatment for imaging and returned to either cold or 22 °C . Five microliters of Botrytis cinerea isolate KB2 in 1/2× grape juice was inoculated on a single leaf at a concentration of 500 spores/μl [88 , 89] . Plants were imaged using an Andor DU434-BV CCD camera ( Andor Technology , South Windsor , CT ) . For image acquisition , PERF18:LUC or PCAF1-like:LUC plants were exposed for 20 min while 4xRSRE:LUC plants were exposed for 5 min . Luciferase activity was quantified for a defined area ( leaf , shoot apex , or whole plant ) as mean counts pixel−1 exposure time−1 using Andor Solis image analysis software ( Andor Technology , South Windsor , CT ) . For statistical analysis of treatment effects , the area under the curve was calculated and compared by Kruskal-Wallis one way ANOVA on ranks , with pairwise multiple comparisons ( Student-Newman-Keuls Method ) , using Sigma Stat v3 . 5 ( San Jose , CA ) . A 1% ( w/v ) solution of citrus pectin ( Sigma-Aldrich ) in 0 . 5 M HCl was refluxed for 3 h at a rolling boil . Following cooling , the sample was neutralized with NaOH , decolorized using activated charcoal , and dialyzed against water using 6 , 000 molecular weight tubing . The dialyzed sample was then lyophilized [90] . Promoter sequences were defined as a fixed distance ( 2 kb for the purpose of motif detection ) upstream of the annotated translation start codon , as described in Hudson and Quail [90] . An enhanced enumerative motif recognition algorithm was developed based on that described by Hudson and Quail [90] . The enhanced method does not require exact motif matching , and permits degenerate bases to occur in any number at any position through the search motif by searching for and enumerating all possible permutations of bases within the specified motif size limits including wildcard ( N ) bases . Significant associations between promoter gene lists and all permutations of motifs are then detected by comparison of per-promoter motif abundance between the promoters of the target coregulated gene list and the promoters of all genes in the Arabidopsis genome . The per-promoter binomial test described [91] is used to rank motifs in order of significance , with the exception that a more efficient factorial handling algorithm was used to perform the binomial test on every motif present in the list of coregulated promoters , rendering the preliminary chi-square filtering step unnecessary . Finally , motifs with related sequence are aligned and automatically clustered for output based on nucleotide-level identity . Programs were written in the Perl programming language utilizing the GMP Multiple Precision Arithmetic Library [92] . The Arabidopsis promoter motif search program is available for use via a web interface at http://stan . cropsci . uiuc . edu/tools . php . Source code is available from MEH on request .
The data discussed in this publication have been deposited in the National Center for Biotechnology Information ( NCBI ) Gene Expression Omnibus ( GEO , http://www . ncbi . nlm . nih . gov/geo/ ) and are accessible through GEO Series accession number GSE8740 . | Plants are sessile organisms constantly challenged by a wide spectrum of biotic and abiotic stresses . These stresses cause considerable losses in crop yields worldwide , while the demand for food and energy is on the rise . Understanding the molecular mechanisms driving stress responses is crucial to devising targeted strategies to engineer stress-tolerant plants . To identify primary stress-responsive genes we examined the transcriptional profile of plants after mechanical wounding , which was used as a brief , inductive stimulus . Comparison of the ensemble of rapid wound response transcripts with published transcript profiles revealed a notable overlap with biotic and abiotic stress-responsive genes . Additional quantitative analyses of selected genes over a wounding time-course enabled classification into two groups: transient and stably expressed . Bioinformatic analysis of rapid wound response gene promoter sequences enabled us to identify a novel DNA motif , designated the Rapid Stress Response Element . This motif is sufficient to confer a rapid response to both biotic and abiotic stresses in vivo , thereby confirming the functional involvement of this motif in the primary transcriptional stress response . The genes we identified may represent initial components of the general stress-response network and may be useful in engineering multi-stress tolerant plants . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
]
| [
"arabidopsis",
"(thale",
"cress)",
"plant",
"biology"
]
| 2007 | Mechanical Stress Induces Biotic and Abiotic Stress Responses via a Novel cis-Element |
Natural immunity or resistance to pathogens most often relies on the genetic make-up of the host . In a LEW rat model of refractoriness to toxoplasmosis , we previously identified on chromosome 10 the Toxo1 locus that directs toxoplasmosis outcome and controls parasite spreading by a macrophage-dependent mechanism . Now , we narrowed down Toxo1 to a 891 kb interval containing 29 genes syntenic to human 17p13 region . Strikingly , Toxo1 is included in a haplotype block strictly conserved among all refractory rat strains . The sequencing of Toxo1 in nine rat strains ( 5 refractory and 4 susceptible ) revealed resistant-restricted conserved polymorphisms displaying a distribution gradient that peaks at the bottom border of Toxo1 , and highlighting the NOD-like receptor , Nlrp1a , as a major candidate . The Nlrp1 inflammasome is known to trigger , upon pathogen intracellular sensing , pyroptosis programmed-cell death involving caspase-1 activation and cleavage of IL-1β . Functional studies demonstrated that the Toxo1-dependent refractoriness in vivo correlated with both the ability of macrophages to restrict T . gondii growth and a T . gondii-induced death of intracellular parasites and its host macrophages . The parasite-induced cell death of infected macrophages bearing the LEW-Toxo1 alleles was found to exhibit pyroptosis-like features with ROS production , the activation of caspase-1 and IL1-β secretion . The pharmacological inactivation of caspase-1 using YVAD and Z-VAD inhibitors prevented the death of both intravacuolar parasites and host non-permissive macrophages but failed to restore parasite proliferation . These findings demonstrated that the Toxo1-dependent response of rat macrophages to T . gondii infection may trigger two pathways leading to the control of parasite proliferation and the death of parasites and host macrophages . The NOD-like receptor NLRP1a/Caspase-1 pathway is the best candidate to mediate the parasite-induced cell death . These data represent new insights towards the identification of a major pathway of innate resistance to toxoplasmosis and the prediction of individual resistance .
Toxoplasma gondii is a widespread obligate intracellular protozoan parasite . One preeminent aspect of its life cycle is the establishment of a chronic infection in humans and many other vertebrate hosts [1] . Toxoplasmosis is most often asymptomatic depending on the parasite's ability to elicit host protective immunity [1] . A serious threat to human health can occur under congenital infection or reactivation of a latent infection in immunodeficient patients [2] . Epidemiological studies have indicated that the phenotypic expression of toxoplasmosis depends on the genetic make-up of both the host and the parasite [3] , [4] . Variations in the outcome of Toxoplasma infection after exposure to similar risk factors [5] , [6] and twin studies [7] support a significant role of the human host genetic background in the susceptibility to toxoplasmosis . Nevertheless , genetic studies in human are hampered by both population heterogeneity and environment variability . In experimental conditions , genetic and environmental factors are under control . Rats , like humans , usually develop subclinical toxoplasmosis . This contrasts with the severity of the disease developed in most strains of mice . Interestingly , an unexpected refractoriness to T . gondii infection was found in the LEW rat strain [8] . Compared to susceptible BN rats , infected LEW indeed displayed negative serology and lack of cyst burden in their brain [9] . Refractoriness of LEW rats was found to be a dominant trait dependent on hematopoietic cells [9] . It is associated with the ability of macrophages to restrict parasite proliferation in vitro [10] . Further genetic studies using LEW resistant and BN susceptible rats and derived reciprocal congenic strains have allowed the mapping of a locus named Toxo1 , which fully controls the refractoriness of LEW rats to toxoplasmosis . Toxo1 has been confined to 7 . 6 megabases , on rat chromosome 10 ( Rn10q . 24 ) [10] . Recently , hNlrp1 a major candidate gene present in the orthologous region to Toxo1 in the human genome ( Hs 17p32 . 2-p13 . 1 ) has been associated with human congenital toxoplasmosis [6] . In the present work , we used genetic dissection with a panel of BN and LEW sub-congenic rats and haplotype analysis of chromosome 10 on nine inbred rat strains either susceptible or resistant to define the localization of the gene or set of genes at work in Toxo1 and to analyze the mechanisms of toxoplasmosis refractoriness . We were able to localize the Toxo1 locus in a 891 kb region highly conserved in all resistant strains of rat . Sequencing of this locus in these nine strains revealed a high concentration of resistant-restricted conserved mutations at the bottom border of Toxo1 around Nlrp1 . Functional studies in ex vivo infected peritoneal macrophages indicate that the Toxo1-mediated restriction of parasite proliferation is associated with the coordinate death of both parasites and host macrophages . The parasite-induced macrophage cell death involved a caspase-1 dependent mechanism and exhibited pyroptosis-like features . The parasite-induced killing of infected macrophages could be blocked by the capase-1 inhibitor without restoration of parasite proliferation . Therefore , we concluded that if at work , the NLRP1a/caspase-1 pathway is not indispensable to restrict parasite proliferation in macrophages .
We previously demonstrated that the Toxo1 7 . 6 Mb interval fully controls the outcome of T . gondii infection independently of the genetic background . The refractoriness to infection conferred by the LEW origin of Toxo1 is characterized by the early elimination of the pathogen resulting in a barely detectable specific immune response and in the absence of brain cysts [10] . In vitro , this Toxo1-LEW mediated refractoriness is associated with the control of parasite proliferation within macrophages [10] . To refine the localisation of the gene ( s ) that control ( s ) these in vivo and in vitro phenotypes , we generated a unique panel of congenic sub-lines . Results from the genetic dissection are shown on Figure 1 . The parasites were found able to proliferate within the macrophages from the congenic BN . LEWc10-Ce , -Cf , -Cga , -Ci and LEW . BNc10-F sub-lines but not within the macrophages from the congenic BN . LEWc10-Cg and -Ch sub-lines ( Figure 1A ) . Thus within the 7 . 6 Mb of the Toxo1 locus a 891 kb region controls the in vitro proliferation of parasites within macrophages . We further investigated refractoriness or susceptibility to T . gondii infection in vivo in rats from the seven congenic sub-lines used for these in vitro studies as well as in rats from the BN and LEW parental strains . The control of refractoriness to T . gondii defined by both the absence or low specific antibody response ( Figure 1B ) and the absence of cyst burden in the brain ( Figure 1C ) , was directed by the same 891 kb region ( Figure 1D ) . Thus , the interval located between the D10GF49 ( 57 . 26 Mb ) and D10GF55 ( 58 . 15 Mb ) microsatellite markers contains the gene or the set of genes that controls the toxoplasmosis outcome . We hypothesized that genetic variation ( s ) underlying Toxo1-mediated innate refractoriness against T . gondii infection result ( s ) from an ancestral polymorphism instead of independent newly acquired mutations in the LEW strain and thus could be identified in various inbred rat strains . To challenge this hypothesis , we investigated toxoplasmosis outcome in seven other inbred rat strains ( LOU , WF , WK , BDIX , OM , F344 and DA ) , in comparison to BN and LEW . Following infection , rats exhibited a dichotomous phenotype either developing high titers of anti-toxoplasma antibodies and cerebral cysts or remaining refractory to parasite infection ( Figure 2 ) . Three strains ( OM , DA and F344 ) displayed , like the BN rat , phenotypes associated to chronic infection with high anti-T . gondii antibody responses ( ≥5000 u . a ) and the detection of brain cysts . Of note , the number of brain cysts was significantly different among these rat strains ( Figure 2B ) reflecting other regulatory mechanism ( s ) for cyst formation . By contrast , the four other rat strains ( LOU , BDIX , WK and WF ) showed the refractory phenotype to T . gondii infection of the LEW strain neither developing specific antibodies nor cerebral cysts ( Figure 2 ) . As expected , parasite proliferation within peritoneal macrophages was observed only in the BN and the three other susceptible strains but neither in LEW nor in the four other refractory strains ( Figure 2C ) . Thus , refractoriness against T . gondii infection is a common trait of several inbred rat strains . To investigate the implication of the Toxo1 locus in resistance against T . gondii infection , we used a targeted method of genotyping to stratify ( LOU x BN ) F2 rat according to their genotype at Toxo1 . Thirty five F2 rats were genotyped at the polymorphic marker D10GF41 , located at the peak of the Toxo1 locus , and their anti-T . gondii Ab titers and brain cyst numbers were quantified following infection with T . gondii ( Figure 3 ) . All rats sharing the two LOU alleles ( ll ) showed no detectable or a weak anti-T . gondii Ab response ( <10 , 000 a . u . ) . Conversely , all rats sharing the two BN alleles ( nn ) had high anti-T . gondii Ab titers ( >20 , 000 a . u . ) . The apparent intermediate antibody response observed in the 15 heterozygous rats ( nl ) was not significantly different from that observed in the panel of rats sharing the LOU alleles ( ll ) . In a similar way , brain cysts were observed in all the ten homozygous BN ( nn ) rats and in none of both the heterozygous BN/LOU ( nl ) and the nine homozygous LOU ( ll ) . These results showed that Toxo1 directs in a dominant manner the toxoplasmosis outcome in LOU rats after T . gondii infection . As Toxo1 controls resistance against T . gondii infection in the LEW and LOU rats , and given the similarities between the response of LEW , LOU , BDIX , WK and WF , we next examined if Toxo1- allelic similarities are conserved among all those resistant animals . Considering that as a result of common ancestry , patterns of allelic similarities and differences among strains can be discerned for every variable locus [11] , we conducted a haplotype study on the nine strains previously described . This analysis , based on allele size data for microsatellite markers , consisted in identifying among the different strains the chromosomal regions with LEW genotype . For this purpose , 41 polymorphic microsatellite markers on chromosome 10 were investigated on these nine strains ( Table S1 ) . The genotype of the nine strains for these markers showed the conservation of an haplotype block in the five resistant strains ( LEW , LOU , WF , WK and BDIX ) between D10Arb7 and D1Rat297 as compared to the four susceptible strains ( BN , OM , F344 and DA ) ( Figure 4 ) . This highly conserved haplotype block extends on 2 . 8 Mb between D10Arb7 and D1Rat297 and overlaps the entire 891 kb Toxo1 locus ( Figure 4 ) . Thus , the data clearly suggested that conserved genetic variations within the Toxo1 locus on chromosome 10 is a common cause of resistance against T . gondii infection in inbred rat strains . According to the genome database ( www . ensembl . org , RGSC3 . 4 version ) , the Toxo1-891 kb interval contains 29 genes . None of these 29 genes could be retained as a candidate on the basis of a significant difference in their level of expression between macrophages from resistant vs . macrophages from susceptible congenic lines ( Table S2 , TextS1 ) . Therefore , the entire Toxo1 locus of the nine rat strains studied in the haplotype analysis was sequenced to identify resistance-correlated variations in coding- and non-coding sequences . A total of 373 SNPs and 21 insertions/deletions were found strictly conserved among the five resistant strains as compared to susceptible strains . The distribution of these mutations along the locus displays a gradient with a densification at the bottom of Toxo1 ( Figure 5 ) . We identified 23 SNPs of which 16 are missense and one is a deletion in the coding sequences leading to the selection of four candidate genes: Inca1 ( 1 SNP ) , Kif1C ( 1 ins/del ) , Nlrp1a ( 13 SNPs ) and Nlrp1b ( 2 SNPs ) . Given that Inca1 and Nlrp1b mRNAs are undetectable in peritoneal macrophages ( Table S2 ) and according to the number of mutations in Nlrp1a vs Kif1C coding sequences , Nlrp1a appeared as the major candidate gene . It encodes the NOD-like receptor ( NLR ) NLRP1 that acts as an intracellular pattern recognition receptor ( PRR ) [12] . Both the mouse Nlrp1b and its ortholog rat Nlrp1a have been described as implicated in the control of a cell death process called pyroptosis that is induced by the lethal toxin ( LT ) from Bacillus anthracis [13] . Based on NLRP1 known function , we investigated the impact of the Toxo1 locus on parasite and host cell fate after ex vivo infection of peritoneal macrophages . Following host-cell invasion , T . gondii replicates within a newly formed non-fusogenic compartment , the parasitophorous vacuole ( PV ) . Using fluorescent parasites and vacuole staining with anti-GRA5 or -GRA3 antibodies which both stain the PV membrane , we compared the fate of intravacuolar parasites by immunofluorescence microscopy after allowing them to invade either non-permissive LEW or permissive BN naive peritoneal macrophages ( Figure 6 ) . The analyses were performed at 2 and 8 hours following invasion using transgenic-YFP2 fluorescent parasites . According to our previous work [10] we found similar rates of parasite invasion in both permissive ( Toxo1-BN ) ( 15%±5% ) and non-permissive ( Toxo1-LEW ) ( 17%±6% ) naive peritoneal macrophages . At 2 hours post-infection , most of intracellular YFP2-parasites were found within a compartment positive for the GRA5 PV membrane marker , in both LEW and BN macrophages ( Figure 6A ) indicating that parasites are able to enter efficiently into resistant macrophages . However , we observed a slight decrease in the percentage of YFP- and GRA5- positive vacuoles within LEW ( 77%±4 ) as compared to BN ( 89%±6 ) macrophages ( Figure 6B ) . At 8 hours post-infection , while YFP2-parasites started to divide within GRA5 positive vacuoles of permissive BN macrophages ( Figure 6A ) , a dramatic drop of YFP staining was observed in GRA5-positive vacuoles of LEW ( 21%±4 ) as compared to BN ( 66%±6 ) macrophages ( Figure 6A and B ) . The loss of YFP emission in GRA5-positive vacuoles was likely to reflect the death of parasites [14] although parasite egress could not be excluded . Interestingly , the difference was not so marked when the parasite surface was stained using anti-SAG1 antibodies . Indeed , in BN macrophages , the percentage of vacuoles containing SAG1-positive parasites ( 62 . 5%±1 ) was not statistically different from the percentage of vacuoles containing YFP2-parasites ( 66%±6 ) . In contrast , in LEW macrophages the percentage of vacuoles containing SAG1-positive parasites ( 43%±3 ) was two-fold the percentage of vacuoles containing YFP2-parasites ( 21%±4 ) , indicating that at least 2/3 of vacuoles in LEW macrophages still contained parasites ( Figure 6C and D ) . Therefore , the decrease of YFP staining could be attributed , at least in majority , to parasite death resulting from the macrophage microbicidal activity , rather than to parasite egress . The parasite death within the PV was further examined using staining of small ubiquitin-related modifier ( SUMO ) as a read-out of the transcriptional activity of live parasites [15] . As shown in Figure 6E and 6F , 2 and 8 hours after infection , a dramatic drop in the percentage of SUMO positive vacuoles was observed in LEW macrophages while no difference was found in BN macrophages , thus supporting the early death of parasites within LEW macrophages . We finally examined whether this phenotype was under the Toxo1 control with our collection of sub-congenic rat strains . For this purpose , peritoneal macrophages from the four susceptible ( BN . LEWc10-Ce , -Cf , -Cga , -Ci and LEW . BNc10-F ) and the two refractory ( BN . LEWc10-Cg and -Ch ) sub-congenic lines were infected with T . gondii . In macrophages from the LEW and BN . LEWc10-Cg , -Ch lines in which Toxo1 is from LEW origin , the inhibition of parasite proliferation correlates with the induction of intracellular parasite death ( LEW: 58±13%; BN . LEWc10-Cg: 69±2%; -Ch: 64±7% ) . Conversely , in macrophages from the BN . LEWc10-Ce , -Cf , -Cga and -Ci lines in which Toxo1 is from BN origin , the parasite proliferation was associated with a decrease of parasite death ( BN . LEWc10-Ce: 17±11%; -Cf: 0±21%; -Cga: 3±7%; -Ci: 1±17%; LEW . BNc10-F: 0±16% ) ( Figure 6G ) . Altogether these results demonstrated that the Toxo1 locus from LEW origin mediates the T . gondii killing by infected peritoneal macrophages . We next examined the fate of infected host cells in a comparative way depending on the Toxo1 genotype . At 8 hours post-infection , we observed that most of LEW macrophages infected with YFP-negative vacuoles presented a condensed nucleus ( Figures 6A , 6C , 6E ) . Indeed , when this phenomenon was quantified we found that 65% of LEW macrophages with YFP-negative vacuoles , but only 14% of BN susceptible macrophages , showed a condensed nucleus ( Figure 7A ) . This observation indicating a parasite-induced killing of LEW macrophages was further investigated using propidium iodide ( PI ) uptake . At 6 hours post-infection , about 40% of LEW macrophages ( 39%±12 ) had lost membrane integrity as compared to less than 10% of BN macrophages ( 6%±8 ) ( Figure 7B , p<0 . 05 ) . The lack of PI uptake under incubation with lysed fibroblasts ruled out the possible triggering by unrelated pathogen-associated molecular pattern ( PAMP ) ( Figure 7B ) . Tracking of both PI uptake and YFP loss in kinetic experiments indicated that the parasite-induced cell death of LEW macrophages is concomitant with the death of intracellular parasites . Indeed , the increase of PI uptake by macrophages and the loss of YFP by intracellular parasites started from 1 hour and increased steadily until 6 hours after infection ( Figure 7C ) . Further quantification of PI and GRA5 PV marker positive LEW macrophages demonstrated that 88% of dying cells were parasite-invaded cells ( Figure 7D ) . Finally , the implication of Toxo1 in the host cell death phenotype was validated using our panel of sub-congenic animals . Specific induction of host cell death after infection of non-permissive peritoneal macrophages was found in all resistant congenic rats ( LEW: 30±10%; BN . LEWc10-Cg: 35±3%; -Ch: 24±3% ) . By contrast , peritoneal macrophages from all animals permissive to T . gondii proliferation failed to initiate such a cell death process after infection ( BN . LEWc10-Ce: 11±1%; -Cf: 8±3%; -Cga: 10±4%; -Ci: 13±7% and LEW . BNc10-F: 17±5% ) ( Figure 7E ) . Altogether , these data demonstrated that T . gondii invasion of resistant macrophages is rapidly followed by the combined deaths of intracellular parasites and infected host macrophages using a mechanism under the control of Toxo1 locus . The rapid loss of membrane integrity of parasite-invaded LEW macrophages ( Figure 7C ) suggested that the T . gondii-induced cell death was not apoptotic . Accordingly , DNA from dying cells did not show the laddering resulting from chromatin fragmentation , observed in classical apoptotic cell death and pyroptosis ( Figure 8A ) . Moreover , caspase-3 activation , which plays a central role in the executive phase of apoptosis , was not observed in infected LEW macrophages ( Figure 8B ) . Additionally , when infected macrophages were incubated with FITC-labelled annexin V , phosphatidylserine exposure on the plasma membrane was not observed prior to the loss of membrane integrity as assessed by PI staining ( Figure 8C ) . Finally , no difference in the number of acidic vacuoles was observed after lysotracker coloration between permissive BN and non-permissive infected LEW macrophages indicating that the death of infected LEW macrophages was not due to autophagy ( Figure 8D ) . Given the described role of NLRP1/Caspase-1 inflammasome pathway in the host response to pathogens [16] , [17] , we hypothesized that the T . gondii-induced cell death of infected resistant LEW macrophages could be associated with both ROS production and caspase-1 activation . The production of intracellular ROS by macrophages was monitored at two time points ( 15 min and 4 hours ) by the dihydro-rhodamine 123 . While T . gondii infection did not induce significant ROS production within permissive LEW . BNc10-F macrophages , a marked increase was recorded in infected resistant LEW macrophages ( Figure 9A ) . We next examined caspase-1 activation within infected macrophages by using the fluorogenic activated caspase-1 specific staining ( FLICA ) . At 4 hours post-infection , the percentage of parasite-induced FLICA positive cells was significantly higher in resistant LEW macrophages ( 29%±2 ) than in permissive LEW . BNc10-F macrophages ( 11%±1 ) ( Figure 9B and C ) . The caspase-1 induction in infected LEW macrophages correlated with the increase of PI-positive cells indicating that caspase-1 was involved in the cell death induction process ( Figure 9D ) . Consistent with these results , the processing of caspase-1 substrate IL-1β that could be prevented by the YVAD caspase-1 inhibitor was detected at 1 h and more evidently at 4 h post-infection in the culture supernatant of infected LEW macrophages and not in that of permissive LEW . BNc10-F macrophages ( Figure 9E ) . The observed difference in the secretion of mature IL-1β was not due to a lack of pro-IL-1β expression since it was found to be induced in response to T . gondii infection within both resistant LEW and permissive LEW . BNc10-F macrophages ( Figure 9E ) . By contrast , pro-IL-1β was not detectable in the cell lysate from LEW macrophages treated with YVAD ( Figure 9E ) indicating that the inhibition of caspase-1 activity resulted in the down-regulation of pro-IL-1β protein expression . Altogether , these results demonstrated that both ROS production and caspase-1/IL1β pathway are involved in the Toxo1-mediated cell death induction of resistant LEW macrophages . The role of caspase-1 in the resistance of LEW macrophages was further investigated using pharmacological inhibitors . Both caspase-1 and pan-caspase inhibitors were able to protect the resistant LEW macrophages from the parasite-induced cell death ( Figure 10A ) . Consistent with our above observations that both parasites and host macrophages killing processes are connected , the two inhibitors also prevented the death of intracellular parasites ( Figure 10B ) . By contrast , these inhibitors failed to restore significant parasite proliferation in non-permissive macrophages ( Figure 10C ) . Altogether these experiments revealed that the resistance of macrophages bearing Toxo1-LEW alleles relies on two pathways , the first one controlling parasite proliferation and the second one controlling the death of both intracellular parasites and host macrophages , via caspase-1 dependent inflammasome activation .
Forward genetics has proved to be a powerful tool to characterize novel biological pathways implicated in host resistance to infection [18] , [19] . In rats , the Toxo1 locus located on chromosome 10 [10] controls the outcome of toxoplasmosis by a still poorly defined mechanism . In the present work , genetic dissection of Toxo1 with a panel of new congenic sub-lines together with haplotype mapping led us to identify a 891 kb interval of rat chromosome 10 that is highly conserved amongst resistant rat strains and that controls both T . gondii infection outcome in vivo and macrophage responses to infection in vitro . We further demonstrated that the Toxo1-mediated refractoriness of macrophages to T . gondii infection is associated with a caspase-1-dependent rapid T . gondii-induced death of both intracellular parasites and host macrophages . The Toxo1-dependent death of infected macrophages displayed the chromatin condensation hallmark of apoptosis but neither caspase-3 activation nor DNA fragmentation were observed . In contrast to apoptosis which is usually a slow process characterized by membrane blebbing , the Toxo1-associated cell death was characterized by a rapid loss of plasma membrane integrity that occurs simultaneously with surface exposure of phosphatidylserine . A non-apoptotic pathway triggered in T . gondii-infected macrophages has been described in mice [20] . It is mediated by IFN-γ-inducible immunity-related GTPases that trigger vacuole membrane disruption and the death of parasites , followed by the necrotic-like death of infected cells . While phenotypically , the IRG-dependent mouse T . gondii/macrophage deaths parallel the Toxo1-mediated T . gondii/macrophage deaths , major differences between the two in vitro models exist , providing strong evidences that mechanistically they are not identical . The mouse IRG-controlled mechanism requires IFN-γ stimulation and occurs in both fibroblasts and macrophages . In contrast , the Toxo1 effect in rats is strictly confined to hematopoeitic cells [9] , [10] and does not depend on IFN-γ stimulation of macrophages in vitro . Moreover , the mouse IRG-controlled resistant system is genotypically restricted to non-virulent genotype II parasites , while in rats , both type I RH and type II PRU strains triggered the LEW T . gondii/macrophage deaths ( Figure S1 ) . In line with this , the effector of the LEW rat resistance is not the kinase parasite effector ROP 18 ( data not shown ) which , by phosphorylating IRGs , disrupts their association with the parasite-containing vacuole and thereby protects the parasite against elimination [21] , [22] . Altogether , the rat Toxo1 locus-mediated resistance to toxoplasmosis does not operate via the IRG-dependent system . Sequencing of the Toxo1 region in all five resistant rat strains ( LEW , LOU , BDIX , WK and WF ) and four susceptible strains ( BN , OM , DA and F344 ) revealed a gradient in the conserved distribution of resistant-restricted mutations that peaks at the bottom of Toxo1 and particularly in the Nlrp1a coding sequence ( 23 SNPs ) . This highly conserved region has been previously demonstrated to be associated with the resistance of rats to the lethal Toxin ( LT ) of Bacillus anthracis [23] . Interestingly , while rat bearing the divergent Toxo1 alleles were susceptible to LT-mediated death , the highly conserved Toxo1-LEW alleles correlated with the total resistance of rats [23] . Moreover , similarly to what we found for the Toxo1 control of T . gondii infectivity , there was a perfect correlation between the in vivo phenotype ( sensitivity vs resistance to LT ) and the in vitro phenotype of macrophages , suggesting that the same gene or set of genes might be at work in the control of these two pathogens . In mice and rats , further genetic mapping associated genetic variants of both mNlrp1b and one of its rat ortholog rNlrp1a to macrophage sensitivity or resistance to LT by a mechanism dependent on a cell death process called pyroptosis , which is also triggered upon infection with intracellular pathogens such as Salmonella , Shigella or Listeria [24]–[26] . NLRP1 is part of the NLR family , which is known to act as an intracellular sensor for cytoplasmic danger signals [12] . After activation , the NLR form a multimeric protein complex called the inflammasome that provides a scaffold for the activation of caspase-1 and target death substrates by a still poorly understood mechanism [12] . The Toxo1 controlled-cell death induction that is triggered in peritoneal macrophages following T . gondii invasion , appeared to be also caspase-1 dependent and associated to IL-1β secretion and ROS production . Together , with the nuclear condensation , it thus features the hallmarks of the pyroptosis programmed-cell death which is uniquely dependent on caspase-1 and inherently proinflammatory . However , while pyroptosis is typically associated to DNA fragmentation [16] , this later event was not observed in the T . gondii-induced cell death possibly due to a yet unexplained undetectable PARP ( Poly ADP-Ribose Polymerase ) expression in rat peritoneal macrophages ( Figure S3 , Text S1 ) . Together these observations combined with the genetics studies tend to support that the NOD-like receptor NLRP1a/Caspase-1 pathway is the best candidate to mediate the Toxo1-dependent parasite-induced cell death . Despite these remarkable genetic and biochemical similarities , LT- and T . gondii-induced phenotypes display striking major differences . First , while Toxo1-conserved LEW alleles are associated to the cell death induction of macrophages upon T . gondii infection , same allelic variants protect the macrophages from LT-mediated cell death [23] . Secondly , while the pharmacological inhibition of host cell death also prevents the concomitant death of parasites , it failed to restore the permissiveness of macrophages to parasite proliferation . Thus , although caspase-1 is induced following T . gondii infection , our data do not argue for its essential role in the macrophage control of parasite proliferation in vitro . They rather suggest that , like it is emerging in the case of intracellular bacteria , the Toxo1-directed resistance of macrophages to T . gondii proliferation may rather be a complex trait resulting from the combined activation of several pathways . For instance , the Naip5/Nlrc4-controlled restriction of Legionella pneumophila growth within non permissive murine macrophages is likely to involve both canonical pyroptosis and a yet undefined caspase-1 independent Naip5 pathway [27] . Very interestingly , in mouse macrophages infected by Shigella flexneri which display NLRC4- and NLRP3-dependent activation of caspase-1 , IL-1β/IL-18 processing and cell death [17] , [26] , it has been demonstrated that inflammasome might negatively regulate pathogen-induced autophagy [26] . Given that other NLRs may affect autophagy [17] , inflammasome inhibition of autophagy might be a generalized mechanism . Moreover , Harris J . and al . demonstrated that autophagy controls IL1-β secretion by targeting pro-IL1-β for degradation [28] . In line with this , the lack of detectable pro-IL1-β in resistant LEW macrophages treated with caspase-1 inhibitor , suggested that inactivation of caspase-1 results in the down-regulation of pro-IL1-β protein expression . It is therefore possible that in our model , the inhibition of caspase-1 could block the negative regulation of inflammasome promoting thus the effect of other pathway ( s ) capable to restrict parasite proliferation without pyroptosis . Altogether , our work provide evidence that natural resistance of rat macrophages to T . gondii is a complex trait relying on a mechanism involving the combined activation of at least two pathways: ( i ) the classical NLRP1a/Caspase-1 pathway , mediating host and parasite cell death and ( ii ) a yet undefined pathway that would directly control parasite proliferation within the parasitophorous vacuole . The parasite effector ( s ) eliciting these pathways and their possible interconnections remain to be investigated . Toxoplasma infection is naturally acquired by the oral route . Following transcytosis across the intestinal barrier [29] , the tachyzoite stage encounters leukocytes in which it replicates to further disseminate into the organism using the migratory properties of infected macrophages and dendritic cells [30] . In rats bearing the LEW-type Toxo1 locus , the refractoriness to toxoplasmosis is evidenced by the absence of both local parasite burden and specific antibody response [9] . Thus resistance likely results from the rapid clearance of the parasite following a vigorous killing response at the site of infection . Our data led us to propose a model where the early death of infected macrophages impairs further dissemination of the parasite , hence constituting an efficient barrier against successful infection . How the Toxo1 locus controls dendritic cell and monocyte responses after infection remains to be characterized . In conclusion , this work highlighted several novel aspects of the host-parasite gene interaction . We unambiguously mapped the Toxo1 locus to an 891 kb region which directs the outcome of toxoplasmosis in the rat . This locus controls parasite infectivity in vivo and is critically associated to macrophage-dependent restriction of parasite intracellular growth and cell death induction of both intracellular parasites and infected macrophages in vitro . It is included within a haplotype block remarkably subjected to strong selection pressure for resistant strains , in contrast to susceptible strains ( Figure S2 , Text S1 ) . The robustness of Toxo1-mediated resistance in controlling parasite infectivity indicates that this resistance might have been conserved among naturally resistant species [11] . Hence , Transmission Disequilibrium Test studies revealed that hNlrp1 , the human ortholog of rat rNlrp1a , has alleles associated with the susceptibility to human congenital toxopolasmosis [6] . In the same way , recent work demonstrated that , in mice , NLRP1 is an innate immune sensor for Toxoplasma infection inducing a host-protective innate immune response to the parasite [31] . Altogether , these data identified a genetically-controlled major pathway of innate immunity to toxoplasmosis allowing predicting resistance of individuals . The results open the way to further investigations towards the gene ( s ) and the mechanisms at work , and could be applied to human toxoplasmosis in regards to the conserved synteny of Toxo1 region between rat and human .
Breeding and experimental procedures were carried out in accordance with national and international laws for laboratory animal welfare and experimentation ( EEC Council Directive 2010/63/EU , September 2010 ) . Experiments were performed under the supervision of M–F . C–D . ( agreement 38 10 38 ) in the Plateforme de Haute Technologie Animale ( PHTA ) animal care facility ( agreement n° A 38 516 10006 delivered by the Direction Départementale de la Protection des Populations ) and were approved by the ethics committee of the PHTA ( permits n° Toxo-PC-1 and n° Toxo-PC-2 ) . Production and genotype analyses of congenic lines were as described previously [10] . The congenic lineages were maintained by regular brother-sister mating . LEW/OrlRj ( LEW ) , BN/OrlRj ( BN ) male rats were obtained from Janvier Laboratory ( Le Genest-Saint-Isle , France ) . WF/N ( WF ) , WKY/NHsd ( WK ) , BDIX/Han ( BDIX ) , F344/Nhsd ( F344 ) , Lou/CNimrOlaHsd ( LOU ) and DA/OlaHsd ( DA ) male rats were obtained from Harlan Laboratory ( Gannat , France ) . OM/Han ( OM ) rats were kindly supplied by the Hannover Medical School ( Germany ) . F2 ( LOU×BN ) progenies were produced in our animal facilities under specific pathogen-free conditions . Cysts from the recombinant T . gondii Prugniaud strain were used to test in vivo the susceptibility to toxoplasma infection . Two-month-old Swiss mice ( Janvier laboratory ) were infected orally with 10 Prugniaud cysts . Their brains were collected 3 months later and ground in a Potter . Cysts were counted in a Thoma's cell and diluted in PBS . Rats were infected orally with 20 cysts . One month later , blood was collected from the retro-orbital sinus for detection of anti-toxoplasma Ab response by ELISA . Rats were euthanized 2 months after infection and brains were collected to determine number of cysts . Tachyzoites of Prugniaud type II strain , and RH , RH-YFP2 ( kindly provided by B . Striepen , Athens ) and RH-mcherry ( kindly provided by A . Bougdour , Grenoble ) type I T . gondii strains were maintained under standard procedures , by serial passage onto human foreskin fibroblast monolayers ( HFFs ) in D10 medium ( DMEM supplemented with 10% heat-inactivated fetal bovine serum , 1 mM glutamine , 500 units . ml−1 penicillin and 50 µg . ml−1 streptomycin ) at 37°C in a humidified atmosphere containing 5% CO2 . The parasites were collected just before the experiment , centrifuged at 500× g for 7 min , suspended in serum-free medium ( SFM , GIBCO ) supplemented with 500 units . ml−1 penicillin and 50 µg . ml−1 streptomycin , and counted . Rat resident peritoneal cells were obtained by injection of sterile PBS into the peritoneal cavity . Collected cells were centrifuged and resuspended in Serum Free Medium ( SFM ) ( Life Technologies , Inc ) and counted . Macrophages were obtained by adhering cells for 1 h at 37°C and 5% CO2 . After 1 h , non-adherent cells were removed by gentle washing with SFM and parasites were added to macrophages settled on coverslips at a ratio of 3∶1 . After incubation for 1 h at 37°C , wells were washed 3 times with SFM to remove extracellular parasites and cells fixed at different times post-infection in 4% formaldehyde . Caspase-1 inhibitor VI ( YVAD ) and caspase inhibitor VI ( pan-caspase , Z-VAD ) were from Calbiochem ( Merck Chemicals , France ) . Macrophages were incubated with 50 µM of Caspase-1 inhibitor or 100 µM of pan-caspase inhibitor for 2 h before infection and during all the infection . Infected macrophages were permeabilized with 0 . 002% saponine or 0 . 1% triton-×100 to detect SAG1 ( mAb Tg05-54 ) , GRA5 ( mAb Tg17-113 ) or GRA3 ( mAb Tg2H1 ) . The rabbit anti-Tg small ubiquitin-like modifier ( TgSUMO ) polyconal antibody was kindly provided by M . A . Hakimi ( Grenoble , France ) [15] . To detect acidic vacuoles , infected macrophages were stained with 50 nM LysoTracker Red for 30 min prior to fixation . Cell death was analyzed by visualizing the uptake of Propidium Iodide ( PI ) ( Molecular Probes ) . Alexa488 and Alexa594 antibody conjugates ( Molecular Probes ) were used as secondary antibodies . Coverslips were mounted in mowiol and observed and counted with a Zeiss Axioplan 2 microscope equipped for epifluorescence and phase-contrast . Kinetic experiments were performed on the IX2 Olympus microscope and analyzed with ScanR software . For analysis of caspase-1 activation , we used a fluorescent caspase-1 activity assay , FLICA , from Immunochemistry Technologies ( Bloomington , MN ) . The assay was performed in 24-well plates , with 5 . 105 cells per well . Cells were incubated with T . gondii and stained with FLICA reagent ( FAM-YVAD-FMK ) as recommended by the manufacturer . Fluorescence was measured on the IX2 Olympus microscope and analyzed with ScanR software . The level of apoptosis of infected LEW macrophages was assessed with Annexin V-propidium iodide staining . Rat peritoneal exudent cells were treated with etoposide ( Sigma Aldrich , 50 µM ) or incubated with parasites in SFM at a MOI of 1∶3 at 37°C , 5% CO2 . After 1 h , cells were collected by centrifugation and incubated 4 h at 37° after addition of fresh medium . Macrophages were stained by the addition of Annexin V ( Santa Cruz biotechnologies , FL-319 ) and Alexa488 secondary antibody ( Molecular Probes ) for 30 min and 1 µg/mL of PI for 5 min . Data acquisition was performed by flow cytometry on 4-colours FACSCalibur ( BD Biosciences ) equipped with 488 nm argon laser and CellQuest Software . Rat peritoneal macrophages ( ∼106 ) were infected with 3×106 parasites for 1 h , washed and returned to 37°C for 6 h , 12 h or 18 h . For positive control of apoptosis , macrophages were treated with 50 µM etoposide ( Sigma Aldrich ) . Macrophages were treated in lysis-buffer ( 100 mM Tris , pH 8 . 0 , 100 mM EDTA , 4% Sodium dodecyl sulfate ( SDS ) with 1 µg/ml RNAse ( Roche ) ( 30 min , 37°C ) followed by treatment with 100 µg/ml proteinase K ( Euromedex ) ( 30 min , 55°C ) . After extraction with phenol/chloroform , DNA was recovered by precipitation and analysed on 1 . 8% agarose gels . Rat peritoneal exudent cells were incubated with parasites in SFM at a MOI of 1∶3 at 37°C , 5% CO2 for 15 min or 4 h and then incubated with 0 . 45 µM of dihydro-rhodamine 123 ( DHR , Sigma Aldrich ) 15 min at 37°C . At the end of incubation , FACS lysing buffer ( BD Bioscience , Pont de Claix , France ) was added to each sample and incubated for 15 min . The samples were then washed in PBS before analysis with a FACSCalibur flow cytometer ( Becton Dickinson ) and the CellQuest Pro software ( Becton Dickinson ) . Cells were lysed in cold RIPA ( 50 mM Tris-Hcl pH 7 , 4 , 150 mM NaCl , 1% NP40 , 0 , 25% Na-deoxycholate ) buffer supplemented with protease inhibitors and centrifuged at 4°C and 13 , 000 g for 10 min . Supernatant of infected or uninfected macrophages ( 5 . 105 cells/well ) were collected and precipitated with TCA ( trihloroacetate ) for 10 min at 4°C prior centrifugation at 16 000 g for 5 min . Pellets were washed two times in acetone then dried and resuspended in laemmli buffer . Protein extracts were subjected to electrophoresis on a 12% Tris-HCl SDS-PAGE and transferred to PVDF membranes ( Amersham ) . Membranes were blocked for 1 h in TTBS ( 100 mM Tris-HCl , 0 . 9% NaCl , and 0 . 05% Tween 20 ) containing 5% skim milk before incubating overnight at 4°C with primary 1/500 anti-caspase 3 ( Cell Signaling ) , 1/1000 anti-IL-1β ( Millipore , AB1832P ) , 1/5000 anti-GAPDH ( Santa Cruz ) or 1/1000 anti-actin ( Sigma Aldrich ) antibodies followed by 1/10000 anti-rabbit secondary horseradish peroxidase ( HRP ) -linked antibodies ( Jackson Immunoresearch ) . Visualization of signals was enhanced by luminol-based chemiluminescence ( ECL , ThermoFisher Scientific ) . The intracellular growth of T . gondii in rat peritoneal macrophages was monitored by selective incorporation of [3H]uracil as previously described [32] . Briefly , 5×105 macrophages were infected with 1 . 5×106 parasites for 1 h in SFM at 37°C and 5% CO2 . After washing to eliminate extracellular parasites , cells were cultured for 20 h in the presence of [3H]uracil ( 5 µCi per well , Ci = 37 GBq ) . Monolayers were washed three times in PBS , disrupted with 500 µl of lysis/scintillation solution ( Optiphase Supermix , Perkin Elmer ) and radioactivity measured by liquid scintillation counting using a Wallac MicroBeta TriLux ( Perkin Elmer ) . Preparation of genomic DNA and genotyping were performed as described [33] . To genotype the 35 ( LOU X BN ) F2 rat progenies , six microsatellite markers ( D10Rat49 , D10Arb2 , D10GF41 , D10Rat27 , D10Mgh4 , D10Rat2 ) were selected to cover chromosome 10 with an average spacing of 20 Mb . For haplotype analysis , the polymorphism of 41 microsatellite sequences around the Toxo1 locus were analysed in the nine strains . Among these 41 markers , 17 newly identified microsatellite markers ( Table S1 ) were used in addition to the 24 microsatellite markers selected from Rat Genome Database ( RGD ) . The anti-Toxoplasma IgG response was measured by specific enzyme-linked immunosorbent assay ( ELISA ) . Total Toxoplasma antigens were prepared as previously described [34] . Immuno plates Maxisorp ( Nunc ) were coated overnight at 4°C with Toxoplasma antigens at 20 µg/ml . After washing , saturation was 1 h at 37°C with PBS containing 5% of skim milk . Sera were diluted at 1/20 and 1/1000 in PBS-0 . 01% Tween 20 and incubated 1h30 at 37°C . Plates were then washed with PBS-0 . 01% Tween 20 , and peroxydase-conjugated anti-rat IgG ( KPL ) secondary antibody diluted at 1/5000 was incubated 1 h at 37°C . Finally , after six washes , 100 µl of substrate TMB-hydrogen peroxyde ( TMB Ultra 1 Step , ThermoScientific ) solution was added to the wells . The color reaction was stopped adding 50 µl of 3 N HCl . Optical densities at 492 nm and 630 nm were measured using an ELx800 absorbance microplate reader ( Bio TeK Instruments ) . Results were expressed as arbitrary units . Each rat brain was removed and homogenized in 16 ml of PBS . Brain suspensions were clarified by gentle incubation in proteinase K buffer ( proteinase K 0 . 4 µg/ml , 10 mM Tris pH 8 , EDTA 1 mM , sodium dodecyl sulfate 0 . 2% , sodium chloride 40 mM ) for 15 min at 56°C . The reaction was stopped by incubation with PMSF 2 mM for 5 min at room temperature . Then , the suspension was washed with PBS and resuspended with FITC-Dolichos biflorus agglutinin ( Vector laboratories , CA USA ) 20 µg/ml for 30 min , at room temperature . After washing in PBS , rat brains were resuspended in 6 ml of PBS , distributed into six-well culture plates ( 1 ml per well ) and cysts counted visually with an Axiovert 40 CFL inverted fluorescence microscope ( Zeiss ) [35] . A custom-made SureSelect oligonucleotide probe library was designed by IntegraGen ( Evry , France ) to capture the chr10 region containing Toxo1 locus ( chr10: 57 , 200 , 000–58 , 200 , 000 ) . The eArray web-based probe design tool was used for this purpose ( https://earray . chem . agilent . com/earray ) . A total of 56 , 450 probes , covering a target of 6830692 bp , were synthesized by Agilent Technologies ( Santa Clara , CA , USA ) . Library preparation , capture enrichment , sequencing , and variants detection and annotation , were performed by IntegraGen ( Evry , France ) . Briefly , 3 µg of each genomic DNA were fragmented by sonication and purified to yield fragments of 150–200 bp . Paired-end adaptor oligonucleotides from Illumina were ligated on repaired DNA fragments , which were then purified and enriched by six PCR cycles . 500 ng of the purified libraries were hybridized to the SureSelect oligo probe capture library for 24 h . After hybridization , washing , and elution , the eluted fraction underwent 14 cycles of PCRamplification . This was followed by purification and quantification by qPCR to obtain sufficient DNA template for downstream applications . Each eluted-enriched DNA sample was then sequenced on an Illumina GAIIx as paired-end 75 bp reads . Image analysis and base calling was performed using Illumina Real Time Analysis ( RTA ) Pipeline version 1 . 10 with default parameters . Sequence reads were aligned to the reference rat genome ( UCSC rn4 ) using commercially available software ( CASAVA1 . 7 , Illumina ) and the ELANDv2 alignment algorithm . Sequence variation annotation was performed using the IntegraGen in-house pipeline , which consisted of gene annotation ( RefSeq ) , detection of known polymorphisms ( dbSNP 125 ) followed by mutation characterization ( exonic , intronic , silent , nonsense etc . ) . For in vivo experiments , data are expressed as means ± SEM and the significance of differences found between groups was initially derived from a Kruskal-Wallis H test and subsequently confirmed by the Mann-Whitney test . For in vitro experiments , data are expressed as means ± SD and the significance of differences found between groups was determined using two-tailed Student's t test . | Toxoplasmosis is a ubiquitous parasitic infection causing a wide spectrum of diseases . It is usually asymptomatic but can lead to severe ocular and neurological disorders . The host factors that determine natural resistance to toxoplasmosis are yet poorly characterized . Among the animal models to study susceptibility to toxoplasmosis , rats develop like humans a subclinical chronic infection . The finding of a total resistance in the LEW rat strain has allowed genetic studies leading to the identification of Toxo1 , a unique locus that controls the outcome of toxoplasmosis . In this report , a panel of recombinant inbred rat strains was used to genetically reduce the Toxo1 locus , on chromosome 10 , to a limited region containing 29 genes . This locus is highly conserved among five resistant , by comparison to four susceptible , rat strains , indicating that refractoriness to toxoplasmosis could be predicted . The Toxo1-controlled refractoriness depends on the ability of macrophages to restrict parasite proliferation and the rapid death of both T . gondii and host macrophages in vitro . The NOD-like receptor NLRP1a/Caspase-1 pathway is the best candidate to mediate the parasite-induced cell death . Our data represent new insights towards the identification of a major pathway of innate immunity that protects from toxoplasmosis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"sequencing",
"techniques",
"genome",
"expression",
"analysis",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"animal",
"genetics",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"microbiology",
"vertebrates",
"animals",
"mammals",
"parasitology",
"animal",
"models",
"genome",
"sequencing",
"model",
"organisms",
"genome",
"analysis",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"pathogenesis",
"molecular",
"biology",
"immune",
"system",
"rodents",
"immunity",
"host-pathogen",
"interactions",
"transcriptome",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"genetics",
"of",
"disease",
"computational",
"biology",
"organisms",
"rats"
]
| 2014 | A Highly Conserved Toxo1 Haplotype Directs Resistance to Toxoplasmosis and Its Associated Caspase-1 Dependent Killing of Parasite and Host Macrophage |
The intra-S phase checkpoint kinase of metazoa and yeast , ATR/MEC1 , protects chromosomes from DNA damage and replication stress by phosphorylating subunits of the replicative helicase , MCM2-7 . Here we describe an unprecedented ATR-dependent pathway in Tetrahymena thermophila in which the essential pre-replicative complex proteins , Orc1p , Orc2p and Mcm6p are degraded in hydroxyurea-treated S phase cells . Chromosomes undergo global changes during HU-arrest , including phosphorylation of histone H2A . X , deacetylation of histone H3 , and an apparent diminution in DNA content that can be blocked by the deacetylase inhibitor sodium butyrate . Most remarkably , the cell cycle rapidly resumes upon hydroxyurea removal , and the entire genome is replicated prior to replenishment of ORC and MCMs . While stalled replication forks are elongated under these conditions , DNA fiber imaging revealed that most replicating molecules are produced by new initiation events . Furthermore , the sole origin in the ribosomal DNA minichromosome is inactive and replication appears to initiate near the rRNA promoter . The collective data raise the possibility that replication initiation occurs by an ORC-independent mechanism during the recovery from HU-induced replication stress .
A major challenge of the cell cycle is to faithfully transmit chromosomes to daughter cells . This is accomplished through the replication and segregation of chromosomes during the respective S and M phases . The integrity of chromosomes is under constant assault from extrinsic and intrinsic sources that directly damage DNA or generate roadblocks for the replication machinery . The resulting DNA damage and replication stress can irreparably harm chromosomes . Checkpoint pathways have evolved to combat these problems , arresting the cell cycle when thresholds are exceeded . The phosphatidylinositol-3-OH kinase family members ATM ( Ataxia Telangiectasia Mutated ) and ATR ( ATM-and Rad3-related ) function as apical kinases in signal transduction pathways that inhibit DNA replication and promote DNA repair [1] . ATM is activated by double strand breaks ( DSBs ) and arrests the cell cycle at the G1/S border , while ATR monitors replication stress and other types of DNA damage during S phase . ATR is recruited to exposed stretches of single strand DNA through association with its binding partner , ATRIP . A key target of ATR is the MCM2-7 complex . Phosphorylation of the replicative helicase suppresses fork elongation , and prevents new origins from firing [2–4] . Once the source of stress is removed , fork elongation rates are restored and new initiation events occur at late replicating or dormant origins [5–7] . While the proteins that elicit checkpoint responses are conserved , there are fundamental differences in how eukaryotes deal with DNA damage . For example , while hydroxyurea ( HU ) inhibits initiation from late firing origins and slows replication fork elongation in S . cerevisiae [8] , mammalian ATR arrests both processes [9 , 10] . Most late firing S . pombe origins are regulated by a checkpoint-independent mechanism [11] . Moreover , the contribution of error prone polymerases and homologous recombination to lesion bypass/repair varies considerably across the eukaryotic lineage . Arabidopsis thaliana elicits a unique response to double strand breaks , in which cell cycle arrest is averted . Instead , the ATR regulated , plant-specific transcription factor , SOG1 , induces genome-wide endoreplication [12] . In this study we explore the intra-S phase checkpoint response in Tetrahymena thermophila . As a representative of the Ciliophora lineage , Tetrahymena contains two nuclei within a single cytoplasm: the diploid germ line micronucleus and the polyploid ( 45 C ) somatic macronucleus [13] . The transcriptionally-silent micronucleus encodes genetic information that is transmitted from parent to progeny during the sexual phase of the life cycle . Its five chromosomes undergo conventional mitosis and meiosis . The actively transcribed macronucleus is polyploid , confers the cellular phenotype and the ~17 , 000 chromosomes are randomly segregate at each cell division . Since the parental macronucleus is destroyed in progeny , a new macronucleus must be generated . This process involves the extensive reorganization of germ line precursor chromosomes . The five micronuclear chromosomes are processed into ~180 unique macronuclear products that endoreplicate to a copy number of 45 C . The 21 kb ribosomal DNA ( rDNA ) minichromosome is further amplified to ~9000 C . Macronuclear chromosomes lack centromeres . While the intrinsic imprecision of amitosis can generate genic imbalances , both DNA content and chromosome copy number are maintained within a narrow range . ‘Excess’ DNA is eliminated in the form of chromatin extrusion bodies [14] and macronuclear chromosomes re-replicate when total DNA content falls below an acceptable minimum [15] . Although the fundamental components for replication initiation ( ORC , MCMs , Cdt1 ) and the DNA damage/replication stress checkpoint response ( ATR , Rad51 , gamma H2A . X ) are conserved in Tetrahymena , their roles in genome maintenance are only partially understood . We previously identified a caffeine-sensitive DNA damage/replication stress checkpoint pathway that is activated by hydroxyurea ( HU ) and methylmethanesulphonate ( MMS ) [16] . Inhibition of this pathway results in aberrant macronuclear division and micronuclear genome instability , consistent with a protective role for chromosomes . A single ATR gene was identified ( TTHERM_01008650 ) , however , no obvious ATM candidates were detected ( Tetrahymena Genome Database ( TGD ) , www . ciliate . org ) . ATR is required for the reorganization of chromosomes during meiosis as well [17] . Our recent analysis of an ORC1 knockdown strain provided unexpected insights into the regulation of MCM2-7 , a major target of ATR [18] . Like the S . cerevisiae orc2-1 mutant [19] , a moderate reduction in Tetrahymena Orc1p induces genome instability , but fails to trigger an intra-S phase checkpoint response [18] . In contrast to the budding yeast mutant , replication initiation is unaffected in Tetrahymena ORC1 knockdown strains . Instead , the rate of replication fork elongation decreases and Mcm6p levels are reduced , suggesting that the replicative helicase becomes rate limiting . Since the MCM complex is much more abundant than ORC in other eukaryotes [20] , this fork elongation defect was unanticipated . Moreover , while moderate reductions in pre-replicative complex ( pre-RC ) components are poorly tolerated during the Tetrahymena vegetative cell cycle , ORC and MCM levels oscillate to a much greater degree during development . Paradoxically , these pre-RC proteins are more abundant early in development when the DNA replication load is comparatively modest , and decline dramatically when the replication load increases [18] . At low ORC concentrations , the rDNA origin is frequently bypassed , raising the possibility that alternative initiation sites ( cryptic origins ) and/or unconventional mechanisms are called into play . In the work presented here , we document an unprecedented checkpoint response to HU treatment that begins with the targeted degradation of Tetrahymena ORC and MCM subunits . We provide evidence for alterations in macronuclear chromosomes that include global deacetylation of histone H3 , and manifest as an apparent time-dependent decrease in DNA content in S phase arrested cells . More remarkably , upon HU removal , DNA synthesis resumes and the entire genome is duplicated prior to replenishment of ORC and MCMs . The rDNA minichromosome is replicated , although its ORC-dependent origin is inactive . Non-rDNA chromosomes also initiate bidirectional DNA replication prior to ORC replenishment . These findings argue for the existence of an alternative mechanism for DNA replication that operates on a genome-wide scale when ORC and MCM proteins are rate limiting .
As previously reported , HU and MMS , arrest cell cycle progression and activate a DNA damage/replication stress checkpoint response in Tetrahymena , including induction of the double strand break repair protein , Rad51p ( Fig 1A , left panel ) [16] . Rad51 accumulation is inhibited by the addition of caffeine , consistent with the involvement of an apical PI3-like kinase . Since MMS activates both ATM and ATR in mammals , we used centrifugal elutriation to distinguish between G1/S ( ATM-like ) and intra-S phase ( ATR ) checkpoint responses . HU or MMS were added immediately to a homogeneous G1 phase population or 1 h later ( midway through S phase ) . Rad51p levels were elevated ~10-fold and 2-fold , respectively; for HU and MMS treated G1 phase cells relative to mock controls ( Fig 1A , right panel , 4 h treatment ) . To better monitor Rad51p induction , cells were examined at 1 h intervals . Whereas the HU-dependent signal increased continuously over 4 h , the Rad51p level quickly plateaued in MMS-treated cells ( Fig 1B , left panel ) . To address whether this difference is due to a block in progression from G1 to S phase in MMS-treated cells , elutriated cultures were grown to mid-S phase prior to drug addition ( Fig 1B , right panel ) . HU and MMS elicited strong and equivalent Rad51p responses in S phase treated cells . These data argue for the existence of distinct ATM-like ( G1 ) and ATR ( S phase ) checkpoint responses . Flow cytometry was used to examine the effect of HU and MMS on bulk DNA synthesis in G1 and S phase treated cells . As expected , DNA replication was rapidly inhibited by either drugs , regardless of the time of drug addition . Whereas the DNA profile of G1 arrested cells remained constant over the examined 4 h interval ( Fig 1C , left panel ) , the profile of S phase treated cells changed in an unanticipated way . An additional peak migrating at the G1 position appeared 1 h after HU addition and became more prominent over time ( Fig 1C , right panel , arrow ) . Two distinct sub-populations were generated , one arrested in S phase and the other displaying an apparent G1 DNA content . MMS treatment did not generate this bimodal distribution ( Fig 1C , right ) . Instead the S phase peak gradually shifted to the left , suggesting that DNA content within the entire population had decreased . Both results were highly reproducible ( n = 8 ) . These flow cytometry profiles were recapitulated in cells synchronized by starvation and re-feeding , with the following variation: bimodal peaks were occasionally observed in MMS-treated S phase cells ( S1B Fig , arrow ) . To investigate this phenomenon further , HU studies were expanded on elutriated S phase cultures . Mock-treated G1 , S and G2 phase populations ( Fig 1D , left panel ) were used as reference points to determine whether HU-treated S phase cells proceeded through G2 prior to generating the G1 profile . The data collected over 6 h are not consistent with this model ( Fig 1D ) . The DNA profile of S phase arrested cells did not shift to the right ( more DNA ) prior to formation of the G1 peak . In another experiment , we overlaid the 4 h HU profile with mock-treated G1 , S and G2 DNA reference points ( Fig 1E ) . Two features stand out . First , the right shoulder of the HU-treated population ( blue profile , smaller peak ) aligned precisely with the right shoulder of the S phase peak ( red profile; T = 0 h for HU addition ) , indicating that that DNA content did not increase in this HU-arrested subpopulation . Second , the left peak and shoulder of the HU-treated profile ( blue profile , larger peak , arrows ) aligned with the G1 profile ( green profile ) . No sub-G1 peak was observed , consistent with the idea that cells did not exit S phase and divide precociously . A subsequent flow cytometry analysis at 10 min intervals showed no evidence for a G2 peak ( S2 Fig ) . The collective data suggest that the DNA content of HU-treated Tetrahymena neither remains constant nor gradually increases , as has been reported for S . pombe [21] and S . cerevisiae [8] , respectively . Instead , the DNA content in the Tetrahymena macronucleus appears to decrease over time . Two mechanisms for decreasing macronuclear DNA content have been previously reported: asymmetric macronuclear division [22 , 23] and elimination of ‘excess’ DNA in the form of chromatin extrusion bodies ( CEBs ) [24] . Multiple criteria were used to examine these and other possibilities in HU or MMS-treated S phase cells . First , we monitored cell number . In contrast to mock-treated controls , cell density did not increase in HU or MMS treated cultures ( Fig 2A ) . Second , we visually assessed cell division . Unlike mock-treated controls , cytokinesis was not observed in HU-treated populations ( Fig 2B ) . Third , we quantified flow cytometry data over time . HU-treated S phase cells displayed a gradual , steady decline in apparent DNA content ( Fig 2C , red profile ) , while the DNA content of mock-treated cells oscillated in conjunction with cell division ( Fig 2C , blue profile ) . Fourth , we assessed cytoplasmic complexity/granularity . Consistent with the absence of cell division , the side scatter parameter increased substantially in the flow cytometry profile of HU-treated cells ( Fig 2D , red graph , and S3C Fig ) relative to mock-treated controls ( Fig 2D , blue graph , and S3A Fig ) . The increased side scatter in HU-treated cells exceeded values generated at any stage of the cell cycle in mock-treated controls , and was not dependent on the time of drug addition ( Figs 2E , S3B and S3C , compare HU addition to G1 or S phase cultures ) . Fifth , cell size and macronuclear size were quantified in DAPI-stained samples . Consistent with cell division arrest , HU-treated cells were on average 1 . 3X larger than mock-treated S phase controls ( Fig 2F graph , P <0 . 001; highly significant ) . Conversely , the size of the macronucleus appeared smaller ( Fig 2F , representative micrograph ) . To better assess the relationship between HU treatment and macronuclear size , mock-treated G1 phase cells were used as our reference point . While HU-treated mid-S phase cells were larger than G1 phase mock controls , their macronuclei were on average 20% smaller ( Fig 2G , P<0 . 1; significant ) ( nuclear/cytoplasmic ratio: 0 . 66X ) . Since , no sub-G1 DNA peak appears in HU-treated S phase cells ( Fig 1E ) , these data suggest that the global architecture of the macronucleus is somehow altered ( see below ) . Finally , DAPI was used to visually assess chromatin exclusion body ( CEB ) formation ( Fig 2H ) . The modest increase in CEBs in HU-treated cells was statistically insignificant , and cannot account for the large sub-population of cells with diminished ( G1 ) DNA content . The collective data demonstrate that HU and MMS-treated Tetrahymena do not undergo cell division or modulate their DNA content through general or ciliate-specific mechanism . Two additional mechanisms were considered to explain the apparent decrease in DNA content in S phase arrested cells: active degradation of DNA and global compaction of chromatin . DNA fiber analysis was used to examine DNA synthesis and determine the fate of pulse labeled DNA over time . As a starting point , we re-fed starved , G1-arrested cells with media containing HU for 4 h . HU-treated cells were simultaneously labeled with IdU for 1–4 h , and DNA fibers were examined for incorporation of IdU into nascent strands . Whereas cells in an asynchronous population generated long IdU tracts in a 20 min labeling ( Fig 3A , upper image , red tracts ) , no incorporation was detected in HU-treated cells , regardless of the timing or duration of labeling ( Fig 3A , middle ( 1 h IdU ) and lower ( 4 h IdU ) images ) . However , CldU was readily incorporated following HU removal ( green tracts ) . We conclude that HU arrests replication forks rather than slowing the rate of fork elongation . To examine the fate of nascent strands in HU-treated cells , synchronized mid-S phase cells were pulse labeled with IdU for 10 min prior to HU addition , and DNA fibers were prepared 1 and 4 h later . As a control , cells were collected directly after IdU removal without HU treatment . Nascent strands were readily detected and the density of labeled fibers was similar in mock and HU-treated cells . The length of labeled tracts was measured in over 350 randomly chosen DNA fibers per sample condition . The median length of labeled DNA tracks did not decrease following HU addition ( Fig 3B ) . We conclude that nascent DNA strands are not selectively targeted for degradation in HU-treated cells . While this experiment does not rule out the possibility that DNA is degraded non-selectively , the rapid resumption of DNA synthesis following HU removal ( S4 Fig ) , indicated that the overall integrity of chromosomes was maintained . Since DNA replication , DNA repair and chromatin compaction are influenced by epigenetic modifications , we next asked whether HU treatment affected two relevant post-translational histone modifications , phosphorylation of histone H2 . AX and acetylation of histone H3 . Phospho-gamma H2A . X accumulates at stalled replication forks in yeast and metazoa [1] , and this modification is a reliable marker for activation of the Tetrahymena ATR checkpoint response [25] . Hyper-phosphorylation of H2A . X is associated with chromatin condensation , DNA fragmentation and apoptosis in metazoan cells [26] . Phospho-gamma H2A . X levels were elevated ~10-fold in HU-treated Tetrahymena ( Fig 3C ) . Whereas , the ATR inhibitor , caffeine did not prevent H2A . X phosphorylation , the histone deacetylase ( HDAC ) inhibitor , sodium butyrate , lowered the gamma H2A . X induction ~2-fold ( Fig 3C , gamma H2A . X ) . Moreover , sodium butyrate inhibited the accumulation of Rad51p in HU-treated cells , and was more effective than caffeine ( Fig 4D ) . These results suggest that histone acetylation might have a protective function . To explore potential links between HU and histone acetylation in Tetrahymena , we directly examined histone H3 with an acetylation-specific monoclonal antibody . Histone H3 acetylation levels were reproducibly reduced ~2-fold in HU-treated Tetrahymena ( Figs 3C and 4C ) , and H3 hypo-acetylation was inhibited when sodium butyrate or caffeine were added at the time of HU addition ( Fig 3C ) . We conclude that the chromatin landscape is altered in HU-treated Tetrahymena . We next asked whether hypo-acetylation contributes to the apparent decrease in DNA content in HU-treated cells . Consistent with this idea , the G1 peak failed to appear when sodium butyrate was added to mid-S phase HU-treated cells ( Fig 3D ) . Instead , a very subtle , modest leftward shift in the mid-S phase DNA content peak was observed in HU + butyrate-treated cells . The cumulative data suggest that chromatin compaction contributes significantly to the altered ( G1-like ) DNA profile in HU-treated Tetrahymena . Since H3 hyper-acetylation promotes precocious activation of late firing and dormant replication origins in other species [27–29] , it is plausible that HU-induced hypo-acetylation could generate the opposite effect in Tetrahymena , inhibiting origin firing as part of a unique intra-S phase checkpoint response . In recently published work we showed that ORC and MCM protein levels change dramatically throughout development , and are co-regulated in vegetative growing cells [18] . To better understand the S phase checkpoint response , the fate of ORC and MCM proteins was examined in synchronized cell populations treated with HU or MMS . Western blot analysis revealed that the abundance of the Mcm6p remained relatively constant when MMS was added at the G1/S border or during mid-S phase ( Fig 1B , G1 and S phase drug addition; Fig 4B , lower panel ) . This observation is consistent with yeast and higher eukaryotes , where MCM activity is regulated by reversible phosphorylation [7 , 30] . HU-treated cells behaved very differently . Mcm6p levels rapidly declined ( Fig 1B , G1 and S phase HU addition; Fig 4B ) . More remarkably , Orc1p ( Figs 4C and 5C , HU arrest ) and Orc2p ( Fig 5C , HU arrest ) levels decreased in a time-dependent manner . Degradation of Orc1p was inhibited by the addition of caffeine to HU-treated cells ( Fig 4D ) , implicating the involvement of ATR . Rad51p was modestly induced in HU + caffeine treated cells , consistent with a role for ATR ( Fig 4D ) . We conclude that Tetrahymena replication stress ( HU ) and DNA damage ( MMS ) checkpoint responses differ in fundamental ways . Two questions immediately arose: are HU-treated cells competent to re-enter the cell cycle and faithfully replicate their chromosomes ? What events must occur for ORC and MCM-depleted cells to progress through S phase ? The precipitous drop in ORC and MCMs raised the possibility that origin licensing might be irreparably compromised in HU-treated cells . We first examined this possibility in cells that had transitioned past the G1/S border prior to cell cycle arrest . Synchronized S phase cells were exposed to HU for 4 h , and subsequently released into drug free media ( recovery period ) . Flow cytometry was used to monitor cell cycle progression , and western blotting assessed the abundance of Rad51p and Mcm6p . Remarkably , the recovering cell population quickly resumed DNA replication and completed S phase within 2 h ( S4A Fig ) , prior to the replenishment of Mcm6p ( S4B Fig , 0–2 h , no caffeine ) . These cells replicated , divided and enter a second S without delay , as Mcm6p levels were eventually restored ( S4B Fig , 4 h , no caffeine ) . DNA synthesis and cell cycle progression during the recovery phase did not require inhibition of the ATR pathway , although the first cell cycle proceeded more quickly in caffeine-treated cells ( S4A Fig ) . Whereas Rad51p levels were relatively constant during the recovery period , Rad51p levels declined and Mcm6p levels were replenished more quickly when caffeine was added ( S4B Fig ) . Moreover , suppression of ATR was not required for new DNA synthesis , in spite of the high levels of damage sustained during HU exposure ( Fig 3C , phospho-gamma H2A . X ) . We next addressed the requirements for ORC during the first S phase following HU removal . G1-arrested cells were used as the starting point for two reasons . First , since Orc1p levels naturally cycle within an unperturbed cell population ( Fig 5B , mock-treated control ) , peaking at G1 and disappearing by late S phase , we could effectively eliminate this variable by adding HU to G1 synchronized cells . Second , this regimen allowed us to follow an entire cell cycle after drug removal . Mock-treated starved/re-fed cells generated a characteristic DNA cell cycle profile ( Fig 5A ) . They entered S phase and down regulated Orc1p within 90 min , and synthesized new Orc1p during G2 phase ( 180 min ) ( Fig 5B ) . The flow cytometry profile is indicative of a healthy synchronous cell cycle . By comparison , DNA content increased within 60 min after HU removal and a new G1 phase population appeared prior to the replenishment of ORC ( Fig 5A and 5B , T = 120 min , T = 150 min ) . In a separate experiment , we probed for Orc2p . Orc1p and Orc2p were down regulated in HU-treated cells ( Fig 5C ) . Like Orc1p , Orc2p levels did not increase prior to the completion of the first S phase during the recovery from HU-induced replication stress . The unexpected behavior of pre-RC components raised the possibility that DNA replication might occur by an ORC-independent mechanism . The Tetrahymena rDNA minichromosome serves as an excellent substrate to examine the effect of depleting ORC subunits . As shown previously , wild type strains initiate replication exclusively from nucleosome-free regions in the 5’ non-transcribed spacer ( NTS ) in this palindromic 21 kb minichromosome ( Fig 6A , Domain 1 and Domain 2 ) [31] . ORC recruitment to rDNA origins is mediated by an integral RNA subunit , designated 26T RNA , that undergoes Watson-Crick base pairing with DNA sequences ( type I elements ) that are present at rDNA origins , but are absent from non-rDNA chromosomes [32 , 33] . Mutations in 26T RNA inhibit ORC recruitment to Domain 1 and Domain 2 origins [32] . Under these conditions , this origin is silenced , but the rDNA minichromosome is still replicated . Hence , our prior studies indicated that a backup program can be called into play when the normal mechanism for ORC-recruitment is compromised . Neutral/neutral ( N/N ) two-dimensional gel electrophoresis was used to assess rDNA origin activity in ORC-depleted Tetrahymena during the recovery from HU-induced replication stress . A characteristic bubble-to-Y arc RI pattern was observed in the HindIII 5’ NTS fragment of mock-treated S phase cells ( Fig 6B , S phase , probe 1 , arrowhead identifies the bubble arc ) . This pattern was not detected in ORC-depleted cells after release from the HU block . Instead , the predominant RI pattern that was generated corresponds to a simple Y arc ( Fig 6C , 0–120 min; filled arrow ) , indicative of passive replication of the 5’ NTS origins . A faint double Y arc pattern ( open arrow ) was also detected during the HU recovery phase , consistent with converging replication forks emanating from initiation events near the rRNA promoter or downstream of the 5’ NTS . Finally , an X spike was detected ( arrowhead ) , indicative of branch migration in stalled replication forks or DNA molecules generated by homologous recombination . We conclude that the rDNA origin is inactivated in ORC-depleted cells . To test for cryptic origins in the transcribed region , the 5 . 5 kb ClaI coding region fragment was examined ( Fig 6A , schematic ) . No bubble arcs were detected in this region in mock-treated controls or cells recovering from HU-induced replication stress . Instead , a simple Y arc was observed ( Fig 6D , probe 2 ) . To examine this question further , neutral/alkaline ( N/A ) gel electrophoresis was used to visualize nascent strands and determine the direction of replication fork progression through this coding region interval . N/A analysis can distinguish between four possibilities: random initiation , initiation from cryptic origins in the coding region , initiation from more distal sites ( 3’ NTS or the telomere ) , and initiation from the promoter-proximal region . Whereas both short and long nascent strands were detected with 5’ NTS proximal probe 3 , only high molecular weight RIs were detected with the more 3’ proximal probe 4 ( Fig 6E ) . The collective data suggest that replication initiates near the rRNA promoter during the HU recovery phase . Further localization of the initiation site could not be achieved due to physical limitations of these methods and the absence of informative restriction sites to bracket the rRNA promoter . DNA fiber analysis was used to expand our analysis of DNA replication to the rest of the macronuclear genome during the HU-recovery phase . Most importantly , this method allowed us to distinguish between two models for genome-wide replication during the recovery phase: elongation of stalled forks generated prior to HU removal versus new initiation events . Cells were synchronized in G1 by starvation and re-feeding , and HU was immediately added to the media . HU was removed 4 h later and nascent DNA strands were sequentially labeled with IdU and CldU . New initiation events will generate DNA fibers in which the red signal ( IdU pulse ) is flanked by green ( CldU , chase ) on both sides . Elongation of existing forks will produce a green-red-gap-red-green pattern , in which the gapped region was replicated prior to the pulse/chase ( HU recovery ) period . Both patterns were detected following HU removal , prior to replenishment of ORC and MCMs ( Fig 7A ) . New initiation events predominated , comprising ~80% of randomly chosen images . We conclude that bidirectional replication initiation is the predominant mechanism for genome-wide duplication of macronuclear chromosomes during the recovery from HU-induced replication stress . DNA fibers were also used to quantitatively examine replication initiation and fork elongation in mock-treated controls ( high ORC and MCMs ) and during the first S phase after HU removal ( low ORC and MCMs ) . The physical distance between replication initiation sites ( inter-origin distance , IOD ) and the speed of elongating replication forks ( fork velocity , FV ) were assessed [34] . In previously published work , we reported a median inter-origin distance of 24 . 3 kb and fork elongation rate of 0 . 83 kb/min for mid-log phase vegetative cells . An experimentally-induced 5-fold reduction in Orc1p and corresponding drop in Mcm6p had no effect on inter-origin distance , but resulted in a decreased fork elongation rate [18] . We performed a similar analysis on cells recovering from HU-arrest , prior to the replenishment of Orc1p , Orc2p and Mcm6p . Whereas the fork elongation rates for mock and HU-recovering cells were indistinguishable ( Fig 7B ) , the median inter-origin distance increased by 16% during the first S phase following HU removal ( from 26 . 0 kb to 30 . 1 kb; p<0 . 01 ) . Hence , although the conventional replication initiation program is perturbed ( Fig 6 ) , new initiation events are still the major contributor to genome-wide replication under low ORC conditions .
A distinguishing feature of ciliated protozoa is the presence of two functionally distinct nuclei within a single cytoplasm . During vegetative growth , the diploid , germ line micronucleus and polyploid , somatic macronucleus replicate at different stages of the cell cycle . Their respective chromosomes segregate at different times and by unrelated mechanisms . The mitotic micronucleus divides during macronuclear G1 and amitotic macronuclear division is coupled to cytokinesis . Since the macronucleus confers the cellular phenotype , macronuclear DNA damage or replication stress would be expected to decrease fitness . However , aberrant macronuclear division with lagging chromosomes , indicative of incomplete DNA replication , is readily tolerated [22 , 33] . The high copy number of macronuclear chromosomes ( 45 C ) may serve as a buffer against errors in DNA replication and chromosome transmission . Indeed , the macronucleus self corrects for genic balances by re-replicating chromosomes [36 , 37] or jettisoning excess DNA in the form of chromatin extrusion bodies [24] . Within the context of these seemingly chaotic cell cycles , Tetrahymena employs an ATR checkpoint kinase to arrest DNA replication in response to damage or replication stress [16] . ATR is also required for micronuclear genome stability and plays a critical role in meiotic chromosome transmission [17] . In this study , we exploited our ability to isolate a pure unperturbed population of G1 phase cells to explore the consequences of inducing DNA damage or replication stress prior to and during S phase . As expected , the administration of MMS or HU midway through macronuclear S phase led to rapid cell cycle arrest ( Fig 1C ) , induction of the break-induced repair protein , Rad51p ( Fig 1A and 1B ) , and phosphorylation of gamma H2A . X ( Fig 3C ) . However , the parallels with yeast and metazoa diverge here . The first unanticipated feature of the Tetrahymena checkpoint response emerged from flow cytometry analyses of cells treated with HU at mid-S phase . In contrast to other eukaryotes where DNA content remains constant or gradually increases during prolonged HU exposure [8 , 11] , the apparent DNA content decreased in HU-arrested Tetrahymena , assuming a value indistinguishable from G1 phase cells ( Figs 1C–1E and S1B ) . DNA fiber analysis indicated that macronuclear DNA was not actively degraded ( Fig 3B ) and microscopic analysis ruled out DNA elimination through CEBs ( Fig 2H ) . Formation of the ‘G1 peak’ was blocked by the HDAC inhibitor , sodium butyrate ( Fig 3D ) , suggesting that chromatin was reorganized in HU-arrested cells . Consistent with this idea , histone H3 acetylation was reduced in HU-treated Tetrahymena ( Figs 3C and 4C ) , and the macronucleus was more compact than mock-treated G1 phase cells ( Fig 2G ) . The collective data suggest that this apparent decrease in DNA content is a reflection of changes in the epigenetic organization of the macronucleus . By analogy , studies in diploid maize cells have demonstrated that chromatin compaction can diminish propidium iodide fluorescence intensity [38] . Weakened interactions between chromosomes and the nuclear envelope promote chromatin compaction in cultured human cells , manifesting as aggregated chromatin clusters [39] . While we cannot distinguish between cause and effect at this time , our data suggest that the altered properties of macronuclear chromosomes described above are closely associated with epigenetic modifications on a genome-wide scale . The second distinguishing feature of the Tetrahymena checkpoint response emerged from western blot analysis of HU and MMS-treated cells . A cornerstone of the intra-S phase checkpoint response in yeast and higher eukaryotes is the reversible phosphorylation-dependent inhibition of the replicative helicase , through ATR/CHK1-dependent phosphorylation of Mcm2p and Mcm3p [2 , 40 , 41] . In stark contrast , a core component of the replicative helicase , Mcm6p , is degraded in HU-arrested Tetrahymena ( Fig 4B ) . Moreover , Orc1p and Orc2p proteins are degraded . The decline in these proteins is not reflective of the natural progression through S phase . First , these proteins are degraded when HU is added prior to S phase entry . Flow cytometry and DNA fiber analysis corroborate the arrest of replication initiation and elongation in HU-treated Tetrahymena ( Figs 3–5 ) . Second , while Orc2p levels do not oscillate across the cell cycle [33] , this protein is degraded in HU-treated cells ( Fig 5C ) . Finally , although MMS and HU elicited intra-S phase checkpoint responses , MMS does not trigger ORC and MCM degradation ( Fig 4B and 4C ) , irrespective of the time of addition . The third and most unexpected behavior of the Tetrahymena checkpoint response was revealed through our analysis of DNA replication after HU removal: an entire S phase occurred without replenishment of ORC and MCM proteins ( Figs 5 and S4 ) . The sole origin in the rDNA minichromosome is silenced ( Figs 6 and 7 ) and no cryptic origins were uncovered in the 21 kb rDNA minichromosome . However , leading strand synthesis proceeds toward the telomeres , suggesting that replication initiates at a new site proximal to the rRNA promoter . New initiation events are commonplace in non-rDNA chromosomes ( Fig 7 ) and the inter-origin distance increases slightly , indicating that stalled forks play a minor role in the replication of chromosomes during the recovery phase . Replication stress has been shown to trigger different , species-specific checkpoint responses and recovery mechanisms . For example , while HU inhibits replication initiation in S . cerevisiae , it only slows replication forks in S . pombe [8] . Furthermore , most late firing origins in S . pombe are regulated by a RAD3 ( ATR ) independent mechanism [11] . Although mammalian ATR inhibits replication initiation and elongation [9 , 10] , replication origins are far from equivalent , since neighboring origins are differentially sensitive to the depletion of dNTP pools [42] . The unifying features in these species are the stabilization of late firing origins and reversible phosphorylation of MCM subunits [2–4] . Once the source of stress is removed , the conventional DNA replication program resumes [5 , 43] . We propose that the DNA replication properties of HU-treated Tetrahymena are a manifestation of DNA replication programs that normally operate during macronuclear development . Recent studies revealed that the abundance of ORC and MCM proteins are dynamically co-regulated [18] . Partial depletion of Orc1p by gene disruption leads to a concomitant decrease in Orc2p and Mcm6p levels during the vegetative stage of the life cycle . Moreover , ORC and MCM naturally fluctuate to a greater degree during conjugation , and poorly correlate with the demands for DNA replication: ORC and MCM are elevated early in development when the replication load is low , and precipitously decline when the replication load increases [18] . Aberrantly migrating ( RNase A + mung bean nuclease-sensitive ) RIs are generated in the endoreplicating macronucleus , suggesting that RNA-DNA hybrids accumulate , similar to yeast senataxin mutants [18 , 44] . Hence , the initiation of DNA replication is dynamically re-programmed during Tetrahymena development . Our analysis of cells recovering from HU-induced replication stress suggests that the capacity to replicate under low ORC and MCM conditions is retained during the vegetative phase of the life cycle . Several models are considered for DNA replication with limiting amounts of ORC . In the first model , replication initiation and elongation are uncoupled in HU-treated Tetrahymena . Stalled replication forks generated during HU arrest are simply elongated during the recovery phase . Whereas initiation and elongation are transiently uncoupled in Drosophila follicle cells [45] , two pieces of data disfavor this model in Tetrahymena . First , during the recovery from HU-induced stress , new initiation events were detected by 2D gel analysis in rDNA minichromosome; however , the 5’ NTS origin did not serve as the initiation site ( Fig 6 ) . More importantly , the predominant CldU/IdU/CldU pattern in non-rDNA DNA fibers generated during HU recovery corresponds to new initiation events rather than the elongation of stalled replication forks ( Fig 7A ) . In the second model , a subpopulation of origins is refractory to ORC and MCM depletion . Initiation events during HU recovery could occur at cryptic or late-firing origins , analogous to origins that are used by S . cerevisiae upon HU removal [46] This model predicts that these origins would function at low ORC concentrations , and consequently should have a high affinity for ORC [47] . Tetrahymena’s intrinsic ability to partially re-replicate the macronuclear genome , despite S phase-specific degradation of Orc1p , is consistent with this model [33 , 36] . A confounding factor is the global decrease in histone acetylation in HU-treated Tetrahymena . Based on studies in yeast [27] , we would predict that this epigenetic state would disfavor ORC-dependent initiation events . Further studies are needed to determine the relative contribution of chromatin-associated and pre-deposition histones to the H3 acetylation profile in HU-arrested Tetrahymena [48] . A third model posits that the genome is duplicated in an ORC-independent manner . Archaebacteria and animal viruses provide the only precedents . Haloferax volcanii strains lacking all four ORC-dependent chromosomal origins exhibit increased fitness relative to strains with 1–4 intact origins [49] . Cell viability in the originless mutant is dependent on RadA recombinase , which is non-essential in wild type strains . In this view , the converging replication forks and branch migration intermediates in the Tetrahymena rDNA 5’ NTS ( Fig 5B ) could represent recombination substrates or intermediates that prime DNA replication . Based on the measured inter-origin distance ( Fig 7 ) , approximately 250 , 000 recombination events would be required to replicate a single polyploid macronucleus . Alternatively , DNA replication could be primed by a covalently bound protein , as in adenovirus [50] , tRNA molecules ( avian sarcoma virus ) [51] , or non-coding RNA . While there is no precedent for these scenarios on genome-wide scale , RNA polymerase I generated RNA-DNA hybrids ( R loops ) have recently been shown to prime unscheduled replication initiation in the rDNA of S cerervisiae [52] . Stable RNA-DNA hybrids have been detected in endoreplicating rDNA molecules , when Tetrahymena ORC concentrations are low [18] . Another possibility is that small non-coding RNAs could be involved . Dicer-generated 27–30 nt RNAs , direct the removal of ~6000 IES elements in the developing Tetrahymena macronucleus [53 , 54] . Small ( 23–24 nt ) non-coding RNAs of unknown function are produced during the vegetative cell cycle [55] . The 287 nt long non-coding RNA , 26T RNA , is an integral component of ORC , selectively directing the complex to complementary DNA sequences at the amplified rDNA origin [32] . The underlying mechanism for replication initiation in ORC-depleted Tetrahymena awaits further studies . The Tetrahymena system has several interesting unanticipated parallels in which endoreplication is a common theme . DSB-inducing agents can trigger an ATR-dependent endoreplication program in Arabidopsis thaliana rather than simply arrest the cell cycle [12] . While Drosophila ORC is normally present in endoreplicating nurse cells [56] , ORC2-deficient mutant cells undergo multiple rounds of endoreplication by an undetermined mechanism [57] . The archaebacteria H . volvanii , which efficiently propagates originless chromosomes is polyploid , and can tolerate significant variation in chromosome copy number [49] . We propose that polyploid cells are inherently more tolerant than diploid cells to DNA damage and replication stress , and can duplicate their chromosomes by unconventional mechanisms that might also function to a lesser degree in haploid/diploid tissues and/or species . High-throughput mapping of replication origins should provide novel insights into underlying mechanism ( s ) for replication initiation site selection in Tetrahymena .
Tetrahymena thermophila strain CU428 was obtained from the Tetrahymena Stock Center . Standard methods were used to cultivate cells and visualize cells by light and fluorescence microscopy [16] . Cell cycle synchronization was achieved by starvation and re-feeding or by centrifugal elutriation as previously described [33] . For DNA damage checkpoint activation , standard growth media was supplemented with either 20 mM hydroxyurea ( HU , Sigma Chemical ) or 0 . 06% methylmethansulphonate ( MMS , Sigma Chemical ) . Stock solutions of caffeine and sodium butyrate ( Sigma-Aldrich ) were prepared in water . The working concentrations of caffeine and sodium butyrate were 1 mM and 50 mM , respectively . For cell division analysis , DAPI-stained cells were examined microscopically at defined intervals after drug treatment and removal . Flow cytometry analysis was performed on Becton Dickinson FACSAria II flow cytometer ( BD Biosciences ) , using propidium iodide to monitor DNA content . Data were analyzed using BD FACSDiva software . Histograms were plotted , in which the y-axis represents the number of events and the x-axis represents the relative DNA content . Forward and side scatter data were graphed to visualize the distribution of individual cells within the population . Whole cell extracts were prepared by lysing cells in 1% SDS lysis buffer ( 50 mM Tris at pH 8 . 0 , 150 mM NaCl , 1 mM EDTA , 1% NP-40 , 1% sodium deoxycholate , 1% SDS ) for 15 min on ice . Protein concentrations were determined using the modified Lowry protein assay reagents ( Bio-Rad ) . Unless otherwise stated , equal amounts of total protein ( 20 μg ) were loaded in each lane and samples were subjected to SDS-PAGE and transferred to nitrocellulose membranes ( Whatman Protran BA85 , GE Healthcare ) . Prior to probing , membranes were stained with Ponceau S staining ( Sigma-Aldrich ) to confirm equivalent sample loading and protein transfer . Immunodetection of Tetrahymena Orc1p ( 1:5 , 000 ) , Orc2 ( 1:5 , 000 ) and Mcm6p ( 1:10 , 000 ) were carried out using polyclonal rabbit antibodies raised against immunogenic Tetrahymena peptides ( Covance ) . Antibodies directed against Rad51 ( 51RAD01 , Thermo Scientific , 1:5 , 000 dilution ) , phospho-gamma H2A . X ( 2F3 , BioLegend , 1:1 , 000 dilution ) , and acetyl-histone H3 ( polyclonal antibody #06599 , Millipore , 1:10 , 000 dilution ) were obtained from the indicated commercial sources ) . Blots were incubated with the primary antibodies at 4°C overnight , washed and incubated for 3 h at 4°C with horseradish peroxidase-conjugated goat anti-rabbit or goat anti-mouse IgG ( Jackson ImmunoResearch ) . Membrane bound secondary antibodies were visualized using ECL reagents ( PerkinElmer ) according to the manufacturer’s instructions . Densitometry of blots was performed using ImageJ software ( version 1 . 47v for Macintosh , National Institutes of Health ) . Total genomic DNA was isolated from Tetrahymena cultures as previously described to preserve DNA replication intermediates ( RI ) [58] . For RI enrichment , 200 μg of genomic DNA were digested with HindIII ( rDNA 5’ NTS analysis ) or ClaI ( rDNA coding region analysis ) for 4 h and applied to the 200-μl packed volume benzoylated naphthoylated DEAE ( BND ) -cellulose ( Sigma-Aldrich ) . Caffeine-eluted DNA samples were precipitated with isopropanol , using 20 μg glycogen as a carrier . Total DNA recovery was estimated to be ~5% of input . Neutral-neutral ( N/N ) and neutral-alkaline ( N/A ) 2D gel electrophoresis were performed as previously described [31] . Approximately 3–10 μg of BND cellulose-enriched DNA was loaded for each 2D gel experiment . Nor N/N analysis , the first dimension gel ( 0 . 4% agarose ) was run in 1X TAE buffer ( 40 mM Tris , 20 mM acetic acid , and 1 mM EDTA ) at 1 . 5 V/cm for 20 h at RT . The second dimension gel ( 1% agarose ) was run in 1X TBE buffer ( 90 mM Tris , 90 mM boric acid , 2 mM EDTA ) containing 0 . 5 mg/ml ethidium bromide at 3 V/cm for 18 h at 4°C . For N/A analysis , the second dimension separation was performed at 4°C in recirculating buffer ( 40 mM NaOH , 2 mM EDTA ) at 1 . 5 V/cm for 20 h . DNA was transferred overnight to a charged nylon membrane ( Hybond-XL , Amersham ) in alkaline buffer by capillary blotting . Membranes were prehybridized at 37°C for 4 h in 1M NaCl , 1% SDS , 10% dextran sulfate , 5 mM Tris [pH 7 . 5] , 100 μg/ml of denatured salmon sperm DNA and 25% formamide . rDNA 5’ NTS and coding region probes were labeled by random priming and added directly to the prehybridization solution . After 18 h , membranes were washed three times in 2X SSC/1% SDS solution for 15 min each at 42°C , and once in 0 . 4X SSC/0 . 1% SDS solution for 15 min at 42°C . Blot were exposed to X-ray film with an intensifying screen at -70°C or analyzed with a phosphorimager . Tetrahymena was cultured in 2% PPYS media to the density of 1 . 5x10e5 cells/ml . For pulse chase experiments , cells were sequentially pulse labeled with 400 μM IdU ( Sigma ) and 100 μM CldU ( MP Biomedicals ) at 30°C . The media was removed and cells were washed once with 1X PBS between each labeling step . For DNA fiber processing , cells were washed twice with PBS , the cell density was adjusted to 1x10e6 cells/ml . Preparation and immunostaining of DNA fibers was performed as described previously described [59] with the following modifications . Briefly , after fixation and HCl treatment , slides were washed three times with 1X PBS , and blocked with 5% BSA in PBS for 30 min . Mouse anti-BrdU ( 1:50 , Becton Dickson ) and rat anti-BrdU ( Accurate Chemical , 1:100 dilution ) antibodies in 5% BSA were then added onto slides . After 1 h incubation , the slides were washed three times again with 1X PBS , and incubated for 30 min with secondary antibodies: Alexa Fluor 568 goat anti-mouse IgG ( Invitrogen/Molecular Probes , 1:100 dilution ) and Alexa Fluor 488 goat anti-rat IgG ( Invitrogen/Molecular Probes , 1:100 dilution ) . Finally , slides were washed three times with 1X PBS , dehydrated with an ethanol series , and mounted with SlowFade Gold antifade solution ( Invitrogen ) . During immunostaining , all antibodies were diluted in 5% BSA in 1X PBS , all incubations were performed at 37°C , and all wash steps were done at RT . DNA fiber images were obtained on a Nikon A1R+ confocal microscope at 600X magnification . Measurements of track length were performed with Nikon NIS-Elements software . Green-red-green tracks represent origins that initiated DNA replication during the IdU pulse . Green-red-gap-red-green tracts represent origins that initiated DNA replication prior to the IdU pulse . Inter-origin distance was defined as the distance between the centers of two red segments in either green-red-green-red-green or green-red-gap-red-green tracks . Fork velocity was determined by measuring the length of the green segment in red-green tracks or the red segments in green-red-gap-red-green tracks . GraphPad Prism software was used to analyze the statistical significance , and the p-values shown in figures were determined by two-tailed unpaired t-test . | DNA damage and replication stress activate cell cycle checkpoint responses that protect the integrity of eukaryotic chromosomes . A well-conserved response involves the reversible phosphorylation of the replicative helicase , MCM2-7 , which together with the origin recognition complex ( ORC ) dictates when and where replication initiates in chromosomes . The central role of ORC and MCMs in DNA replication is illustrated by the fact that small changes in abundance of these pre-replicative complex ( pre-RC ) components are poorly tolerated from yeast to humans . Here we describe an unprecedented replication stress checkpoint response in the early branching eukaryote , Tetrahymena thermophila , that is triggered by the depletion of dNTP pools with hydroxyurea ( HU ) . Instead of transiently phosphorylating MCM subunits , ORC and MCM proteins are physically degraded in HU-treated Tetrahymena . Unexpectedly , upon HU removal the genome is completely and effortlessly replicated prior to replenishment of ORC and MCM components . Using DNA fiber imaging and 2D gel electrophoresis , we show that ORC-dependent mechanisms are bypassed during the recovery phase to produce bidirectional replication forks throughout the genome . Our findings suggest that Tetrahymena enlists an alternative mechanism for replication initiation , and that the underlying process can operate on a genome-wide scale . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | Checkpoint Activation of an Unconventional DNA Replication Program in Tetrahymena |
In this paper we used a general stochastic processes framework to derive from first principles the incidence rate function that characterizes epidemic models . We investigate a particular case , the Liu-Hethcote-van den Driessche's ( LHD ) incidence rate function , which results from modeling the number of successful transmission encounters as a pure birth process . This derivation also takes into account heterogeneity in the population with regard to the per individual transmission probability . We adjusted a deterministic SIRS model with both the classical and the LHD incidence rate functions to time series of the number of children infected with syncytial respiratory virus in Banjul , Gambia and Turku , Finland . We also adjusted a deterministic SEIR model with both incidence rate functions to the famous measles data sets from the UK cities of London and Birmingham . Two lines of evidence supported our conclusion that the model with the LHD incidence rate may very well be a better description of the seasonal epidemic processes studied here . First , our model was repeatedly selected as best according to two different information criteria and two different likelihood formulations . The second line of evidence is qualitative in nature: contrary to what the SIRS model with classical incidence rate predicts , the solution of the deterministic SIRS model with LHD incidence rate will reach either the disease free equilibrium or the endemic equilibrium depending on the initial conditions . These findings along with computer intensive simulations of the models' Poincaré map with environmental stochasticity contributed to attain a clear separation of the roles of the environmental forcing and the mechanics of the disease transmission in shaping seasonal epidemics dynamics .
A plethora of deterministic epidemic models involving susceptible , infected and recovered individuals have been proposed [1] , [2] , carefully analyzed [3]–[8] and confronted with data sets in the biomathematics and ecology literatures [9]–[12] . A well defined topic within this mathematical ecology research area is the study of -type models with seasonal forcing [13]–[16] . These models have proved to be useful for understanding the observed patterns and the natural processes behind human and non-human epidemics [17]–[21] . Here , we restrict our attention to the and models in which we introduce seasonal forcing while varying the structural form of the incidence rates . Two hypotheses pertaining the RSV and the measles transmission mechanisms were modeled with two simple functional forms of the incidence rates . We show that in doing so , we are able to attain a clear separation of the roles of the environmental forcing and the mechanics of the disease transmission in shaping the epidemics dynamics . The construction of deterministic incidence rates functions is a critical building block of epidemiological modeling . In a seminal paper , Hethcote [1] showed that because there are many choices for the form of the incidence , demographic structure and the epidemiological-demographic interactions , there really is a plethora of incidence rate functional forms to choose from . Not surprisingly , the biomathematics literature abound in qualitative mathematical analyses of many of these functional forms [22]–[26] . However , biological first principles derivations of incidence rate functional forms are not too common . As we show in this study , using such first principles derivations greatly enrich the reaches of the practice of confronting models with data while testing biological hypotheses . Thus , despite the big amount of available functional incidence rates forms [1] , we believe that the set of models chosen to be confronted with data should be restricted to those forms derivable from first principles . To illustrate this argument , in this study we first show that a simple probabilistic setting wherein infectious encounters are modeled with a pure birth stochastic process leads to a general nonlinear incidence form proposed previously by Liu [24] and later analyzed by Hethcote and Van Den Driessche [23] ( hereafter we refer to the Liu , Hethcote and Van Den Driessche incidence rate as the LHD incidence rate ) . The LHD incidence rate leads to models with qualitatively different dynamics compared with the ones obtained using the classical incidence rate . In the SIRS model with either incidence rate and seasonal forcing , becomes a periodic function of time and the trajectory “pursuits” a moving target thus giving rise to limit cycles . That moving target is the former endemic equilibrium that bounces back and forth between two points . In either model , the target switches between that moving point and the disease free equilibrium when crosses 1 , giving rise to a period doubling bifurcation . In the SIRS model with classical incidence rate this mechanism does not depend on the initial conditions . In this work we show that the disease free equilibrium ( DFE ) is unconditionally an attractor in the SIRS model with LHD incidence rate . This leads to a scenario where two regions of attraction can coexist . The trajectory will either reach the disease free equilibrium or have periodic solutions depending on the initial conditions . Furthermore , after carrying a formal model selection we show that the SIRS model with LHD incidence rate leads to a significant fit improvement over the classical SIRS model with the same seasonal forcing . Finally , we compared the applicability and generality of the classical and LHD incidence rates functions by fitting them to two measles time series data sets . Using the later function leads to a vast improvement of model fit in both cases . Since we were fitting a deterministic SEIR model , we chose to use the data from the two largest cities in the measles data set ( London and Birmingham , see http://www . zoo . cam . ac . uk/zoostaff/grenfell/measles . htm ) , where the effects of demographic stochasticity are expected to be less influential in the dynamics of the epidemics [10] . Varying the form of the contact rate function while including environmental stochasticity in the SIRS and SEIR models leads to a better understanding of the dynamics of an infectious disease transmission . Depending on the model and contact rate , the disease free equilibrium ( DFE ) is either a saddle point or an attractor . In the first case , if a trajectory located originally in the basin of attraction of the endemic equilibrium ( EE ) basin of attraction is perturbed with environmental noise , it may transiently visit the DFE stable submanifold and then return to the EE basin of attraction . If however the DFE and the EE coexist as stable equilibria , a trajectory initially at the EE basin of attraction may end up in the DFE basin of attraction . The interaction between stochasticity and the different contact rate models was studied using computer intensive simulations of the Poincaré map [27] .
The classical model has been extensively studied in order to predict and understand various disease dynamics behaviors , as well as their spread and persistence [28] . For many infectious diseases , the pool of susceptible individuals is replenished due to the waning of immunity [17] , [18] . To account for the lost of immunity , the classical susceptible , infected and recovered model is adjusted by allowing a fraction of the recovered individuals to move back into the susceptible pool at a rate . This susceptible , infected , recovered and susceptible model is expressed as ( 1 ) ( 2 ) ( 3 ) where is the rate of loss of infectiousness and the total population size remains constant ( i . e . ) . The constant represents both , the birth and mortality rates . Assuming that birth and mortality rates are equal is justified on the grounds that the annual infection rate is considerably higher than the population growth . The constant is the contact rate , the average number of individuals with whom one infected individual makes sufficient contact to pass on the infection [29] . The fraction represents the average number of infections per susceptible individual and hence represents the expected number of infections when susceptible individuals are available [5] . Note that the above definition of as a per individual constant leads to a consistency of the units within each of the model equations and assumes homogeneous mixing . In the following sections we will discuss different ways to model the incidence rate . The equations for the classic SEIR ( Susceptible-Exposed-Infectious-Recovered ) model are as follows [30]: ( 4 ) ( 5 ) ( 6 ) ( 7 ) where represents both , the birth and mortality rates per capita . The mean latent and infectious periods of the disease are and . As written , the SEIR model has a stable endemic equilibrium provided . Further biological realism to model recurrent epidemics can be incorporated to both this SEIR model and the SIRS model above by assuming that the transmission rate varies seasonally . Indeed , Earn et al [30] study the range of the dynamical behavior of the SEIR model with seasonality and find it useful for explaining the measles numerous transitions between regular cycles and irregular , possibly chaotic epidemics . Also , Alonso et al . [31] show that noise amplification provides a possible explanation for qualitative changes from regular to irregular oscillations of lower amplitude . In this paper , we follow the suggestion made by Hethcote [1] and couple Liu , Hethcote and Van Den Driessche's incidence rate with seasonal forcing in both the SIRS and SEIR models . To incorporate the claim that epidemics of recurrent infections is driven by seasonality , it is customary to depart from the standard incidence rate by assuming that the average number of incidences sufficient for transmission per infected individual , is a periodic or quasi-periodic function of time ( ) . Often , the incidence rate is assumed to have a sinusoidal form of the type ( 8 ) where stands for the strength of the seasonality and year . Various authors have shown that such a generic description of the seasonal variation in transmission rates is not as revealing as a detailed description of the actual processes underlying the seasonal drivers of transmission through mechanistic seasonal forcing functions [11] , [18] , [30] , [32] , [33] . However , as we show in the results section , in some cases this sinusoidal function may unequivocally represent a linear transformation of a weather covariate . Although other authors have used a more flexible Haar step function for the seasonal forcing ( e . g . [30] ) , we restrict ourselves to the incorporation of the sinusoidal form above ( eq . 8 ) as the seasonal forcing . This has the advantage of ease of interpretation and qualitative analysis . In any case , the main purpose of incorporating the forcing is to explore the main qualitative characteristics of coupling the seasonally varying disease transmission and different incidence rate functional forms . Brauer [34] generalizes the incidence rate definition in the following way: if the average member of the population makes contacts in one unit of time with , and if is the probability of choosing one infected individual from the population at random , then is the rate of new infections per unit of time . The mass-action incidence rate model is recovered using and the classic incidence rate is recovered by picking . A general incidence rate function was proposed by Hethcote and van den Driessche [23]:where and are constants . Consider the special case where and . Using Brauer's generalization and idea , Hethcote and van den Driessche's model is recovered using the function . Then , the incidence rate function becomeswhere and . Although the mathematical properties of the general function are known in general [23] , [35] , [36] a mechanistic , first principles derivation of it is still lacking . Such a derivation can be obtained using a probabilistic reasoning analogous to the argument used by [37] to model the Allee effect through stochastic mating encounters: Through physical movement or any other means of dispersion , an infected individual will have contact with a given number of susceptible individuals in the population . The potential to effectively disperse the disease ( virus ) could be thought of as being proportional to that number of susceptibles with whom the infected individual makes contact: indeed , the more contact the infected individual has with susceptibles , the more likely he is to effectively transmit the disease . It then follows that the magnitude of the realized disease dispersion could be measured for example , in terms of the dispersion ability ( i . e . vagility ) of the infected individual . Accordingly , every infected individual will be expected to realize a certain virus ( or micro-parasite ) dispersion potential . Let the realized disease dispersion made by one infected individual be denoted by . Then , the number of successful transmission encounters per infectious individual can be modeled with a random variable . By writing , we are stressing the fact that the infection process is a function of the magnitude of the realized dispersion . Furthermore , we assume that the probability that an infected individual encounters and infects a susceptible individual given a realized change in dispersion is proportional to the previous number of successful infection encounters times a function of the number ( or density ) of the infected individuals in the population . Often [7] , a non-linear function is chosen to account for factors such as crowding of infected individuals , multiple pathways to infection , stage of infection and its severity or protective measures taken by susceptible individuals . These assumptions allow us to specify a new infection event as the conditional probability ( 9 ) where is a non-negative function such that is a constant . Towards the end of this section we discuss possible functional forms for . We remark that if counts the number of successful transmission encounters of an infected individual that recently invaded a population consisting only of susceptible individuals , then the expected value of is in fact equal to the mean number of secondary infections in the context of the SIRS model . If the SEIR model dynamics is in place , then , when there is only one infected individual in the population , . Assuming that the probability that more than one successful infectious encounter occurs after an extra dispersion amount is negligible , then can be modeled using a simple homogeneous birth process where the quantity being born is the number of successful virus transmission encounters . The probabilistic law of this stochastic process is completely defined by the terms . To solve for these terms , first note that according to eq . ( 9 ) which leads toIn the limit when , the above equation leads in turn to the following system of differential equations:Then , it is well known [38] that solving this system of equations leads to ( 10 ) ( 11 ) Furthermore , approximating using a Taylor series expansion around leads to specific quantitative definitions of the stochastic process . For example , if or if , the one-step transition probability mass function ( pmf ) of adopts the negative binomial and Poisson forms respectively [37] . The Negative Binomial transition pmf would bring into the picture over-dispersion ( higher variance to mean ratio ) as a key qualitative property of the moments of the pure birth process describing the evolution of the number of successful transmission encounters . In any case however , the probability that one infected individual successfully passes on the infection isThis expression is readily interpretable: for a fixed value of , the probability of successfully passing on the infection converges to as the product grows large . Therefore , in this expression we are recovering the model property that the probability of successfully passing on the infection increases with the realized disease dispersion effort . Each individual's realized dispersion is in turn related to the individual's ‘effort’ to transmit the infection . In a given population , the magnitude of the realized disease dispersion for each infected individual can be expected to vary widely . Accounting for this demographic source of heterogeneity could be achieved by assuming that each individual's dispersion ability is drawn from a given probability distribution . That is , we would be modeling the variation in disease dispersion per infected individual with a random variable whose pdf has support on . Without loss of generality , here we model randomness in the product instead of just in the realized disease dispersion . Then , the probability that an infected individual chosen at random from the population realizes more than one successful secondary infection is found by averaging over all the possible realizations of . That is , A suitable probabilistic model for with empirical and theoretical support can be difficult to find ( see for instance the models in [39] ) . A flexible positive , continuous distribution such as the gamma distribution could therefore be used . Here , we assume that the magnitude of the disease dispersion brought about by an infected individual is distributed according to a special case of the gamma pdf , the exponential distribution . Accordingly , letting we get that the probability of successfully transmitting the infection is ( 12 ) As mentioned before , various biological hypotheses pertaining the behavior of the transmission as a function of the abundance of infected individuals have been advanced to justify various functional forms of . Suitable candidates for should satisfy the conditionsThese conditions guarantee the basic requirement that the probability of a new infective encounter ( eq . 9 ) is null in the absence of infected individuals and that the overall chance that a new infection occurs increases proportionally with when is small . Furthermore , if such proportionality decreases in magnitude as grows large ( that is , is concave down ) . Consider the following two functional forms: Many other functional forms for the incidence rate could be derived using the above arguments . If for instance other heavy-tailed distributions are used instead of the exponential distribution , other incidence rate functional forms will arise and this could certainly be the topic of further research . However , in this work we limit ourselves to the exploration of the reaches of using the LHD model because it explicitly incorporates heterogeneity in transmission potential , because of its bi-stability properties ( see “qualitative analysis of the SIRS models” section ) and to formally test if it arises as a better explanation for bi-annual epidemic patterns using data from different localities and diseases . Thus , from this point on , in this work we will only consider the LHD incidence rate function and the classical incidence rate . In his seminal paper , Hethcote [1] also mentions that the LHD general incidence rate function could be eventually coupled with any seasonal forcing function . Motivated by this comment , in the results section we explore the reaches of doing so .
The two different SIRS models were fitted to time-course data of reported cases of syncytial virus infections . The data come from Gambia and Finland ( Figure 1 ) . Two ML formulations were used . The first one consisted of a Poisson likelihood that only required the available observed counts of infected individuals ( eq . 13 ) . The second formulation consisted of the joint likelihood of the counts and of the observed weather covariate and thus used information present on the time series of reported cases and on the corresponding time series of mean monthly temperature range for both locations . The ML estimates according to the first formulation for each model and data set combination are displayed in Table 1 . Both information criteria used indicate that for Finland , the best model was the SIRS model with LHD incidence rate function . For Gambia , both information criteria for the SIRS model with classic incidence rate function are lower by three points approximately . This implies that given the data and the two information criteria ways of penalizing the likelihood score , both models are nearly indistinguishable for any practical purpose [49] . In Gambia , the extra parameter introduced by the LHD model is penalized: given the data set at hand , incorporating one extra parameter does not lead to a clear improvement In Figure 2 we plotted the model predicted number of infected individuals versus the observed values for the classical and the LHD SIRS model respectively . Note that , even though the best model is deterministic , the dynamics displayed by the data ( small epidemics followed by a big epidemic peak ) is very well recapitulated by the predicted solutions . The results of the second ML formulation are qualitatively identical to the results with the Poisson likelihood ( see table in the Text S1 ) . For Finland , the BIC statistic for the classical model was 10376 . 2000 and for the LHD model 9893 . 5780 . For Gambia , the BIC for the classical model was 729 . 1133 whereas the LHD model had a BIC of 733 . 2750 . Hence , here again , for Finland the LHD is the best model whereas for Gambia , the classic model is better . Because the BIC can be used only to compare models for which the numerical values of the dependent variable are identical for all estimates being compared , it cannot be used to select between the two ML formulations . Indeed , in the second likelihood formulation the data fitted consist not only of the time series of infected counts but also of the monthly temperature range , thus it uses twice as much data for parameter estimation . Zeng et al [55] mention that an indication of which likelihood formulation is better can be obtained by comparing the per datum BIC score . Take for instance the BIC for the LHD model for Finland , 9893 . 5780 . Dividing that BIC by the total number of data points used ( , we get a per datum BIC of 48 . 4979 . Now , the BIC for the LHD model for the Poisson likelihood formulation is ( Table 1 ) 3196 . 9330 . Dividing that number by the number of data points used ( ) we get 31 . 34248 . Thus , the Poisson likelihood formulation yields a better per datum BIC for Finland . For Gambia , the Poisson likelihood formulation seems to be better than the Poisson-Normal sampling model: for the classic model with Poisson likelihood this statistic is , whereas for the classic model with Poisson-Normal likelihood it is . The SEIR model with classic and LHD incidence rate were fitted to measles time series data from London and Birmingham . In both cities , the SEIR-LHD model was selected as best ( see Table 2 ) . Notably , the difference in AIC and BIC is at least 2000 points in each case . The predictions for each model and city combination are shown in Figure 3 . We remark that assessing and comparing the quality of the model predictions visually may be misleading . Indeed , according to our likelihood formulation , the parameter estimation process does not weight equally a deviation from the model prediction at low and high infected counts . In fact , the variance of the Poisson sampling error varies according to the mean predictions . In this section we discuss the differences in the qualitative behavior of the SIRS model ( 1 ) – ( 3 ) with both classical and LHD incidence rates with and without seasonal forcing . We refer the interested reader to the Text S1 for proofs of the following claims . By construction , the set is a positively invariant set of the SIRS model ( 1 ) – ( 3 ) . If we set the coefficients constant , the Dulac criterion guarantees that the SIRS model with neither the classic nor the LHD incidence rate function has periodic solutions in . Regarding the classical incidence rate , the SIRS model has two stationary solutions: a disease free equilibrium ( ) and an endemic equilibrium ( ) . It is well known that is a threshold for this model: If the disease remains endemic , while implies that the disease dies out . On the other hand , the SIRS model with LHD incidence rate has one disease free equilibrium and two endemic equilibria and . The is unconditionally a local attractor . However , only one of the endemic equilibria denoted as , lies inside the positively invariant set . If the endemic point is locally an attractor . Thus , when the LHD model exhibits bi-stability . Introducing seasonal forcing has the following effects on the SIRS dynamics with classic incidence: first , it is well known that by letting the contact rate to be a periodic function of the form ( 8 ) where is small , the SIRS model with classical incidence rate has a periodic solution with period . This behavior is shown in Figure 4 A and B . Also , when seasonal forcing is introduced , the basic reproductive number becomes a periodic function of time , , that oscillates between the values and . The endemic point also becomes a periodic function of time that bounces back and forth between two extreme points , and . The expressions for and are given in the Text S1 . The associated limit cycle of the model's solution inherits the stability behavior of the endemic point: if , then the limit cycle is asymptotically stable . A stable limit cycle is displayed in Figure 4 C . Because the function can cross the boundary of periodically depending on the value of , the dynamic behavior of the model's trajectory with respect to the nature of the endemic point ( stable/unstable ) can be described with a race analogy: The model's solution can be thought of as a hopeless ‘pursuer’ engaged in a race against the endemic solution who plays the role of the fast ‘leader’ that cannot be caught upon . Just as in a cycling race , as soon as the leader changes its strategy , so does the pursuer behind the leader . In that way , if is such that and only while , the leader ( ) is deemed as stable and the solution's trajectory pursues the endemic point . As soon as becomes less than , the leader ‘changes its strategy’ and is deemed unstable whereas the becomes stable . At that moment , the trajectory switches its objective and pursues the and keeps doing so while . That sudden change of objective gives rise to a period doubling bifurcation of the limit cycle as seen in Figure 4 D . This change of objective ( period doubling bifurcation ) happens as grows large . We remark that at least one route to chaos in the associated Poincaré map of this model when is taken as the bifurcation parameter has been shown [53] , [56] , [57] . Finally , in the SIRS model with LHD incidence rate ( see Figure 5 A and B ) , if we let the contact rate to be a periodic function of the form ( 8 ) , a limit cycle also arises ( see Figure 5 C ) . Here again , as increases , the trajectory engages in the same pursuer/leader dynamics and the limit cycle undergoes a period doubling bifurcation ( Figure 5 D ) . However , contrary to what happens in the classical SIRS model with seasonal forcing , periodicity or extinction of the epidemics depends also on the initial conditions: if the initial proportion of infected individuals is too high , the disease will die from a subsequent depletion of the susceptible pool of individuals . Only if the epidemic begins with a small number of individuals will it slowly work its way up and attain a persisting limit cycle .
Multiple lines of evidence show that the forced SIRS and SEIR models with LHD incidence rate function constitute a better explanation of the seasonal epidemic patterns than the corresponding classical models with seasonal forcing , for the data sets and cases explored here . The first line of evidence is statistical in nature: when confronted with different time series of seasonal epidemics , the LHD model was selected as best in three out of four cases and in the fourth case , the LHD model was nearly indistinguishable from the classic model . By formulating the fitting and the model selection problems using likelihood-based inference and information theoretic model selection criteria we were able to conclude that given the data and the models at hand our model embodies the most likely explanation of how the observed data arose . Our model's nonlinear incidence rate takes into account heterogeneity in the ability to transmit the infection while modeling the infectious process as a pure birth stochastic process and hence , it is a more realistic model formulation . This new level of model complexity was achieved by incorporating only one extra parameter . The emphasis we give to a first principles derivation that hinges on interpretability and simplicity is not always sought in other SIR-type model formulations and modeling exercises [6] , [17] , [18] , [24] , [58] . Hence , our results show that a careful exploration of other incidence rate functions before resorting to mathematically more complex , high-dimensional models may bring new insights into the current understanding of the functioning of epidemics . Another line of evidence in favor of the LHD model comes from its qualitative predictions . The classical SIRS model without the seasonal forcing predicts somewhat artificially that regardless of the initial proportion of infected and susceptible individuals , provided , the endemic equilibrium will be reached [28] . On the other hand , the LHD model without seasonal forcing predicts that the disease-free equilibrium is always an attractor , thus exhibiting bi-stability ( see qualitative analysis section ) . Hence , if the initial proportion of infected individuals is too high , the disease will die from a subsequent depletion of the susceptible pool of individuals , contrary to what the classical model predicts . For the disease to persist in the population , the initial proportion of infected individuals has to be very low . Only then the infection process will proceed steadily to the endemic solution . This qualitative prediction matches the virus transmission strategy that the syncytial virus seems to have evolved: recall that in our model the extra parameter is the density of infected individuals at which the probability of successfully transmitting the infection is . In every locality , the ML estimates of were in the order of to , thus indicating that a very low density of infected individuals is needed in order to maximize the transmission rate of the measles and RSV diseases . Incorporating weather covariates to our nonlinear SIRS model further improves the biological insights that can be concluded from the parameter estimation and model analysis exercises . A simple look at the strong auto-covariation patterns and at the pure weather trends , in particular for Gambia ( Figure 1 ) indicate that modeling weather and weather effects with a sinusoidal function seems a natural add-on to the classic SIRS model , for this data set . For Gambia , the fact that the per datum BIC for the LHD model with the joint Poisson-Normal likelihood is very similar to the per datum BIC for the classic model indicates that the weather can indeed be viewed as a simple rotation and translation ( eq . 15 ) of the weather effects ( eq . 8 ) . Thus eq . 15 may not always be viewed only as a phenomenological artifact [18] . For Finland , however , this was not the case . The per datum BIC favors much more clearly the Poisson likelihood formulation . Hence , we consider that in Finland the weather effects model ( eq . 8 ) would be better expressed as some unknown nonlinear transformation of the weather . In other words , in this country with more extreme weather , a change in the temperature range of a certain size is not translated as an equivalent change in the weather effects in the transmission rate . Also embedded within our weather effects model formulation ( eq . 8 ) is the hypothesis that weather affects incidence rates in a nonlinear fashion . In particular , when the strength of seasonality is high enough , the limit cycles predicted by both weather forced models undergo a period doubling bifurcation such that relatively small epidemic outbreaks are followed by big ones . Notably , these effects of the strength of seasonality were detected in Finland , the locality where the amplitude of the relative weather oscillation is larger . The model selection exercise should by no means be the ending point of the analysis . Instead , if appropriateness of one model vs . the other cannot be resolved , a near-tie in a model selection situation should lead to the search and reformulation of each model's scientific predictions in a way that can be clearly tested in further experiments . Hence , the model selection results presented here should be rather viewed as the starting point of further analyses ( see [59] ) . Even for simple deterministic models , parameter estimation for dynamic data can be non-trivial . Dynamic models often present multimodal likelihoods thus complicating the parameter estimation process [42] . In these cases , the type of inferences possible is limited due to the presence of wide confidence sets that include parameter values with different qualitative predictions . If for instance the ML estimate of a bifurcation parameter is in a 2 limit-cycles region but its confidence interval includes parameter values for which these cycles do not appear , then there is not enough evidence in the data at hand to properly infer something about the size of the parameter of interest and hence , about the dynamic properties displayed by the data . In our case however , the precision of our parameter estimates and in particular , of the bifurcating parameter ( Tables 1 and 2 ) is enough to identify the bifurcation region where the strength of seasonality lies for the data at hand . Although in the two models studied here a period doubling bifurcation appears in the limit cycle , the LHD incidence rate model still provides very different qualitative predictions . In the classical model , the value of the basic reproduction number as a function of time acts as a stability switch for the DFE , so that any trajectory that begins with biologically realistic initial conditions will eventually enter the limit cycle . This is not the case for the LHD model , for which the periodicity or extinction of the epidemics depends very naturally on the initial conditions . Other studies have incorporated seasonal forcing in SIRS-type models [17] , [60] , [61] , but since all have used the classical incidence rate function , they constrain their disease persistence and epidemics predictions to whether the basic reproductive number can or cannot be periodically above 1 . Nonlinear incidence rate forms derived from first principles constitute a promising starting point to review the interaction between demographic and environmental stochasticity and nonlinear seasonal effects . Indeed , recent studies have considered including in the classic SIRS model stochasticity in the seasonal process , besides sampling and/or observation error [17] . After showing that a simple pure observation error fit of our LHD model brings about a considerable fit improvement , we explored the qualitative differences between the models by coupling the deterministic skeletons with environmental noise . In Figure 6 , the depicted stochastic trajectories show that in the classical model increasing the environmental noise results in transient visits to the disease free equilibrium stable submanifold ( panel c ) ) , whereas in the LHD model , with a large enough perturbation the trajectory visits the disease free equilibrium basin of attraction and remains there . Hence , the fact that regardless of the value of the basic reproduction number the DFE is always an attractor opens the door to stochastic phenomena whereby the trajectory exits the endemic solution basin of attraction and hits just by chance the DFE basin of attraction , only when the LHD incidence rate is used . By the same token , the trajectory periodically wanders in the direction of the DFE stable submanifold ( similar to the “saddle fly-by” reported by Cushing et al [54] ) . The results presented here are not by any means an exhaustive exploration of the interplay between nonlinear dynamics and stochasticity , both critical factors shaping seasonal epidemic patterns . However , our results may be viewed as the starting point of multiple research avenues . Three such research topics could be: first-principles derivation of non-linear incidence rate functions , the role of bi-stability and demographic stochasticity for disease persistence and the simulation of environmental and demographic stochasticity in the Poincaré map . | Nonlinearity in the infection incidence is one of the main components that shape seasonal epidemics . Here , we revisit classical incidence and propose a first principles derivation of the infection incidence rate . A qualitative analysis of the SIRS model with both the classical and the proposed incidence rate showed that the new model is physically more meaningful . We conducted a statistical analysis confronting the SIRS and SEIR models formulated using both incidence rate functions with four data sets of seasonal childhood epidemics . Two data sets were hospital records of cases of syncytial respiratory virus ( RSV ) . The other two data sets were taken from the well-known UK measles epidemics database . We found that seasonal epidemics is better explained using our incidence rate model embedded in a Poisson sampling process . The results presented here are not by any means an exhaustive exploration of the interplay between nonlinear dynamics and stochasticity . Our results may be viewed as the starting point of multiple research avenues . Three such research topics could be: the first-principles derivation of non-linear incidence rate functions , the role of bistability and demographic stochasticity for disease persistence and the simulation of environmental and demographic stochasticity in the Poincaré map . | [
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion"
]
| [
"ecology/theoretical",
"ecology",
"computational",
"biology/population",
"genetics",
"genetics",
"and",
"genomics/microbial",
"evolution",
"and",
"genomics",
"mathematics/statistics"
]
| 2011 | First Principles Modeling of Nonlinear Incidence Rates in Seasonal Epidemics |
Human papillomaviruses infect stratified epithelia and link their productive life cycle to the differentiation state of the host cell . Productive viral replication or amplification is restricted to highly differentiated suprabasal cells and is dependent on the activation of the ATM DNA damage pathway . The ATM pathway has three arms that can act independently of one another . One arm is centered on p53 , another on CHK2 and a third on SMC1/NBS1 proteins . A role for CHK2 in HPV genome amplification has been demonstrated but it was unclear what other factors provided important activities . The cohesin protein , SMC1 , is necessary for sister chromatid association prior to mitosis . In addition the phosphorylated form of SMC1 plays a critical role together with NBS1 in the ATM DNA damage response . In normal cells , SMC1 becomes phosphorylated in response to radiation , however , in HPV positive cells our studies demonstrate that it is constitutively activated . Furthermore , pSMC1 is found localized in distinct nuclear foci in complexes with γ-H2AX , and CHK2 and bound to HPV DNA . Importantly , knockdown of SMC1 blocks differentiation-dependent genome amplification . pSMC1 forms complexes with the insulator transcription factor CTCF and our studies show that these factors bind to conserved sequence motifs in the L2 late region of HPV 31 . Similar motifs are found in most HPV types . Knockdown of CTCF with shRNAs blocks genome amplification and mutation of the CTCF binding motifs in the L2 open reading frame inhibits stable maintenance of viral episomes in undifferentiated cells as well as amplification of genomes upon differentiation . These findings suggest a model in which SMC1 factors are constitutively activated in HPV positive cells and recruited to viral genomes through complex formation with CTCF to facilitate genome amplification . Our findings identify both SMC1 and CTCF as critical regulators of the differentiation-dependent life cycle of high-risk human papillomaviruses .
Human papillomaviruses are the causative agents of cervical and other anogenital malignancies . HPV-16 , 18 , 31 along with at least ten other types are referred to as high-risk as they are associated with the development of genital cancers[1] [2] . These high-risk viruses infect squamous epithelial cells in the genital tract and link their productive life cycles to differentiation . HPVs infect cells in the basal layers of stratified epithelia and establish their genomes as nuclear episomes at about 50 to 100 copies per cell [3] . In infected basal cells , viral genomes are replicated along with cellular DNA and distributed equally to the two daughter cells[4] . While one daughter cell remains in the basal layer , the other migrates away and undergoes differentiation in suprabasal layers . HPVs do not encode their own polymerases and rely on cellular enzymes to replicate their genomes . Normally , daughter cells that migrate from the basal layer exit the cell cycle , however , in HPV infections these cells remain active in the cell cycle and re-enter S/G2 in suprabasal layers to productively replicate their genomes in a process called amplification[5] [6] . Amplification is coincident with activation of the late viral promoter and synthesis of capsid virions [4] [7] [8] . Two processes that regulate the differentiation-dependent viral late phase include maintenance of cell cycle competence in suprabasal cells and activation of the ATM DNA damage response . In normal cells , the Ataxia-Telangiectasia Mutated ( ATM ) DNA damage pathway mediates the repair of double strand breaks . A trimeric MRN complex consisting of MRE11 , RAD50 , and NBS1 first recognizes double strand breaks induced by irradiation or chemicals[9] . The MRN complex recruits ATM proteins to sites of double strand breaks resulting in its activation through autophosphorylation [10] . The ATM kinase then activates a series of downstream effectors that can be grouped into three pathways: p53/p21 , CHK2/CDC25 and NBS1/SMC1 [10] [11] [12] . Phosphorylation of these proteins results in cell cycle checkpoint arrest in G1 , G2 , or S depending on which pathway is utilized . Induction of p53 phosphorylation leads to G1/S arrest while activation of CHK2/CDC25 as well as NBS1/SMC1 leads to G2/S arrest [11] [13] [14] . ATM also induces the phosphorylation of a specific type of histone , called γ -H2AX , that is recruited to regions that are tens of kilobases in size that flank sites of double strand breaks [15] [16] . While activation of the ATM pathway is important for HPV genome amplification , it is not clear which set of the ATM factors provide necessary functions . SMC1 proteins are members of the Structural Maintenance of Chromosomes ( SMC ) family of proteins that play critical roles in organizing and stabilizing chromosomal segregation during mitosis . SMC1 forms a heterodimeric cohesin complex with SMC-3 that encircles and mediates sister chromatid cohesion for proper chromosome segregation[17] . Two other members of this family , SMC2 and SMC4 , associate to form the condensin complex that regulates chromosome condensation[18] . In addition to its role in chromosome segregation , SMC1 can be phosphorylated by ATM and plays a central role in mediating G2/M checkpoint arrest as well as homologous recombination repair [12] [14] . As shown by Yazdi et al . SMC1 is part of the NBS1 arm of the ATM pathway , which is distinct and acts independently from the CHK2/CDC25 pathway[12] . Previous studies have shown that ATM and CHK2 are important regulators of HPV genome amplification[5] [19] [20] [21] . In the present study , we investigated if SMC1 is important for HPV genome amplification in differentiating cells . Our studies demonstrate that HPV proteins induce the constitutive activation and phosphorylation of SMC1 ( S957 ) in both undifferentiated and differentiated cells . In HPV positive cells , active , phosphorylated SMC1 proteins localize to foci or puncta in the nucleus in complexes with other members of the ATM pathway . We further demonstrate that SMC1 is recruited to HPV genomes through association with cellular CTCF insulator proteins . Importantly , inhibition of SMC1 or CTCF activity through shRNA knockdowns impairs differentiation-dependent genome amplification . This study identifies SMC1 and CTCF as critical components of the ATM pathway that are important for differentiation-dependent genome amplification .
To investigate if SMC1 structural maintenance proteins play any role in the life cycle of HPV , we first examined if the levels of total and phosphorylated SMC1 are altered between normal and HPV positive keratinocytes . Western blot analysis demonstrated that levels of total , and phosphorylated SMC1 , are increased by approximately two fold in HPV positive cells as compared to normal keratinocytes ( Fig 1A ) . Similar activation of SMC1 is also seen in HPV-18 positive cell lines . Consistent with previous observations , the levels of active pCHK2 are also significantly increased in HPV positive cells ( Fig 1A ) [19] . It was next important to determine if epithelial differentiation alters the levels of total or phosphorylated SMC1 proteins . For this analysis we used calcium induced differentiation of epithelial cells as this allowed for the isolation of uniformly differentiated populations of cells . HPV amplification begins about 48 hours after addition of calcium and plateaus at approximately 72 to 96 hours[19] . Our studies show that the levels of both total SMC1 and pSMC1 in HPV positive cells remain elevated at similar levels throughout differentiation in contrast to normal keratinocytes ( Fig 1B ) . HPV E7 oncoproteins are able to induce pSMC1 in the absence of complete replicating viral genomes ( S1 Fig ) . Following DNA damage , a subset of SMC1 proteins become phosphorylated and associate with regions adjacent to double stranded breaks[22] , [23] , [24] , [25] . We next investigated if HPV alters the localization of SMC1 proteins . For this analysis , we examined HPV positive CIN 612 cells as well as normal keratinocytes by immunofluoresence to observe localization of SMC1 and pSMC1 . In normal keratinocytes , total SMC1 proteins are distributed uniformly throughout the nucleus but few pSMC1 foci are observed ( Fig 2A ) . In contrast , in HPV positive cells , high levels of pSMC1 positive foci are observed that are localized to the nucleus ( Fig 2A ) . Consistent with previous observations , pCHK2 is found only in HPV positive cells and is localized to foci in the nucleus [20] . Upon differentiation of HPV positive cells , an increase in the number and size of nuclear foci containing pSMC1 is seen ( Fig 3B ) while no such foci were seen in normal keratinocytes . Specifically the number of nuclei with 3 or more pSMC1 foci increases significantly upon differentiation to approximately 80% of cells ( Fig 2C ) . We next investigated if ATM or another kinase was responsible for the phosphorylation of SMC1 . For this analysis we treated cells with the specific ATM kinase inhibitor , KU60019 , and screened for levels of pSMC1 as well as pCHK2 , a known target of the ATM kinase , by Western blot analysis[26] . Treatment with KU60019 was found to completely abolish phosphorylation of CHK2 , however , it only partially inhibited phosphorylation of SMC1 ( Fig 3A ) . This indicates that phosphorylation of SMC1 is partially dependent upon ATM kinase activity as well as an additional kinase which is most likely ATR . We next treated cells with the DNA damage inhibitor , caffeine , and found that it significantly reduces levels of pSMC1 but not pCHK2 ( Fig 3B ) . We conclude that both ATM and ATR kinases target SMC1 in HPV positive cells and is consistent with previous reports examining normal cells following DNA damage[12] , [14] , [27] . In HPV positive cells , members of the ATM pathway such as pCHK2 and γ-H2AX are recruited to nuclear foci that also contain viral genomes[20] . To determine if pSMC1 is co-localized to ATM-positive foci , dual immunofluoresence assays were performed on HPV positive cells . We observed both pCHK2 and γ-H2AX to co-localize to the same foci as pSMC1 in differentiated HPV positive cells ( Fig 4 ) . It was next important to determine if pCHK2 and γ-H2AX formed complexes with pSMC1 or whether they merely localized to similar regions in the nucleus . For this analysis we used proximity ligation methodologies , which is also referred to as on-slide co-immunoprecipitation , as this assay allowed for a demonstration of association as well as the identification of where these complexes are located in cells[28] . The proximity ligation assay involves the use of secondary antibodies that are linked to oligonucleotides with overlapping regions of complementarity . If the proteins are in a complex or close proximity , the oligonucleotides can hybridize and act as primers for polymerase mediating amplification with incorporation of fluorescent nucleotides . The appearance of red signal in immunofluorescence analysis is indicative of complex formation . We performed this analysis using HFKs along with HPV positive cells and found complex formation between γ−H2AX and pSMC1 as well as pCHK2 and pSMC1 in nuclear foci only in HPV positive cells ( Fig 5 ) . In normal keratinocytes exposed to the same set of antibodies , only background signals were observed demonstrating that these interactions are specific to HPV positive cells . We conclude that pCHK2 , γ-H2AX and pSMC1 associate into complexes at nuclear foci in HPV positive cells . Since our studies indicated that SMC1 proteins form complexes with CHK2 , and γ-H2AX at HPV replication centers , it was important to investigate if SMC1 was necessary for HPV amplification . For this analysis we used lentiviral vectors that expressed shRNAs that were effective in partially knocking down levels of SMC1 . SMC1 is an essential cellular protein and complete knockdowns are lethal[11] . In our studies , SMC1 levels in HPV 31 positive cells were reduced by approximately 50% through the use of lentivirus shRNA vectors ( Fig 6A ) . HPV positive cells were infected with the recombinant lentiviruses and the reduction in SMC1 levels was confirmed by Western analysis . The cells in which SMC1 had been reduced with shRNAs continued to proliferate at a rate comparable to control cells and exhibited similar percent viability ( S2 Fig ) . The ability to amplify viral genomes upon differentiation was next examined by Southern blot analysis and indicated that reduction of SMC1 levels by approximately 50% is sufficient to inhibit HPV genome amplification ( Fig 6B ) . Similar results were observed in four independent experiments using two different shRNA vectors against SMC1 as well as the combination of the two . Interestingly , reduction of SMC1 levels did not alter the ability of γ-H2AX to form foci in HPV positive cells . We conclude that SMC1 is an important regulator of HPV genome amplification . As part of the cohesin complex , SMC1 proteins associate broadly with cellular chromosomes to mediate sister chromatid cohesion . SMC1 can also form complexes with CTCF DNA binding proteins to bind to specific sequences [29] , [30] , [31] . CTCF is a zinc-finger protein that is required for transcriptional insulation as well as DNA looping and chromatin organization [32] . The association of CTCF with SMC1 has been shown to be independent of SMC1’s role in sister chromatid cohesion[33] , [34] . Similarly , SMC1 binding to CTCF is not necessary for transcriptional insulation and not all CTCF sites are occupied by SMC1 [29] . We investigated if pSMC1 and CTCF localize to the same nuclear foci in HPV positive cells through co-immunofluorescence analyses . As shown in Fig 7A , pSMC1 and CTCF co-localize to distinct foci in the nuclei of differentiated HPV positive cells that also contain CHK2 and γ-H2AX . Similar CHK2 and γ-H2AX foci in HPV positive cells have been described previously [19 , 20] . Numerous CTCF foci are detected in the nuclei of HPV positive cells but only a subset co-localize with pSMC1 . In contrast all pSMC1 foci co-localize with CTCF ( Fig 7A ) . Importantly , no pSMC1/CTCF foci were detected in normal keratinocytes demonstrating that such complexes are absent in HFKs . We next investigated if pSMC1 and CTCF formed complexes in these cells through the use of proximity ligation assays . As seen in Fig 7B , complexes are present in HPV positive cells but no signal above background was observed in normal keratinocytes . CTCF binds consensus motifs ( CCCTC ) and we identified similar sequences in the L2 and L1 open reading frames of HPV 31 as well as a potential non-consensus motif in the E2 ORF ( Fig 8A ) . Examination of other HPV types identified similar sequence motifs in the L2 ORFs of over 85% of types while the rest contain CTCF motifs in E2 , E4 or L1 ( Fig 8B ) . Additional CTCF sites may also be present that do not consist of canonical CCCTC motifs . In order to identify which regions of the HPV genome were bound by SMC1 proteins , a series of chromatin immunoprecipitation analyses ( ChIP ) were performed using HPV positive cells . In this analysis we screened for binding to sequences around the origin or replication in the HPV 31 URR as well as sequences around the canonical CTCF binding sites in the L2 open reading frame . Little or no binding of SMC1 was detected at the URR sequences in either undifferentiated or differentiated cells ( Fig 8C and 8D ) . In contrast , significant binding was detected to the L2 region in undifferentiated cells and similar levels were found upon differentiation ( Fig 8E and 8F ) . In addition , we observed low level binding of γ-H2AX to both URR and L2 regions in undifferentiated cells and this increased significantly at both regions upon differentiation ( Fig 8C , 8D , 8E and 8F ) . This is consistent with the reported broad distribution of γ-H2AX binding around sites of DNA damage and is in agreement with our previous report of increased binding of γ-H2AX to viral sequences upon differentiation [16] . An analysis of SMC1 and γ-H2AX binding to HPV sequences as a percentage of input is shown in S3 Fig and is consistent with our analyses using fold binding over IgG ( Fig 8 ) . We conclude that SMC1 preferentially binds to the late region of HPV genomes and this is consistent with a role for CTCF in recruiting SMC1 to viral DNAs . It was next important to examine if CTCF played a role in genome amplification . The levels of CTCF were first examined by Western blot analysis of undifferentiated and differentiated HPV positive cells and compared to those seen in normal HFKs . While CTCF levels were observed to decline in HFKs upon differentiation , they were retained at higher levels in differentiated HPV positive cells ( Fig 9A ) . To determine if CTCF played a role in genome amplification , HPV positive CIN 612 cells were infected with lentiviruses expressing shRNAs against CTCF and screened for levels of viral episomes following differentiation by Southern blot analysis . Since CTCF is an essential gene , we were only able to transiently reduce CTCF levels by approximately 50% as measured by Western blot analysis ( Fig 9B ) . No significant change in the levels of pSMC1 was detected upon reduction in CTCF levels ( S4 Fig ) , In addition , reduced binding of SMC1 to viral genomes was detected by ChIP analysis following CTCF knockdown ( S5 Fig ) Importantly , we observed a corresponding inhibition of genome amplification when CTCF levels were reduced . Similar results were seen in three independent experiments using two different shRNA vectors as well as the combination of both . CTCF can provide a number of functions including transcriptional regulation and so we performed Northern analysis of HPV transcripts upon CTCF knockdown and found a reduction in levels ( S6 Fig ) indicating that CTCF acts in multiple ways to regulate the HPV life cycle . Interestingly , reduction of CTCF levels did not alter the ability of γ-H2AX to form foci in HPV positive cells ( S7 Fig ) . These studies implicate CTCF as an important factor regulating genome amplification . To confirm that CTCF/SMC1 binding was critical to genome amplification , we mutated the three canonical CTCF binding sites in L2 in the context of complete HPV genomes so as to retain the L2 coding sequence intact . The mutant genomes were then sequenced to insure no other mutations were present in the viral DNAs . The mutant viral DNAs as well as wild type genomes were excised from bacterial sequences , religated and transfected into normal human foreskin keratinocytes along with a drug selectable marker . Drug resistant colonies were selected and expanded . In multiple independent transfections we observed that colonies containing mutant genomes grew more slowly than wild type transfected cells ( Fig 10A ) . The colonies were expanded and the state of viral DNA was examined by Southern blot analysis . Cells with wildtype genomes exhibited high levels of episomal DNAs that were stably maintained and amplified upon differentiation . In contrast , cells containing CTCF mutant genomes contained very low levels of episomes that were rapidly lost with passage . The cells with mutant genomes contained primarily integrated copies and failed to amplify upon differentiation ( Fig 10B ) . This suggests that the CTCF binding sites in L2 play a central role in the HPV life cycle including an unexpected potential role in stable maintenance of episomes . Finally , it was important to confirm that the mutations introduced into HPV 31 genomes abrogated the binding of CTCF and SMC1 proteins . ChIP analysis was performed on the wild type and mutant genome containing cells and no binding of CTCF or SMC1 to the L2 region was detected at both an early passage following transfection as well as in subsequent passages ( Fig 10C and S3 Fig ) . In contrast , both factors were found to bind to L2 in cells containing wild type viral genomes . Interestingly , pSMC1 still formed foci in cells with integrated HPV genomes though they appear localized to only one or two foci ( S7 Fig ) . These studies demonstrate that CTCF and SMC1 binding to the L2 region is critical for genome amplification as well as a identifying a novel role in stable genome maintenance .
Our studies demonstrate that the cohesin protein , SMC1 , is a critical factor in regulating HPV genome amplification in differentiating cells . SMC1 is a member of the cohesin complex that mediates sister chromatid cohesion , which is necessary for proper chromosomal segregation in mitosis [35] . The phosphorylated form of SMC1 also functions in the ATM DNA damage repair pathway and facilitates homologous recombination repair [24] , [23] . Our studies indicate that HPV genomes increase the levels of total and phosphorylated SMC1 in the absence of irradiation or other DNA damaging agents . This contrasts with what occurs in normal cells where exposure to irradiation or other damaging agents is required for induction of SMC1 phosphorylation[24] . The constitutive activation of SMC1 phosphorylation in HPV positive cells suggests it has an important role in the viral life cycle . The activation of the ATM DNA damage pathway is necessary for HPV genome amplification upon differentiation but it has little effect on transient or stable maintenance replication in undifferentiated cells [19] . Previous studies indicated that HPV genome amplification is dependent on the activities of the CHK2 arm of the ATM pathway , however , it was unclear if the NBS1/SMC1 factors , which form a separate and independent arm of the ATM pathway , played any role[19] . Both arms act independently and both mediate cell cycle arrest in late S/G2 , which this is consistent with the timing of HPV genome amplification [9] . In our studies , the partial knockdown of SMC1 was sufficient to block genome amplification upon differentiation and indicates that both the CHK2 and SMC1/NBS1 arms of the ATM pathway provide essential activities . This is further supported by our observations that pSMC1 forms complexes with γ-H2AX and CHK2 in HPV positive cells together forming replication centers associated with the viral DNA . Recent studies have also confirmed the presence of NBS1 at HPV-31 viral foci and its importance in HPV 31 genome amplification [36] . Proximity ligation analyses together with chromatin immunoprecipitation experiments indicate that SMC1 is recruited to HPV genomes through complex formation with CTCF insulator proteins . SMC1 is preferentially recruited to the L2 region of HPV 31 that contains three canonical CTCF binding sites . In normal cells , SMC1 forms complexes with only a subset of CTCF proteins and CTCF can bind DNA in the absence of cohesins[37] , [38] . In our studies , we identified CTCF sequence motifs in L2 , L1 and a potential site in E2 . The three motifs in L2 are canonical sites that are separated by spacing typically seen in active CTCF sites[39] . Chromatin immunoprecipitation analyses demonstrated similar levels of SMC1 binding to the L2 region in undifferentiated and differentiated cells while the levels of γ-H2AX binding increased . Importantly , mutation of these sites results in abrogation of binding of both SMC1 and CTCF and viral genomes containing these mutations quickly lose the ability to be stably maintained as episomes as well as to amplify upon differentiation . Finally , transient knockdown of CTCF with shRNAs inhibited amplification and transcript levels . Since we were not able to establish cells in which CTCF is stably knocked down , we were unable to screen these cells for an effect on long-term stable maintenance of episomes . Overall , these findings implicate SMC1 interactions with CTCF as critical for amplification of HPV genome upon differentiation . The inability of genomes with mutated CTCF sites to be stably maintained as episomes in undifferentiated cells was unexpected and indicates that CTCF may have roles in the HPV life cycle beyond recruiting pSMC1 to viral DNAs as part of the ATM DNA damage response . CTCF complexes mediate insulator function by blocking adjacent enhancer function and CTCF complexes bound to HPV genomes may also act as insulator elements that shield cryptic viral promoters from activation by the HPV enhancer element in the URR[34] , [29] . This could provide specificity to the activation of the HPV enhancer , permitting only unidirectional transcription from early and late promoters around the E6 and E7 ORFs . In support of a role of CTCF in regulating HPV transcription , we observed reduced levels of HPV early transcripts in CTCF knockdown cells . CTCF has also been shown to play a role in regulating chromatin assembly , gene regulation and looping in herpesviruses[40] . The Lieberman group has shown that in KSHV and EBV , CTCF is responsible for organizing chromatin loops that regulate latent gene expression and mutation of CTCF binding sites reduces stable genome copy number[41] , [42] . It is likely that CTCF functions to organize intramolecular loops on HPV genomes in a similar manner . These loops could form between different CTCF sites as well as potentially with E2 factors bound to URR sequences . These sites could be located on the same genome or on nearby sister genomes . The E2 proteins have been shown to form complexes with ChlR1 to promote proper viral genome segregation . ChlR1 has also been shown to bind SMC1 as well as other members of the cohesin complex [43] . This suggests that CTCF’s role in the stable maintenance of HPV episomes could be through its association with SMC1 and tethering factors such as ChlR1 . In this way intermolecular looping with other viral genomes as well as cellular CTCF sites may occur . DNA looping mediated by CTCF/cohesion complexes is a dynamic process that changes through the cell cycle , and could play a role in modulating the structure of HPV genomes during the differentiation dependent life cycle . Further analysis of looping of viral DNAs can shed light on the role CTCF/cohesin complexes in genome maintenance . Our studies suggest that the role of SMC1/CTCF complexes in HPV genome amplification is part of the ATM-dependent homologous recombination repair pathway . In normal cells homologous recombination repair of DNA occurs in G2 after chromosomes have replicated and sister chromatids are brought into proximity through the action of cohesins . In HPV positive cells , SMC1 cohesin complexes may be functioning in a similar manner during amplification by aligning multiple HPV genomes together in G2 . HPV amplification has been shown to be dependent on homologous DNA repair factors such as RAD51 and BRCA1[20] . In addition , amplification occurs in G2 , which is when homologous recombination occurs[6] . SMC1 is also a component of the RC-1 homologous recombination complex that contains single strand binding protein RPA and RAD51 , as well as the repair polymerase ε and DNA ligase III [18] . It is possible that SMC1 helps to recruit homologous repair polymerases to viral genomes to mediate replication of viral genomes during the amplification process . Repair polymerases such as DNA polymerase epsilon and ligases such as DNA ligase III could be the principal mediators of viral amplification rather than polymerase α and polymerase δ that replicate host chromosome DNAs in early to mid S-phases . Since HPV genomes are only approximately 8 kilobases in size , replication could easily be accomplished by repair polymerases . In normal cells , SMC1 proteins are bound at discrete sites on chromosomes and upon DNA damage they become phosphorylated and are recruited to sites of double strand breaks . It is possible that double strand breaks are generated in HPV genomes prior to or during amplification and that SMC1 cohesin protein complexes migrate from the sites in L2 to double strand breaks . The mechanism responsible for the generation of double strand breaks in HPV genomes is unclear and to date we have no evidence that double strand breaks are present in HPV genomes . If these sites exist and are generated in a random fashion , then SMC1 binding might not exhibit sequence specificity other than to the L2 sites . If double strand breaks are generated at distinct sites during HPV genome amplification , then there should be specificity of recruitment of SMC1 and other ATM factors to these sites that should be detected by Southern analysis . A rolling circle model of replication has been suggested for HPV genome amplification and this would require site-specific cleavage of viral DNA concatamers[44] . This model however remains controversial and it is equally possible that HPV genomes amplify through theta structure replication . In theta replication the resolution of interlocked replicative intermediates would also require the generation of double strand breaks during cleavage/religation of circular intermediates and such an activity has been suggested for ATMs role in SV40 replication[45] . This mechanism should result in the formation of large replicative structures upon addition of ATM inhibitors but we have not observed such complexes of HPV genomes in our preliminary studies . Further understanding of the mechanism by which homologous recombination factors together with SMC1 and CTCF contribute to HPV genome amplification will require a detailed structural analysis of viral DNAs undergoing amplification . During the submission of this study , we became aware of a study from the Parish lab ( Paris et al . ) that identifies CTCF sites in HPV 18 and demonstrates CTCF to be a critical regulator of HPV 18 gene expression and splicing . These studies are in agreement with our work and demonstrate that CTCF is an important regulator of HPV biology . Overall , our studies identify SMC1 and CTCF as critical regulators of viral replication during the differentiation-dependent life cycle of human papillomaviruses .
The human keratinocytes used in this study were obtained from discarded foreskin circumcisions from anonymous donors by the Keratinocyte Core in the Northwestern University Skin Disease Research Center ( SDRC ) and are not classified as human subjects research . These specimens were not specifically collected for this study and lack all identifiers . Human foreskin keratinocytes ( HFKs ) were isolated from neonatal human epidermis . HFKs containing HPV genomes were cultured in E-medium supplemented with mouse epidermal growth factor ( EGF ) ( Becton Dickenson , Franklin Lakes , New Jersey , Cat . #354010 ) . The HFKs were cultured on J2 fibroblasts arrested with mitomycin-c as previously described[46] . CIN-612 cells , which maintain the HPV-31 viral genome episomally were cultured in E-medium + EGF , with mitomycin-c treated J2 feeders . pBR-322min containing HPV 31 was used for the generation of HPV 31 cell lines ( HFK-31 ) along with pSV-Neo2 a neomycin resistance plasmid and used for stable cell line selection . Validated shRNA constructs in the pLKO . 1 lentiviral background targeting SMC1 and CTCF were purchased from Sigma-Aldrich ( St Louis , Missouri ) as was pLKO . 1 shGFP . pVSVG and pGag-Pol-Tat-Rev were used for the generation of lentiviral particles . Recircularized HPV-31 was created by the removal of the pBR-322 backbone from pBR-322min-HPV31 and subsequent religation . Recircularized HPV-31 genomes were contransfected with a neomycin resistance expression vector using FuGene-6 ( Roche ) , after versene treatment to remove J2 fibroblast feeders . One million cells were seeded on each 100-mm dish and transfected with plasmid DNA once at 60% confluency . After 24 hours , cells were selected with G418 and expanded as previously described[47] . The three CTCF consensus sequences within the L2 coding region of the HPV-31 genome ( pBR-322 ) were mutated using site-directed mutagenesis so as to maintain L2 coding sequence intact and verified by sequencing ( Clontech , Mountain View , California , R045A ) . Site one starting at nucleotide 4387 5’-GTC|CCT|CTT-3’ was changed to 5’-GTC|CCG|CTT-3’ . Site two starting at nucleotide 4477 5’-CCC|TCT|-3’ was changed to 5’-CCA|AGC|-3’ . Site three starting at nucleotide 4543 5’-CAC|CCT|CCT|AC-3’ was changed to 5’-CAT|CCG|CCG|AC-3’ . Stable cell lines containing the WT or the 3X L2 mutant were transfected into the same HFK background and stably selected as described above . To induce differentiation , HPV negative and positive HFKs as well as CIN-612 cells were switched to Invitrogen’s M154 with HKGS and Pen-Strep and . 03mM filter-sterilized calcium chloride for 24 hours and allowed to grow to 80% confluency . The media was changed to Invitrogen’s M154 without HKGS plus Pen-Strep and 1 . 5mM filter-sterilized calcium chloride , and cultured for 72 hours to induce differentiation as previously described [48] . HPV negative and positive HFKs as well as CIN-612 cells were cultured as outlined above . To induce differentiation , 3 million cells were harvested following versene treatment to remove J2 fibroblast feeders . The cells were then resuspended in 1 . 5% methylcellulose that was dissolved in E-media . Following resuspension , the cells were allowed to grow for 48 hours . They were subsequently harvested for protein lysates or total genomic DNA as previously described [49] . Whole cell lysates were extracted using RIPA lysis buffer , following versene treatment for the removal of J2 3T3 feeder fibroblasts . Protein was quantitated using the Bio-Rad ( Hercules , California ) bradford assay and 40 g of protein was loaded per well , and run on 8 . 0% , 10 . 0% , or 4–20% SDS-polyacrylamide gels . The gels were wet transferred to Immoblion-P , PVDF ( Millipore , Billerica , Massachusetts ) and probed with the primary and secondary antibodies . ECL ( GE , Pittsburgh , Pennsylvania ) was used as the chemiluminescent substrate . Phospho-CHK2-Thr-68 ( Cell Signaling , Danver , Massachusetts , Cat . #2661 ) , CHK2 ( Cell Signaling , Cat . #3440 ) , Phospho-SMC1 ( S957 ) ( Abcam , Cambridge , Massachusetts , Cell Signaling , Cat . #ab1275 , 4805 , Bethyl , Cat . #A304-147A-M , Montgomery , Texas ) , SMC1 ( Abcam , Cell Signaling Cat #ab9262 , 4802 ) , CTCF ( BD , Franklin Lakes , New Jersey , Cell Signaling , Millipore Cat #612149 , 2899 , 07–729 , 17–10044 ) , γ-H2AX ( Millipore , Cell Signaling , Cat #05–636 , 5438 ) , GAPDH ( Santa Cruz , Paso Robles , California ) . Cells were seeded at subconfluence onto coverslips ( No . 1 ) . After 24 hours , the cells were treated with 4% Methanol-free formaldehyde in PBS , differentiated in calcium as described above . Next , the cells were permeabilized in PBS + 0 . 1% Triton X-100 ( PBT ) and blocked in Normal Goat Serum ( NGS+T ) ( Invitrogen ) with 0 . 1% Triton X-100 . Primary antibodies were added to the NGS+T according to manufacturer’s specifications or at 1:50 if not specified . Washes were performed in PBST . Secondary antibodies ( Alexa Fluor , Invitrogen ) were added to NGS+T . All incubations were performed in a humidity chamber . Slides were mounted using Pro-Long Gold Antifade + DAPI ( Invitrogen ) , or Gelvatol ( homemade ) and DAPI and imaged using a Zeiss Axioscope . Image analysis performed in Axiovision and Fiji Is Just Image J ( FIJI ) . Cells were seeded at subconfluence on coverslips ( No . 1 ) . All cell-types were seeded without J2 fibroblast feeders . After 24 hours , the cells were treated with 4% Methanol-free formaldehyde in PBS . Subsequently the cells were permeabilized in PBS + 0 . 1% Triton X-100 ( PBST ) and blocked in Normal Goat Serum ( NGS+T ) ( Invitrogen ) with 0 . 1% Triton X-100 . Primary antibodies were added to the NGS+T according to manufacturer’s specifications or at 1:50 if not specified . Washes were performed in PBST . Secondary antibodies , covalently linked with plus/minus oligos were diluted in NGS + T as per the manufacturer’s instructions ( OLINK DuoLink , Upsalla Sweden ) . Ligations , and polymerase amplifications , and washes were performed as per the manufacturer’s instructions . All incubations were performed in a humidity chamber . Chromatin immunoprecipitations were performed using CIN612 or stable HPV-31 ( WT or L2 Mutant 3X ) cells grown as described above . Before fixation , cells were treated with versene to remove J2 fibroblast feeders . Cells were then seeded and grown to 80% confluency . Chromatin-immunoprecipitation was performed as previously described using SMC1 ( Abcam , Cat . #ab9262 ) , γ-H2AX ( Millipore Cat . #05–636 ) , CTCF ( Millipore , Cat . #07–729 ) , Normal Rabbit IgG ( Santa Cruz , Cat . #SC-2027 ) , Normal Mouse IgG Antibody ( Santa Cruz , Cat . #SC-2025 ) , Normal Mouse IgG ( Millipore , Cat . #12-371B ) and Protein-G Dynabeads ( Invitrogen Cat . #100-03D ) [50] . Real-time , touchdown PCR was performed with the Lightcycler 480 ( Roche ) and the HPV-31 CTCF Trio Primers and URR Primers; Forward: 5’-TTTGGTGGGTTGGGTATTGG-3’ , Reverse:5’-GTAGGAGGCTGCAATACAGATG-3’ . Forward , 5’-AACTGCCAAGGTTGTGTCATGC 3’ , Reverse , 5’-TGGCGTCTGTAGGTTTGCAC-3’ . Two MISSION pLKO . 1 shRNA constructs for SMC1 ( Sigma ) , two constructs targeting CTCF , or shGFP ( shCTRL ) were transfected ( 3g ) into 50% confluent 293T cells along with pVSVG and pGag-Pol-Tat-Rev via X-tremeGENE HP DNA Transfection Reagent ( Roche , Indianapolis , Indiana , Cat . #06366236001 ) . 24 hours post-transfection , the media was changed and the cells were allowed to grow for another 24 hours . The viral supernatants were collected , and concentrated using a Millipore concentrator ( Cat . #UFC910096 ) . For infections , 2 million cells were seeded and allowed to grow for twenty-four hours . The concentrated viral particles were either pooled or singly distributed equally to cells in the presence of 0 . 8 g /ml Polybrene ( Sigma-Aldrich ) in 3 mls of media . After 6 hours , an additional 4 mls of media were added . Twenty-four hours post-transduction , the media was changed , and the cells were allowed to grow for another 24 hours . Cells were then either harvested or differentiated using methylcellulose . Knockdown was confirmed using western blot analysis . CIN612 cells were treated with or without ATM-kinase inhibitor KU-60019 ( 10uM , Tocris , Bristol , UK , Cat . #4176 ) or Caffeine ( 100mM Stock ) ( Sigma , Cat . # C0750 ) or their respective vehicle ( DMSO , PBS ) in E-medium supplemented with mouse epidermal growth factor ( EGF ) for the indicated times as previously described [26] , [51] . The cells were subsequently harvested for protein . CIN-612 or HPV-31 ( WT or L2 Mutant 3X ) cells were treated with versene to remove J2 3T3 feeder fibroblasts . Southern blot lysis buffer ( 400 mM NaCl , 10 mM Tris-HCl , 10 mM EDTA ) was used to induce lysis of undifferentiated or calcium or methycellulose-differentiated cells . Cells were subsequently treated with RNase A ( Sigma-Aldrich ) and proteinase K ( Sigma-Aldrich ) . Total DNA was isolated by phenol-chloroform extraction , and Southern analysis was carried out as previously described[52] . Total RNA was collected in STAT-60 and isolated by phenol-chloroform extraction . Northern analysis was carried out as previously described [52] . | Over 120 types of human papillomavirus ( HPV ) have been identified , and approximately one-third of these infect epithelial cells of the genital mucosa . Infection by a subset of HPV types is responsible for the development of cervical and other anogenital cancers . The infectious life cycle of HPV is dependent on differentiation of the host epithelial cell , with viral genome amplification and virion production restricted to differentiated suprabasal cells . While normal keratinocytes exit the cell cycle upon differentiation , HPV positive suprabasal cells are able to re-enter S-phase to mediate productive replication . HPV induces an ATM-dependent DNA damage response that is essential for viral genome amplification in differentiating cells . In this study we demonstrate that a protein that mediates sister chromatid association prior to mitosis , SMC1 , plays a critical role in the differentiation-dependent replication of HPV through the recruitment of DNA damage proteins to viral genomes . SMC1 binds specifically to CTCF binding sites in the late region of HPV through association with the DNA insulator protein CTCF . Knockdown of either SMC1 or CTCF abrogates viral genome amplification . Further , mutation of CTCF sites in the late region of the HPV genome results in loss of both episomal maintenance and the ability for SMC-1 and CTCF to interact with the genome . Our findings identify an important regulatory mechanism by which HPV controls replication during the productive phase of the life cycle , and this can lead to new targets for the development of therapeutics to treat HPV induced infections . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | Human Papillomaviruses Activate and Recruit SMC1 Cohesin Proteins for the Differentiation-Dependent Life Cycle through Association with CTCF Insulators |
This article presents the integration of brain injury biomechanics and graph theoretical analysis of neuronal connections , or connectomics , to form a neurocomputational model that captures spatiotemporal characteristics of trauma . We relate localized mechanical brain damage predicted from biofidelic finite element simulations of the human head subjected to impact with degradation in the structural connectome for a single individual . The finite element model incorporates various length scales into the full head simulations by including anisotropic constitutive laws informed by diffusion tensor imaging . Coupling between the finite element analysis and network-based tools is established through experimentally-based cellular injury thresholds for white matter regions . Once edges are degraded , graph theoretical measures are computed on the “damaged” network . For a frontal impact , the simulations predict that the temporal and occipital regions undergo the most axonal strain and strain rate at short times ( less than 24 hrs ) , which leads to cellular death initiation , which results in damage that shows dependence on angle of impact and underlying microstructure of brain tissue . The monotonic cellular death relationships predict a spatiotemporal change of structural damage . Interestingly , at 96 hrs post-impact , computations predict no network nodes were completely disconnected from the network , despite significant damage to network edges . At early times ( ) network measures of global and local efficiency were degraded little; however , as time increased to 96 hrs the network properties were significantly reduced . In the future , this computational framework could help inform functional networks from physics-based structural brain biomechanics to obtain not only a biomechanics-based understanding of injury , but also neurophysiological insight .
The finite element method is often used to study neurotrauma [1]–[9] and continues to emerge as a useful tool in the field of neuroscience [10]–[13] . Models continue to advance in biofidelity by incorporating an increased level of anatomic detail [4] , [13] , [14] , improved representation of the material behavior at various loading rates [15]–[20] , and advanced measures and predictions of injury [8] , [21] , [22] . Finite element models are commonly used to understand the biomechanics of brain and skull deformation when the head is subjected to insult , leading to improved insight into mechanisms of acute injury . For example , modeling axonal injury mechanisms within white matter of the brain has been the focus of some recent efforts and provides a means to relate an insult to a cellular injury mechanism [8] . By using diffusion tensor imaging ( DTI ) fiber tractography , the structural orientation of neuronal axonal bundles can be incorporated into the finite element model and can be used to compute the axonal strain during brain white matter deformation [8] , [23] . Then , by using an axonal injury threshold , the occurrence of diffuse axonal injury ( DAI ) is predicted [8] . Wright and Ramesh [8] show that the degree of injury predicted is highly dependent on the incorporation of the axonal orientation information and the inclusion of material anisotropy into the constitutive model for white matter . By modeling the underlying mechanism of DAI , an enhanced understanding of the neurotrauma is attained through a spatiotemporal description of tissue deformation . In general , finite element simulations may help to elucidate the injury mechanisms of neurotrauma . As finite element models advance , experimentally based models of neurotrauma also continue to become more sophisticated , ranging from the macroscopic [24] , [25] to the cellular level [26]–[29] . Various biomechanical and physiological injury thresholds for neurotrauma have been proposed in the past , including intracranial pressure [21] and strain [8] , [30] , [31] . While these thresholds offer an immediate prediction of injury , they lack a long-term description of functional degradation . A time-evolving injury model is attractive since the biological response occurs on a slower time scale than an injurious stimulus as a consequence of mechanotransduction cascades [27] . There have been significant efforts to develop empirically based time-evolving cellular injury thresholds , which many times use in vitro cellular and tissue culture models . Morrison III et al . [27] suggest that for brain biomechanics , neuronal culture models that accurately mimic specific brain features can be used to explore tissue properties and tolerances or thresholds to mechanical loading . Cell death has been primarily used as a definition of injury within the neuronal culture model community and has been applied to determine tissue-level tolerance criteria using local values of axonal strain , strain rate , and time from “insult” [32] , [33] . Furthermore , as Morrison III et al . [27] point out , empirical functions for cellular death based on culture models could be incorporated into finite element analyses , thereby enabling biological predictions to supplement mechanical predictions of local tissue stress and strain . In the study presented here , this concept is further explored and used as a bridge to a network-based analysis of the brain . As physics-based models become more capable of predicting tissue-level injury mechanisms from improved computational and experimental resources for biomechanics , there remains a need to understand how structural damage in a given location of the brain evolves , and how it may influence functional or cognitive performance over time . Such a goal is complex and difficult . For example , as Kaiser et al . point out [34] , in some instances the brain can be robust to physical damage , and in other instances physical damage can cause severe functional deficits . Nevertheless , Kaiser et al . [34] pursue the important and unanswered question: Are the severity and nature of the effects of localized damage predictable ? To help explore the answer to this question , tools are being developed for the quantitative analysis of brain network organization , based largely on graph theory [35] , [36] . Typically , network nodes represent brain regions , often obtained from high resolution magnetic resonance imaging ( MRI ) . The network links or edges between brain regions represent interregional pathways that convey neuronal signals and are commonly obtained from non-invasive DTI or diffusion spectrum imaging ( DSI ) [37] , [38] . The connection matrix of the network of the human brain forms the so-called “human connectome” [35] , [36] , [39] , [40] . Recently , Jirsa et al . [41] used connectomics to establish a framework for a “virtual brain” , in which network modeling is used to understand the intact and damaged brain . Similar to previous work , hypothetical or random deletion of nodes or edges were used to degrade the structural connectome [42] . More recently , structural and diffusion images from 14 healthy subjects were used to create spatially unbiased white matter connectivity importance maps that quantify the amount of disruption to the overall brain network that would be incurred if that region were compromised [43] . In this study , we attempt to extend the capabilities of neurocomputational models by providing a physics-based approach for predicting degraded regions of interest . Physics-based injury predictions may help inform structural connectome analysis . In this study , neurotrauma is investigated by using finite element simulations of a single individual subjected to a simulated head impact . Tissue damage is computed using empirically based damage models that provide a link from macroscopic biomechanical deformation to mesoscopic damage . Axonal bundle tracts are explicitly modeled using a multiscale description of white matter tracts obtained from diffusion tensor imaging . Then , using the physics-based injury predictions for white matter tissue from finite element simulations , the structural brain connectivity or connectome is degraded , and various network measures are computed . This is an important contribution because finite element simulation predictions of tissue damage provide physics-based reasoning for removing nodes or degrading edges to create the “damaged” brain connection matrix . In turn , this approach may provide further insight into mild traumatic brain injury by shedding light on the relationship between mechanical stimulus to the brain and neurobiological processes that result . Furthermore , if successful , the computational framework presented herein could supplement ongoing efforts to evaluate the use of non-invasive medical imaging tools , such as diffusion tensor and spectrum imaging , to detect white matter disruption for neurotrauma diagnostics [44] , [45] by providing a time-evolving history of tissue injury .
T1 and diffusion tensor magnetic resonance images are taken from a single individual ( the corresponding data can be found in [37] ) . The T1 image ( Figure 1a ) is segmented into different head materials ( Figure 1b ) using the software Amira [46] and the Connectome Mapper Toolkit [47] . The segmented geometry is then used to create a biofidelic three-dimensional finite element volume mesh ( Figure 1c ) . In order to create a corresponding structural connectome or network , the T1 image is parcellated into 83 regions of interest ( ROI ) representing the location of anatomical regions of the brain based on the Desikan-Killiany atlas extended to include subcortical regions [48] . Since diffusion tensor images ( Figure 1e ) are used to generate axonal bundle fiber tractography based on the direction of peak water diffusion in each voxel [49] , the DTI fiber tractography ( Figure 1f ) represents the approximate location of neuronal axonal bundles [50] . Fibers are filtered for connectome creation to include only fibers that begin and end within ROI . The structural connectome is assembled using the Connectome Mapper Toolkit [47] and is composed of nodes representing ROI generated at the centroid , connected by edges that represent structural pathways for which the DTI tractography traverses . Following the segmentation enabled by the Desikan-Killiany brain atlas , there are 83 network nodes and 1029 network edges . Furthermore , DTI fiber tractography is incorporated into the finite element model by using a transverse isotropic material model specifically developed for representing white matter tissue [23] , [51] where each finite element within white matter regions is assigned an orientation based on the superposition of fiber tractography [52] . Details of the numerical implementation are published elsewhere [52] , but it is important to point out that various different cases needed to be considered while assigning one orientation per finite element . For example , in the case of multiple fibers that overlap the spatial bounds of a single element , the multiple fiber orientations were averaged . Herein , finite element simulations of the human head are designed to mimic experimental conditions for cadaveric impact tests , which are conducted to understand the dynamics of a frontal impact and the associated compression-tension damage [53] . In this study , the explicit dynamic finite element method [54] is used to capture the transient response of head impact . Supplementary information about the finite element method is included in the supporting information , Text S1 . In addition to a volume mesh for the head , the finite element model also requires material constitutive descriptions and properties for all its components: skull , cortex , brain stem , cerebrospinal fluid , and soft tissue , which represents a homogenized mixture of muscle and skin . A detailed list of the material constitutive laws used to relate tissue strain to stress is included in the supporting information , Text S2 . Once again , we point out that diffusion tensor imaging is used to inform the constitutive model for white matter tissue , which has been applied in the past for studying non-human primates [31] and human injury [8] , [22] , [55] . The entire finite element model consisted of 1 , 394 , 945 tetrahedral elements and 237 , 115 finite element nodes . For boundary conditions , the bottom of the neck is fixed , and a force was applied to a circular area on the forehead ( about ) in the anteroposterior direction . The input force-time curve was a sinusoidal shape with a peak force of at . The finite element simulation computes the time-evolving mechanical strain and stress in the direction of axonal fiber bundles . The strain in the direction of axonal bundles is referred to as the axonal strain . One limitation of the current finite element model is the exclusion of viscoelasticity in the constitutive description of brain matter . The authors acknowledge that to accurately model the progression of damage , the constitutive model should be extended to account for the time-dependent behavior of brain tissue . The exclusion may have an effect on the outcome of our results , leading to larger shear stresses , but smaller shear strains , thus , less predicted damage . For example , Chafi et al . [56] shows that viscoelasticity plays a major role in the dynamic response of the brain under blast loading . Future efforts are focused on improving the mechanical description of brain tissue . In order to model damage using a physics-based approach , either an explicit failure mechanism should be modeled or an empirically based failure threshold is required . For this study , measures of axonal strain and strain rate computed for white matter regions are used as input for empirically based injury threshold predictions that are obtained from cellular culture experiments . Specifically , experimental results for cellular death are described using a mathematical function for tolerance criteria that relates strain to resultant cell death evaluated for up to four days post-injury [27] , [32] , [33] . Experimental data exists for the CA1 and CA3 regions of the hippocampus , dentate gyrus , and cortex for the rat . Experiments suggest that some regions of the rat brain are sensitive to loading rate , while other brain regions are not . That is , Elkin et al . [33] found that cortical cell death was dependent on applied strain rate , whereas hippocampal cell death was not . The relations used in the present model , obtained from Morrison III et al . [27] , are: where is time from insult , is the local strain , and is the local strain rate [27] . Similar to Morrison III et al . [27] , the units of time are days , strain is dimensionless and strain rate is inverse seconds . The damage parameter , D , is defined as the percent area of cell death . Using Equations 1–3 , the axonal strain and strain rate from the finite element simulations , as well as time , are used to calculate percent cell death for a tissue region . Due to insufficient resolution of the MRI data used in this study , segmentation of the hippocampus into CA1 , CA2 and dentate gyrus regions was not possible , thus the more conservative equation for was used to compute cell death for the entire hippocampus . Equations 1–3 are monotonic functions that only increase with time , thus cellular repair mechanisms and regeneration are not currently captured . Potential issues with monotonic cellular death predictions , as well as using rat brain injury thresholds instead of human cellular injury thresholds , will be discussed later . Since the explicit dynamic finite element method is used , the transient wave propagation for solid mechanics is resolved; however , this is computationally costly and limits the total time of biomechanical prediction . The explicit dynamic simulation runs to 15 ms , thus the “long-term” structural mechanics of brain swelling , relaxation , etc . , are not captured in the current model , although could be adapted in the future by using quasi-static , implicit finite element solvers . The reader should understand that we use the local tissue strain and strain rates predicted from finite element simulations of short duration , about , as input to experimentally based cellular death models that were developed over a hour period . This assumes that the tissue damage due to large deformations occurs immediately and initiates an injury process that grows with time . This assumption is based on the observation that cell death was not immediate in response to deformation , but instead , increased over 4 days after injury [27] . It should be noted that the cellular death estimates that Morrison III et al . [27] developed were not based on axonal strain but instead on nominal strain applied to the back of the substrate on which the neuronal cells were attached . Thus , Equations 1–3 may not accurately describe the actual relationship between axonal strain and injury . Furthermore , the neuronal cell bodies in the experiments were not aligned in a specific orientation , so the response is not strictly representative of axonal strain injury . There are research efforts examining stretching of individual axons but have not proposed empirical relationships for cellular death in terms of applied strain , rate of loading and time from insult [57] , [58] . Experimental measurements of the cellular response of white matter , especially axonal bundles , would be interesting to explore and could provide a more accurate representation of cell damage in the future . Furthermore , the experiments performed by Morrison III et al . [27] used the rat hippocampus , which is mostly comprised of gray matter . Thus , it should be noted that the empirical data described in Equations 1–3 was not intended to describe white matter injury response so there may be limitations in applying Equations 1–3 to predict cellular white matter injury . Cellular injury threshold data for isolated white matter is currently limited; however there have been some efforts to characterize mechanical damage to axons [57]–[59] . We hope a computational framework described here will help to motivate efforts to obtain different white matter and gray matter empirical functions , which could then be used for each of region of the brain separately . As mentioned before , the application of this approach raises the possible need for understanding properties of white matter fiber bundles . In order to map the finite element results to network-based analysis tools , output data from each finite element that represents white matter is mapped to a corresponding voxel in the MRI data that is used to create the DTI tractography . This mapping is referred to as the element-to-voxel map . Multiple finite elements within a single voxel are averaged . The element-to-voxel map enables voxels to be assigned additional data , including axonal strain and strain rate from the finite element simulation . Alstott et al . [42] chose to generate brain lesions by altering the structural connectivity matrix of the brain by deleting nodes using various methods . In the present work , instead of deleting nodes , we degrade the edges of the network based on the computed cellular death at the voxel level . To understand how the structural network is degraded , consider the schematic shown in Figure 2 . Edges in the network are constructed from voxels that connect two different ROIs . The edge strength is relative to the number of tracts between two ROI . Cellular death for each voxel is computed using Equations 1–3 and may grow to reach the chosen critical value of cellular death , ( discussed shortly ) . Voxels with a predicted cellular death greater than are shown in red in Figure 2 . For this study , if a tract traverses a voxel that is greater than the threshold , the entire tract is considered damaged , thus the edge strength is decreased . This procedure is similar to that used by Kuceyeski et al . [43] who simulated lesions at each voxel and removed tracts passing through the lesion in order to create a damaged network to enable analysis of changes in network measures and the evaluation of the importance of each voxel . Similar to Honey et al . [60] , the connection strengths were resampled to a Gaussian distribution with a mean of and a standard deviation of . As Alstott et al . point out [42] , this transformation does not alter the rank-ordering of strong to weak pathways , but simply compresses the scale of connection strengths . There are other possible ways to degrade the structural network that will be discussed later . Since Equations 1–3 predict some degree of cell death for non-zero values of axonal strain , strain rate , and time , the additional critical value of cell death , , was required as a rule in order to degrade a voxel . The critical cellular death of was chosen because it predicted similar levels of damage as compared to other proposed thresholds , including 18% axonal strain , which was used previously as a threshold that indicated degradation in electrophysiological function [30] . During the calculation of cellular death , positive values of strain and strain rate were used ( negative strain was not used to calculate cellular death ) . The choice of injury threshold , as well as the micromechanics of axonal fiber bundles , is an active area of research that would assist making the current methodology more accurate in the future . During the dynamic simulation , strain and strain rate data for each voxel is output at increments . Cell death is calculated at each voxel for each of these increments . Then , the maximum cell death value calculated in each voxel is used to predict cell death up to . It is important to note that cell death calculations are the result of the combination of variables at each increment rather than considering each variable independently .
The deformed configurations of the head , along with contours and response curves for various locations within the brain ( frontal , parietal , occipital , temporal , corpus callosum and cerebellum ) are shown in Figures 3a–f . Output variables including pressure , axonal strain , and effective strain rate , useful for understanding the anisotropic biomechanical response for white matter , are shown . Confidence in the finite element model is established by comparing output from the computations to pre-existing experimental data on cadaveric head impact [53] . The values of axonal strain and strain rate were taken from specific locations in each of the six regions plotted in Figure 3 . Four out of six of the locations represented the locations of the pressure transducers for the experiments of Nahum et al . [53] . Locations within the cerebellum and corpus callosum were also added for the strain and strain rate analysis . A variation of results will exist in each region , and taking an average of values in a anatomical region could help resolve this problem , but including simulation data points that are located farther from the experimental sampling points could also make the results less accurate . The head impact simulation took about 30 hours on 32 processors in order to reach 15 ms . As seen in Figure 3b , the intracranial pressure quickly increases to positive values in the frontal and parietal regions , while quickly increasing to negative pressures in the occipital and posterior fossa regions . The anteroposterior pressure gradient ( seen in Figure 3a ) is commonly observed in experimental and computational studies - giving rise to the so-called coup and contrecoup loading scenario , for which there are an associated number of proposed injury mechanisms . Short duration intracranial pressure gradients with high positive pressures are observed at the coup region , with negative pressures at the contrecoup region . The maximum positive pressure of approximately is reached in about in the frontal lobe , closest to the impact , while the maximum negative pressure of approximately is reached in about in the posterior fossa region . The computed pressure response is directly compared to Nahum et al . 's [53] experimental results in Figure 3b and show similar trends . In addition , validation of the strain response against cadaveric experiments of Hardy et al . [61] is also described in the Supplemental Text S3 . Axonal strain at the various brain regions responds slower than the pressure response . The axonal strain begins to substantially grow at 1 ms and shows a gradual rate of change of strain , with a maximum of 33% at over the duration of the simulation within the temporal region . Unlike the pressure , an obvious transcranial gradient is not apparent for the axonal strain . While the frontal region had the highest predicted pressure , the temporal and occipital brain locations have the largest values of axonal strain in the regions specifically examined . Later , in the structural network analysis , these areas are associated with the largest amount of cellular death . Interestingly , if a threshold of injury of 18% axonal strain is chosen [8] , [30] , our results show the onset of injury occurs at approximately 9 . 1 ms within the temporal lobe , despite the intracranial pressure reaching approximately in . The time scales at which the pressure and strain grow will be discussed later in the context of injury cascades . The effective strain rate , also commonly referred to as rate of loading or loading rate , has a maximum value of approximately in the temporal lobe . From the contours shown in Figure 3e there appears to be a strain rate focusing , with lower strain rates closer to the skull and higher values more central to the brain . Shear strain focusing has been reported earlier [3] , [62] and is attributed to the partial conversion of energy of the axial impact into a shear mechanical stress wave as a result of the material response of the brain , cerebrospinal fluid , and skull [62] . The maximum strain rate is observed before maximum axonal strain because the axonal strain accounts for magnitudes of deformation , while the strain rate relates to the rate of change of strain . The axonal strain and strain rate output from the finite element simulations are used to compute the amount of cellular death , according to Equations 1–3 . Figure 4 shows the evolution of damaged tractography up to 96 hrs post-impact , using a critical cellular death of as a threshold for injury and the corresponding evolution of the degraded structural network for sagittal and transverse views . The edges in the network , which are fully damaged , are shown in red for visualization . In reality , each edge is weighted and has degraded values before it is fully damaged . The network nodes are scaled by the percentage of connections that were removed , so that larger nodes have lost more connections compared to original values ( ) . The sagittal and transverse views are shown to provide insight about where the damage is predicted and how it progresses through time . At 24 hrs , the top four regions affected include the cuneus ( medial surface of left cerebral hemisphere in the occipital lobe ) , fusiform gyrus ( temporal lobe ) , lingual gyrus ( occipital lobe ) , and peri-calcarine . Together , damage in these brain regions have 559 tractography fibers removed from network edges . Note that the total number of fiber tracts prior to impact was 497 , 442 , so this damage corresponds to a 0 . 011% degradation of tractography . The percentage of fully degraded edges for 24 , 48 , 72 , and 96 hrs are listed in Table 1 . Using an 18% strain criterion , 17 . 3% of the edges were fully degraded . Figure 4b shows that structural damage occurs first in left cuneus and the right superiotemporal regions , resulting in one fully damaged edge at . The edge between the two regions had only three tract fibers with less than one voxel or between them . The edge also had a high mean fiber length of ( the original total mean fiber length was ) . This suggests that network edges associated with few tract fibers and long fiber length are more susceptible to damage . Asymmetry of the predicted damage occurs due to the asymmetry of the underlying anatomy associated with the finite element mesh , as well as potential asymmetry of the fiber tractography , which has been studied in recent work [63] . At , predicted damage seems to show a high density of damaged fibers in the anteroposterior direction in both hemispheres of the brain . Lateral tractography damage is also predicted within the corpus collusum . As time progresses , fully damaged rostrocaudal tracts are predicted and continue to increase . Structural changes to the tractography and resulting network arise because of the underlying voxel condition . That is , if a tract goes through a voxel that has reached the critical cellular death value , , the tract is removed . Therefore , it is useful to examine how the inherent voxel properties influence results . Figure 5a shows the distribution of voxels above the predicted critical cellular death of 3% as a function of angle between the axonal fiber bundle direction and the direction of the head impact ( i . e . , frontal impact on the anteroposterior axis ) . If the tract fiber direction within a voxel is parallel with the impact direction , the angle is zero . This shows the distribution of damage as a function of angle with respect to the impact direction . Because of the high degree of mechanical rotation observed through the shear focusing in the temporal and occipital brain regions , tract fibers with large angles with respect to the impact direction are damaged initially . This is seen in Figure 5a at 24 hrs ( see red bars in plot ) and in Figure 4a–b . This data can be further analyzed by normalizing the number of damaged voxels for a given orientation by the total number of undamaged voxels with a given orientation from the impact direction . This measure is referred to as the angle-normalized number of damaged voxels . The distribution of angle-normalized number of voxels is shown in Figure 5b , and is important because it takes into account the original distribution of the number of tract fibers eligible to be damaged with respect to the impact direction . For example , Figure 5a shows 18 voxels above the 3% cell death threshold at 48 hrs; thus , one might assume that axonal bundles oriented from the impact direction have little importance . However , when the original number of voxels available to be damaged is taken into consideration ( Figure 5b ) , angles have the highest percentage of damage at 48 hrs . In other words , depending on the underlying white matter microstructure , i . e . , the axonal bundle orientation , with respect to the direction of impact , the node , and edge degradation in the structural network may be affected differently . This is directly related to the structural mechanics ( as opposed to structural connectomics ) of the underlying constitutive or material law used within the finite element simulation . However , because a monotonic function is used to describe the empirical cellular death prediction , this trend becomes more dilute as time progresses ( but should not be extrapolated past 96 hrs since Equations 1–3 are not validated beyond that time ) . From Figure 4g–h , at 96 hours the predicted damaged axonal pathways can be seen in many directions and across all white matter brain regions indicating diffuse structural degradation . To summarize , at 24 hrs fibers in the areas of large rotational tissue strain are susceptible , while at greater times , neuronal tracts that align with the direction of impact seem more susceptible to damage ( using the assumed threshold of cellular death ) . In the future , additional empirical cellular death predictions that are non-monotonic ( if valid ) , or cellular regeneration and repair models , would be useful to explore . Figure 6a–e shows the evolution of the structural connectivity strength matrices predicted as a result of head impact simulations . The original structural connectivity strength is defined as a Gaussian distribution , created by using a mean of and standard deviation of distributed over the total number of edges in the network . Each figure represents a snapshot in time starting with and ending at 96 hours . Within the 96 hr period ( for which the monotonic cellular death criteria are validated ) , as long as regions have non-zero strain and strain rate computed from the finite element simulation , edges in the network become degraded . This decline in network strength is evident in Figure 6e . Since a 3% critical cellular death value is only one possible injury criteria that could be chosen , an additional criterion was examined . Results using two different damage thresholds are shown in Figure 6e and f . Figure 6e is the connection matrix using the 3% threshold at 96 hrs , while Figure 6f is connection matrix using the 18% axonal strain threshold at , as used in previous work [8] , [31] . Connection strengths show qualitatively similar trends in magnitude and location of edge degradation . While the two criteria offer a similar prediction of network damage , the cellular death Equations 1–3 are , perhaps , more useful because effects of strain , strain rate , and time have been decomposed into separate multiplicative terms that allow a compartmentalized study of extrinsic biomechanical conditions . In general , a network's global efficiency represents how well-connected the network is compared to a perfectly connected network [64] and captures the network's capacity for communication along short paths [65] . Herein , values are reported as normalized global efficiency , which is the global efficiency of the network divided by the efficiency of an ideal network . Ideal network efficiency is calculated as a network where all nodes are connected at the minimum cost . The cost for each edge is defined as unity minus Gaussian strength . While strength offers a measure of the capacity to send information , the cost indicates the resistance to sending information - connections with high strength are low in cost . Local efficiency is a network measure that Latora [65] suggests helps to reveal the fault tolerance of the network system and shows how efficient communication between first node neighbors is . For this study , local efficiency is calculated for the same set of nodes used to make the local network of the undamaged network , even if a node loses its edge to the primary node . This method is used to provide a more direct comparison of the network after damage by accounting for edges that were removed . As a result , edges can no longer be part of a short path between nodes and cannot contribute to the efficiency , while also taking into account all nodes that should be a part of the local network without damage . The normalized global and mean local efficiencies as a function of time are shown in Figure 7 . The normalized global efficiency is approximately 0 . 14 at and is about the same at 24 hrs , indicating the network remains capable to send information . However , at 48 hrs an 8 . 8% reduction in normalized global efficiency is predicted and continues to reduce with time . A similar trend is observed in the mean local efficiency . At 96 hrs the normalized global efficiency was reduced 24% , while the mean local efficiency was reduced 27% . By 96 hrs , all but sixteen nodes had greater than 20% reduction of local efficiency . Also , note that there were no nodes completely disconnected from the network , although there were edges completely removed . The total number of edges at was 1029; at , 203 edges were removed using the cellular death threshold . Using an injury threshold of 18% strain , 161 edges were removed . Watts and Strogatz [66] define the small-world network based on the clustering coefficient of the network and the characteristic path length of the network . The clustering coefficient , , is the fraction of triangles around a node and is weighted by the geometric mean of weights associated with edges of a triangle [64] . It measures how well the first neighbors of a node are connected to each other . The characteristic path length , , is the average shortest path between all nodes of the network . To qualify as small-world , the networks clustering coefficient should be greater than that of an equivalent random network , whereas the characteristic path length should be approximately equal [66] . The small-world coefficient is defined as [64] . This value was found to be 1 . 87 , 1 . 87 , 2 . 03 , 2 . 31 , and 2 . 64 for t = 0 , 24 , 48 , 72 , and 96 hrs , respectively . When using axonal strain as the threshold variable , a value of 2 . 38 was computed . The increase seen in the small-world coefficient is due to the more rapid decrease in clustering coefficient of the random network compared to the damaged network that it is based on . In addition , the increase seen in the small-world coefficient also shows the ability of the network to maintain its modular structure more effectively than the random network that it was compared to . The percent reduction of local efficiencies and the associated betweenness for the top 10 regions affected by impact at 96 hrs are listed in Table 2 . Betweenness is defined as the number of times a node is part of a shortest path between nodes , and is one indication of a node's network centrality that help to describe the “importance” of a node [64] , [65] . Hubs within the connectome are identified using betweenness centrality . The maximum betweenness of the original network at is 226 , while the minimum value is zero . Recall from Figures 4a–b and 5a , white matter disruption is observed in the left cuneus and the right superiotemporal regions , which has a betweenness centrality of 41 and 17 , respectively . The moderate to low betweenness of the regions first affected at 24 hrs helps to explain why the normalized global efficiency was not significantly reduced during this time and shows the importance of a robust brain network . At 96 hrs , the right hemisphere lateral orbitofrontal region shows the largest degradation in local efficiency of approximately 45 . 4%; however , that region also has a zero betweenness centrality . On the other hand , the right hemisphere medial orbitofrontal region has a betweenness of 110 and shows a 37 . 4% reduction of local efficiency , indicating it may have a more significant effect to the brain network . At 96 hrs , there are 10 network nodes with betweenness greater than 100 that have at least a 20% reduction of local efficiency . Local efficiency and betweenness for all ROIs are included in Supplemental Figures S2 and S3 , respectively . The results reported thus far are based on a critical cell death value of , as well as a critical strain threshold of 18% . However , it is also useful to evaluate the sensitivity of the results to the choice of . An analysis of network properties was performed for multiple critical cell death values , in the range from , and are shown in Figure 8a–b . Figure 8a shows the percent reduction in total edge strength as a function of the critical cell death threshold choice . Figure 8b shows the percent reduction in global efficiency as a function of the critical cell death threshold choice . Results show that the choice of the critical threshold significantly affects both network measures that were examined . For example , for a change of between 3% or 4% , the reduction in global efficiency is 23 . 9% and 16 . 4% , respectively . In other words , from only a 1% change in the choice of results in a 7 . 5% change in global efficiency reduction . Therefore , the choice of will be important in obtaining accurate results and highlights the need for future experimentation to characterize cellular injury criteria for all areas of the brain in order to improve the accuracy of this modeling approach .
There are exciting possibilities for future work , as well as limitations to the current modeling approach . The current model did not attempt to model the coupled effects to the functional network; instead , we provided an example for a single individual in order to establish a methodology to link physics-based predictions of tissue damage with structural network analysis for frontal impact neurotrauma . It is important to note that the empirical relationships may only be accurate for the rat ( not the human ) , and most likely , there are many more regions that need to be characterized . Our results and conclusions may be altered according to these injury thresholds . Although human injury thresholds are currently limited , as additional brain region injury thresholds are experimentally characterized and improved they can be included in the future . Due to the computational cost of the finite element simulations , the brain was only segmented into 83 different regions . In the future , increased segmentation of regions of interest and improved biofidelity of the finite element model would increase the resolution of the analysis . While this study does not address the resulting functional outcome from structural degradation , coupled structure-function relationships as a result of neurotrauma would be interesting to explore . For example , a coupled analysis may enable functional stimulus that may prohibit or enhance further cell death . For the prediction of tissue damage , additional physics , such as electrochemical reactions , may be useful by incorporating diffusional properties . In addition , increased resolution of the biomechanical response may also be improved by further developing white matter material response descriptions that use multiple fiber tract orientations within a single element , thereby enabling the capability to use diffusion spectrum imaging . There is also an opportunity to use this framework to explore additional injury mechanisms or thresholds from empirical or experimental data , such as intracranial pressure . Note that the current methodology degrades network edges , instead of nodes . In the future , it may be useful to investigate methods for degrading nodes in addition to edges because the all of the nodes represent cortical gray matter , and the gray matter does experience significant strains . This would require a choice of strain measurement other than axonal strain since the gray matter is treated as isotropic . With this type of criteria , it may be possible to degrade a node based on a ratio of damaged voxels within a region of interest compared to the voxel volume of the region as a whole . Further work should also work to validate this approach in humans . There are at least two distinct areas associated with connectome damage that should be explored in order to validate the approach described herein . The first deals with how well the location of damage within the network description is captured using physics-based predictions . The second validation strategy should address the nature in which edges and nodes in the network are degraded . There are various approaches that may be useful for addressing both areas . For example , in order to validate how well the location of damage is captured , further understanding about how cellular level changes effect fractional anisotropy , that results in altered fiber tractography would be useful to develop . Some of this information may be obtained from various ongoing studies that are examining the ability of DTI to diagnose mTBI , which could also be extended to create degraded structural connectomes . One way to do this may be to use DTI studies pre- and post-injury from typical loading conditions that cause rotation-induced diffuse axonal injury . By providing similar loading profiles within the simulation and comparing computed DTI tractography damage with the clinical data set , maybe results can be compared . Perhaps this may be accomplished using sports-related impact injury , such as American football , where many helmets have sensors built-in to record impact loads . In conclusion , this work has explored “connectome neurotrauma mechanics” by using physics-based finite element simulations to help elucidate injury mechanisms associated with neurotrauma by using various cellular injury thresholds to define tissue damage , and established a coupled computational framework to inform structural connectome analysis . | According to the Centers for Disease Control and Prevention in the United States , approximately 1 . 7 million people , on average , sustain a traumatic brain injury annually . During the last few decades , brain neurotrauma biomechanics has been an active area of research involving medical clinicians and a broad range of scientists and engineers . In addition , advances and fast growth of human connectomics continues to reveal new insights into the damaged brain . With recent advances in computational methods and high performance computing , we see the need and the exciting possibility to merge brain neurotrauma biomechanics and human connectomics science to form a new area of investigation - connectome neurotrauma mechanics . For neurotrauma , the idea is simple - inform human structural connectome analysis using physics-based predictions of biomechanical brain injury . If successful , this technique may be further used to inform human functional connectome analysis , thus providing a new tool to help understand the pathophysiology of mild traumatic brain injury . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"tissue",
"mechanics",
"medicine",
"mechanics",
"classical",
"mechanics",
"dynamics",
"(mechanics)",
"neuroscience",
"biomechanics",
"biophysics",
"simulations",
"neuroimaging",
"biology",
"head",
"injury",
"biophysics",
"continuum",
"mechanics",
"connectomics",
"physics",
"neuroanatomy",
"biophysic",
"al",
"simulations",
"neurological",
"disorders",
"neurology",
"computational",
"biology"
]
| 2012 | Combining the Finite Element Method with Structural Connectome-based Analysis for Modeling Neurotrauma: Connectome Neurotrauma Mechanics |
The Caenorhabditis elegans left and right AWC olfactory neurons communicate to establish stochastic asymmetric identities , AWCON and AWCOFF , by inhibiting a calcium-mediated signaling pathway in the future AWCON cell . NSY-4/claudin-like protein and NSY-5/innexin gap junction protein are the two parallel signals that antagonize the calcium signaling pathway to induce the AWCON fate . However , it is not known how the calcium signaling pathway is downregulated by nsy-4 and nsy-5 in the AWCON cell . Here we identify a microRNA , mir-71 , that represses the TIR-1/Sarm1 adaptor protein in the calcium signaling pathway to promote the AWCON identity . Similar to tir-1 loss-of-function mutants , overexpression of mir-71 generates two AWCON neurons . tir-1 expression is downregulated through its 3′ UTR in AWCON , in which mir-71 is expressed at a higher level than in AWCOFF . In addition , mir-71 is sufficient to inhibit tir-1 expression in AWC through the mir-71 complementary site in the tir-1 3′ UTR . Our genetic studies suggest that mir-71 acts downstream of nsy-4 and nsy-5 to promote the AWCON identity in a cell autonomous manner . Furthermore , the stability of mature mir-71 is dependent on nsy-4 and nsy-5 . Together , these results provide insight into the mechanism by which nsy-4 and nsy-5 inhibit calcium signaling to establish stochastic asymmetric AWC differentiation .
Cell fate determination during development requires both the induction of cell type specific genes and the suppression of genes that promote an alternative cell fate [1]–[4] . For example , both inductive signaling , mediated by an EGFR-Ras-MAPK pathway , and lateral inhibition , mediated by LIN-12/Notch activity and microRNA ( miRNA ) , are required for six multipotential vulval precursor cells to adopt an invariant pattern of fates in C . elegans [5] . Notch signaling-mediated lateral inhibition also plays a crucial role in the neuronal/glial lineage decisions of neural stem cells; as well as the B/T , alphabeta/gammadelta , and CD4/CD8 lineage choices during lymphocyte development [6] , [7] . In the Drosophila eye , the kinase Warts and PH-domain containing Melted repress each other's transcription in a bistable feedback loop to regulate the two alternative R8 photoreceptor subtypes expressing Rhodopsin Rh5 or Rh6 [2] . In the C . elegans sensory system , two sets of transcription factors and miRNAs reciprocally repress each other to achieve and stabilize one of the two mutually exclusive ASEL and ASER taste neuronal fates [8]–[10] . Notch signaling acts upstream of the miRNA-controlled bistable feedback loop to regulate ASE asymmetry through a lineage-based mechanism in early embryos [11] . The C . elegans left and right sides of Amphid Wing Cell C ( AWC ) olfactory neurons specify asymmetric subtypes through a novel mechanism independent of the Notch pathway in late embryogenesis [12] . Like ASE neurons , the two AWC neurons are morphologically symmetrical but take on asymmetric fates , such that the AWCON neuron expresses the chemoreceptor gene str-2 and the contralateral AWCOFF neuron does not [12]–[14] . Asymmetric differentiation of AWC neurons allows the worm to discriminate between different odors [15] . In contrast to reproducible ASE asymmetry , AWC asymmetry is stochastic: 50% of animals express str-2 on the left and the other 50% express it on the right . Ablation of either AWC neuron causes the remaining AWC neuron to become AWCOFF , suggesting that AWCOFF is the default state and the induction of AWCON requires an interaction or competition between the AWC neurons [12] . The axons of the two AWC neurons form chemical synapses with each other; AWC asymmetry is established near the time of AWC synapse formation [16] , [17] . In addition , axon guidance mutants are defective in inducing the AWCON state . These results suggest that the synapses could mediate the AWC interaction for asymmetry [12] . nsy-4 , encoding a claudin-like tight junction protein , and nsy-5 , encoding an innexin gap junction protein , act in parallel to downregulate the calcium-mediated UNC-43 ( CaMKII ) /TIR-1 ( Sarm1 ) /NSY-1 ( MAPKKK ) signaling pathway in the future AWCON cell [18] , [19] . Both AWCs and non-AWC neurons in the NSY-5 gap junction dependent cell network communicate to participate in signaling that coordinates left-right AWC asymmetry . In addition , non-AWC neurons in the NSY-5 gap junction network are required for the feedback signal that ensures precise AWC asymmetry [18] . Once AWC asymmetry is established in late embryogenesis , both the AWCON and AWCOFF identities are maintained by cGMP signaling , dauer pheromone signaling , and transcriptional repressors [12] , [20] , [21] . unc-43 ( CaMKII ) , tir-1 ( Sarm1 ) , and nsy-1 ( MAPKKK ) are also implicated in the maintenance of AWC asymmetry in the first larval ( L1 ) stage [22] . Although multiple genes were identified to be involved in the establishment and the maintenance of AWC asymmetry ( for a review , see [23] ) , it is still unknown how the calcium-regulated signaling pathway is inhibited by nsy-4 and nsy-5 in the AWCON cell . The TIR-1/Sarm1 adaptor protein assembles a calcium-signaling complex , UNC-43 ( CaMKII ) /TIR-1/NSY-1 ( ASK1 MAPKKK ) , at AWC synapses to regulate the default AWCOFF identity [16] , thus downregulation of tir-1 expression may represent an efficient mechanism to inhibit calcium signaling in the cell becoming AWCON . In support of this idea , a prior large scale examination of potential miRNA targets indicated that tir-1 and unc-43 may be downregulated by this class of RNAs [24] . Here , we analyze the function of the miRNA mir-71 in stochastic AWC asymmetry by characterizing its role in downregulation of the calcium signaling pathway in the AWCON cell . We show that mir-71 acts downstream of nsy-4/claudin and nsy-5/innexin to promote AWCON in a cell autonomous manner through inhibiting tir-1 expression , in parallel with other processes . We also show that nsy-4 and nsy-5 are required for the stability of mature mir-71 . Our results suggest a mechanism for genetic control of AWC asymmetry by nsy-4 and nsy-5 through mir-71-mediated downregulation of calcium signaling .
The calcium-regulated UNC-43 ( CaMKII ) /TIR-1 ( Sarm1 ) /NSY-1 ( ASK1 MAPKKK ) signaling pathway suppresses expression of the AWCON gene str-2 in the default AWCOFF cell [12] , [16] , [25] , [26] . To establish AWC asymmetry , the calcium-mediated signaling pathway is suppressed in the future AWCON cell . miRNAs are small non-coding RNAs that are robust in mediating post-transcriptional and/or translational downregulation of target genes [27] . In C . elegans , miRNAs are processed from premature form into mature form by alg-1/alg-2 ( encoding the Argonaute proteins ) and dcr-1 ( encoding the ribonuclease III enzyme Dicer ) [28] . Gene expression profiling revealed increased levels of unc-43 and tir-1 in dcr-1 mutants [24] , suggesting that unc-43 and tir-1 may be downregulated by miRNAs . Thus , we hypothesized that miRNAs may play a role in downregulation of the UNC-43/TIR-1/NSY-1 signaling pathway in the cell becoming AWCON . To test this hypothesis , we took a computational approach to identify miRNAs predicted to target the 3′ UTRs of known genes , including unc-2 , unc-36 , egl-19 , unc-43 , tir-1 , nsy-1 , and sek-1 , in the AWC calcium signaling pathway . Only the miRNAs that fit the following criteria were selected for further analysis: 1 ) At least 6 nucleotides in the seed region ( position 1–7 or 2–8 at the 5′ end ) of a miRNA is perfectly matched to the target 3′ UTR; 2 ) The seed match between a miRNA and its target 3′ UTR is conserved between C . elegans and a closely related nematode species C . briggsae , since evolutionary conservation between C . elegans and C . briggsae genomes is useful in identifying functionally relevant DNA sequences such as regulatory regions [29] , [30]; and 3 ) A miRNA is predicted by both MicroCosm Targets ( formerly miRBase Targets; http://www . ebi . ac . uk/enright-srv/microcosm/htdocs/targets/v5/ ) [31]–[33] and TargetScan ( http://www . targetscan . org/worm_12/ ) [34] . Based on these criteria , we identified six potential miRNAs ( mir-71 , mir-72 , mir-74 , mir-228 , mir-248 , mir-255 ) predicted to target unc-2 , unc-43 , tir-1 , nsy-1 , and sek-1 ( Figure S1A ) . A subset of these identified miRNA-target pairs were also predicted by other miRNA target prediction programs , including PicTar ( http://pictar . mdc-berlin . de/ ) [35] and mirWIP ( http://146 . 189 . 76 . 171/query . php ) [36] . Since most miRNAs are not individually essential and have functional redundancy [37]–[40] , loss-of-function mutations in a single miRNA may not show a defect in AWC asymmetry . To circumvent potential problems that may be posed by functional redundancy , we took an overexpression approach to determine the role of these six miRNAs in AWC asymmetry . We generated transgenic strains overexpressing individual miRNAs in both AWCs using an odr-3 promoter , expressed strongly in AWC neuron pair and weakly in AWB neuron pair [41] . Wild-type animals have str-2p::GFP ( AWCON marker ) expression in only one of the two AWC neurons ( Figure 1A and 1E ) . Since loss-of-function mutations in the AWC calcium signaling genes ( unc-2 , unc-36 , unc-43 , tir-1 , nsy-1 , and sek-1 ) led to str-2p::GFP expression in both AWC neurons ( 2AWCON phenotype ) ( Figure 1B and 1E ) [12] , [16] , [25] , [26] , we proposed that overexpression of the miRNA downregulating one of these calcium signaling genes would also cause a 2AWCON phenotype . We found that mir-71 ( OE ) animals overexpressing mir-71 , predicted to target tir-1 and nsy-1 , had a strong 2AWCON phenotype ( Figure 1C , 1E , and Figure S1B ) . This result suggests that mir-71 may downregulate the expression of tir-1 and nsy-1 to control the AWCON fate and that mir-71 is sufficient to promote AWCON when overexpressed . However , overexpression of the other five miRNAs individually caused a mixed weak phenotype of 2AWCON and 2AWCOFF ( Figure S1B ) . Since the activity of the nsy-1 3′ UTR in AWC was independent of mir-71 ( OE ) ( Figure S2B ) , we focused on the investigation of the potential role of mir-71 in promoting AWCON through negatively regulating tir-1 expression . The genetic interaction between mir-71 and tir-1 was characterized by double mutants ( Figure 1E ) . tir-1 ( ky648 ) gain-of-function ( gf ) mutants had two AWCOFF neurons ( 2AWCOFF phenotype ) ( Figure 1D and 1E ) [22] . We found the tir-1 ( ky648gf ) 2AWCOFF phenotype was significantly reduced in the tir-1 ( ky648gf ) ; mir-71 ( OE ) double mutants ( p<0 . 001 ) ( Figure 1E ) . These results support the hypothesis that mir-71 downregulates tir-1 to control the AWCON fate . To further determine the requirement of mir-71 in AWC asymmetry , we analyzed str-2p::GFP expression in the mir-71 ( n4115 ) deletion null allele [40] . mir-71 ( n4115 ) mutants displayed wild-type AWC asymmetry ( Figure 1E ) , suggesting that mir-71 may function redundantly with other miRNAs or non-miRNA genes to regulate calcium signaling in AWC asymmetry . In addition to mir-71 , mir-248 was also predicted to target tir-1 by three programs ( Figure S1A ) . mir-71 and mir-248 have different predicted target sites in the tir-1 3′ UTR . Since mir-248 mutants are not available , we analyzed the effect of mir-248 overexpression on AWC asymmetry . Unlike the highly penetrant 2AWCON phenotype caused by mir-71 overexpression , mir-248 overexpression generated a mixed weak phenotype of 2AWCON and 2AWCOFF ( Figure S1B ) . To test whether mir-71 and mir-248 have a synergistic effect on AWC symmetry , we made transgenic animals overexpressing both mir-71 and mir-248 in AWCs . The 2AWCON phenotype was not significantly higher in mir-71 ( OE ) ; mir-248 ( OE ) animals than in mir-71 ( OE ) ( data not shown ) . These results suggest that mir-71 may not act redundantly with mir-248 to regulate tir-1 expression in AWC asymmetry . To knockdown mir-248 expression , we made an anti-mir248 transgene expressing short hairpin RNA ( shRNA ) , consisting of both sense and antisense sequences of mir-248 , in AWC . The anti-mir-248 transgene caused an AWC phenotype similar to mir-248 ( OE ) ( data not shown ) , suggesting that the effect of the anti-mir-248 transgene on AWC asymmetry is not through knockdown of mir-248 but mainly due to overexpression of sense mir-248 in the shRNA construct . Functional redundancy of miRNAs and other regulatory pathways has been suggested by a previous study in the Drosophila eye [42] . To overcome functional redundancy of mir-71 , we crossed mir-71 ( n4115 ) into sensitized backgrounds including tir-1 ( ky388 ) , nsy-4 ( ky616 ) , and unc-76 ( e911 ) mutants . tir-1 ( ky388 ) is a temperature-sensitive ( ts ) allele that caused a 2AWCON phenotype in 29% of animals at 15°C ( Figure 1E ) [16] . The 2AWCON phenotype of tir-1 ( ky388ts ) mutants was significantly suppressed by mir-71 ( n4115 ) , such that 20% of mir-71 ( n4115 ) ; tir-1 ( ky388ts ) double mutants had a 2AWCON phenotype ( p<0 . 05; Figure 1E ) . These results further support the hypothesis that mir-71 antagonizes the function of tir-1 in the calcium signaling pathway to promote the AWCON fate . mir-71 is located within a large intron of the F16A11 . 3a ( ppfr-1 ) gene , encoding a protein phosphatase 2A regulatory subunit ( Figure 1F ) . It is possible that the 181 bp deletion mutation within the intron of ppfr-1 in mir-71 ( n4115 ) mutants may affect ppfr-1 activity leading to suppression of the tir-1 ( ky388ts ) 2AWCON phenotype . To test this possibility , we analyzed AWC phenotypes in ppfr-1 ( tm2180 ) ; tir-1 ( ky388ts ) double mutants . ppfr-1 ( tm2180 ) has a 1027 bp deletion removing the first three exons and therefore is a potential null allele ( Figure 1F ) [43] . The 2AWCON phenotype of ppfr-1 ( tm2180 ) ; tir-1 ( ky388ts ) double mutants was not significantly different from that of tir-1 ( ky388ts ) single mutants ( Figure 1E ) . This result suggest that ppfr-1 is not required for AWC asymmetry and that suppression of the tir-1 ( ky388ts ) 2AWCON phenotype was most likely caused by loss of mir-71 activity in mir-71 ( n4115 ) mutants . The nsy-4 claudin-like gene and the unc-76 axon guidance pathway gene induce the AWCON state by inhibiting the downstream calcium-signaling pathway . Loss-of-function mutations in nsy-4 and unc-76 cause a partially penetrant 2AWCOFF phenotype ( Figure 1E ) [12] , [19] . mir-71 ( n4115 ) mutations significantly enhanced the 2AWCOFF phenotype of nsy-4 ( ky616 ) and unc-76 ( e911 ) mutants ( p<0 . 001 ) . On the other hand , the 2 AWCON phenotype of nsy-4 ( OE ) trasnsgenic animals overexpressing nsy-4 in AWCs was significantly suppressed in nsy-4 ( OE ) ; mir-71 ( n4115 ) double mutants ( p<0 . 001; Figure 1E ) . These results are consistent with a role of mir-71 function in promoting the AWCON fate , and suggest that mir-71 may act in parallel with other regulatory molecules to antagonize the calcium-regulated signaling pathway to generate the AWCON identity . The predicted mir-71 target site in the tir-1 3′ UTR is 96 bp downstream of the stop codon; the prediction is strongly supported by four different programs , including MicroCosm Targets , TargetScan , PicTar , and mirWIP ( Figure S1A ) . The nucleotides at position 1–8 in the seed region of mir-71 perfectly match the target site of the tir-1 3′ UTR; the seed match is conserved between C . elegans and C . briggsae ( Figure 2A ) . To determine whether mir-71 acts directly through the predicted binding site in the tir-1 3′ UTR , we made GFP sensor constructs with the AWC odr-3 promoter and different 3′ UTRs: wild-type tir-1 3′ UTR or the tir-1 3′ UTRmut with mutated mir-71 target site ( Figure 2B ) . Transgenic animals expressing each sensor construct were crossed to mir-71 ( OE ) animals . The GFP intensity of each sensor construct in an individual AWC neuron was normalized to the nucleus-localized TagRFP intensity of the transgene odr-3p::2Xnls-TagRFP::unc-54 3′ UTR in the same cell . The unc-54 3′ UTR does not contain any strongly predicted mir-71 sites . The normalized GFP intensity of each sensor construct was compared between mir-71 ( OE ) animals and their siblings losing the mir-71 ( OE ) transgene in the L1 stage , during which tir-1 is functional for the maintenance of AWC asymmetry [22] . We found that mir-71 ( OE ) animals , compared with wild type , had a significantly reduced normalized expression level of GFP from the tir-1 3′ UTR sensor construct ( p<0 . 005; Figure 2B upper panels ) . However , the normalized expression level of GFP from the tir-1 3′ UTRmut was not significantly different between wild-type and mir-71 ( OE ) animals ( Figure 2B bottom panels ) . These results suggest that mir-71 directly inhibits gene expression through the predicted target site in the tir-1 3′ UTR . However , we did not observe a significant difference in the GFP expression level from the tir-1 3′ UTR between wild-type animals and mir-71 ( n4115lf ) mutants ( Figure S2A ) . This result suggests potential functional redundancy of mir-71 in the regulation of tir-1 expression . Interactions between the 5′ and 3′ UTRs have been shown to regulate translation in mammalian cells [44] , bacteria [45] , and RNA viruses [46] . To determine if the tir-1 5′ UTR plays a role in regulating the inhibitory effect of mir-71 on the tir-1 3′ UTR , we included the tir-1 5′ UTR in the GFP sensor constructs ( Figure S3 ) . Similar to the tir-1 3′ UTR sensor constructs without the tir-1 5′ UTR ( Figure 2B ) , the normalized expression level of GFP from the tir-1 5′ UTR/tir-1 3′ UTR sensor construct was significantly decreased in mir-71 ( OE ) animals compared with wild type ( p<0 . 04; Figure S3A ) . However , the normalized expression level of GFP from the tir-1 5′ UTR/tir-1 3′ UTRmut sensor construct was not significantly different between wild-type and mir-71 ( OE ) animals ( Figure S3B ) . These results suggest that the tir-1 5′ UTR does not affect mir-71 ( OE ) -mediated suppression of gene expression through the tir-1 3′ UTR . The nsy-1 3′ UTR was also predicted to contain a mir-71 binding site by the four programs used in this study ( Figure S1A ) , but the GFP expression level from the nsy-1 3′ UTR was not significantly different between wild-type and mir-71 ( OE ) animals ( Figure S2B ) . This result suggests that the predicted mir-71 site in the nsy-1 3′ UTR may not be functional in AWC cells , therefore we did not further investigate the regulation of nsy-1 expression by mir-71 . tir-1 ( OE ) animals overexpressing tir-1 in AWC had a 2AWCOFF phenotype [16] . We used the tir-1 ( OE ) 2AWCOFF phenotype as readout to determine if mir-71 acts through the tir-1 3′ UTR to suppress the AWCOFF fate . We made tir-1 ( OE ) sensor constructs by replacing GFP in the GFP sensor constructs ( Figure 2B ) with tir-1 and crossed transgenic animals expressing each tir-1 ( OE ) sensor construct into mir-71 ( OE ) animals ( Figure 2C ) . The fold change in tir-1 ( OE ) 2AWCOFF phenotype was determined by dividing the 2AWCOFF percentage of tir-1 ( OE ) animals with the 2AWCOFF percentage of their tir-1 ( OE ) ; mir-71 ( OE ) siblings , which was then normalized to the relative tir-1 ( OE ) transgene copy number determined by qPCR . The higher normalized fold change in tir-1 ( OE ) 2AWCOFF indicates more suppression of 2AWCOFF phenotype by mir-71 ( OE ) in tir-1 ( OE ) ; mir-71 ( OE ) animals . The normalized fold change in tir-1 ( OE ) 2AWCOFF of tir-1 3′ UTR was significantly higher than that of the tir-1 3′ UTRmut ( p = 0 . 03; Figure 2C ) . These results suggest that mir-71 suppresses the AWCOFF fate by downregulating tir-1 expression through its 3′ UTR . To determine if mir-71 is expressed in AWC neurons , we generated transgenic animals expressing YFP under the control of a 2 . 4 kb promoter upstream of the mir-71 transcript ( Figure 1F ) . The expression of YFP was detected in several head neurons and the body wall muscle in L1 ( Figure 3A ) , which is consistent with previously reported expression pattern of mir-71 [47]–[49] . The mir-71p::YFP transgenic animals were crossed into an odr-1p::DsRed strain , expressing DsRed primarily in AWC and AWB neurons ( Figure 3B ) . YFP was coexpressed with DsRed in AWC and AWB neurons ( Figure 3C ) , suggesting that mir-71 is expressed in these neurons . We found that 52% of animals had visible mir-71p::YFP in both AWC cells , 28% had visible YFP in only AWC left ( AWCL ) , and 20% had visible YFP in only AWC right ( AWCR ) ( Figure 3D ) . These results suggest that the expression of mir-71 , when detected in one of the two AWC neurons , does not have a side bias towards AWCL or AWCR , which is consistent with stochastic choice of the AWCON fate . We then investigated whether mir-71 , when detected in both AWC neurons , has differential expression levels between AWCON and AWCOFF . Transgenic animals expressing mir-71p::GFP , ceh-36p::myr-TagRFP ( myristoylated TagRFP marker of AWCON and AWCOFF ) , and str-2p::2Xnls-TagRFP ( nucleus-localized TagRFP marker of AWCON ) were generated and analyzed in the L1 stage ( Figure 4A , 4A′ , 4A″ , 4B , 4B′ , and 4B″ ) . The ceh-36 promoter is expressed in AWCL , AWCR , ASEL , and ASER [50] , [51] . mir-71p::GFP expression was significantly higher in the AWCON cell than in the AWCOFF cell in 71% of the animals ( p<0 . 001; Figure 4C ) . To confirm this result , we generated transgenic animals expressing mir-71p::NZGFP , odr-3p::CZGFP , and str-2p::2Xnls-TagRFP in which reconstituted GFP ( recGFP ) [52] expression from two split GFP polypeptides , NZGFP and CZGFP , was restricted mainly in the two AWC cells . Consistent with the mir-71p::GFP result , recGFP expression was significantly higher in the AWCON cell than in the AWCOFF cell in 81% of the animals ( p<0 . 001; Figure S4 ) . Together , these results suggest that mir-71 is expressed at a higher level in the AWCON than in the AWCOFF cell . The higher expression of mir-71 in the AWCON cell is consistent with the role of mir-71 in promoting the AWCON fate . The suppression of gene expression by mir-71 through the tir-1 3′ UTR ( Figure 2B and 2C ) and the role of mir-71 in promoting the AWCON fate ( Figure 1C and 1E ) suggest that gene expression through the tir-1 3′ UTR may be downregulated in the AWCON cell . To investigate this possibility , transgenic animals expressing odr-3p::GFP::tir-1 3′ UTR ( GFP reporter of the tir-1 3′ UTR regulation in both AWCs ) , odr-3p::2Xnls-TagRFP::unc-54 3′ UTR ( nucleus-localized TagRFP marker of both AWCON and AWCOFF ) , and str-2p::myr-mCherry ( myristoylated mCherry marker of AWCON ) were generated and analyzed in the L1 stage ( Figure 4D , 4D′ , 4D″ , 4E , 4E′ , and 4E″ ) . The GFP intensity was normalized to the nucleus-localized TagRFP intensity measured in the same AWC cell to account for variation in focal plane and promoter activity . Normalized GFP intensity was significantly lower in the AWCON cell than in the AWCOFF cell in more than 85% of the animals ( p<0 . 02; Figure 4F ) . These results suggest that the expression of tir-1 is downregulated in the AWCON cell , consistent with a higher expression level of mir-71 in AWCON and downregulation of tir-1 expression by mir-71 . To determine the site of mir-71 action , mosaic animals in which the two AWC neurons have differential mir-71 activity were used to ask whether mir-71 acts in the future AWCON cell or the future AWCOFF cell . Mosaic animals were generated by random and spontaneous mitotic loss of an unstable transgene expressing the mir-71 ( OE ) construct odr-3p::mir-71 and a mosaic marker odr-1p::DsRed that showed which AWC cells retained the transgene . We specifically looked for the mosaic animals in which only one of the two AWC neurons expressed the mir-71 ( OE ) transgene; this cell was identified by expression of the DsRed marker . Mosaic analysis was first performed in transgenic lines expressing the mir-71 ( OE ) transgene in a wild-type background . Expression of the mir-71 ( OE ) transgene in both AWC neurons resulted in a 2AWCON phenotype ( Figure 5A and 5C ) . When the mir-71 ( OE ) transgene was retained in only one of the two AWC neurons , the mir-71 ( OE ) AWC neuron became AWCON and wild-type AWC neuron became AWCOFF in the majority of these mosaic animals ( p<0 . 0001; Figure 5B and 5D ) . This result is consistent with a significant cell-autonomous requirement for mir-71 in the AWCON cell to regulate its identity , which is opposite to the cell autonomous function of tir-1 in regulation of the AWCOFF identity . This result suggests that the AWC cell with higher mir-71 activity can prevent the contralateral AWC cell from becoming AWCON and that mir-71 may play a role in a negative-feedback signal sent from pre-AWCON to pre-AWCOFF . Similar results were obtained from previous mosaic analysis of nsy-4 and nsy-5 [18] , [19] . NSY-4 claudin-like protein and NSY-5 gap junction protein are the two parallel signaling systems that antagonize the calcium signaling pathway to specify the AWCON identity [18] , [19] . To determine whether mir-71 acts downstream of nsy-4 and nsy-5 to promote AWCON , mosaic analysis was performed with the mir-71 ( OE ) transgene in nsy-4 ( ky627 ) and nsy-5 ( ky634 ) mutants . Loss-of-function mutations in nsy-4 and nsy-5 caused a 2AWCOFF phenotype ( Figure 5E ) [18] , [19] , opposite to the mir-71 ( OE ) 2AWCON phenotype . Overexpression of mir-71 in both AWC neurons significantly suppressed the 2AWCOFF phenotype of nsy-4 ( ky627 ) and nsy-5 ( ky634 ) mutants . In addition , nsy-4 ( ky627 ) ; mir-71 ( OE ) and nsy-5 ( ky634 ) ; mir-71 ( OE ) animals resembled the mir-71 ( OE ) parent more closely than the nsy-4 ( ky627 ) or nsy-5 ( ky634 ) parent , but mixed phenotypes were observed ( Figure 5E ) . These results suggest that mir-71 mainly acts at a step downstream of nsy-4 and nsy-5 to promote AWCON . In the majority of the mosaic animals retaining the mir-71 ( OE ) transgene in only one of the two AWC neurons , the mir-71 ( OE ) AWC neuron expressed str-2p::GFP and the other AWC neuron did not ( Figure 5F ) . This significant cell-autonomous requirement for mir-71 in the future AWCON neuron in nsy-4 ( ky627 ) and nsy-5 ( ky634 ) mutants is the same as in the wild-type background . These results suggest that mir-71 acts cell autonomously downstream of nsy-4 and nsy-5 to promote the AWCON identity . alg-1 mutants had overaccumulation of premature mir-71 and underaccumulation of mature mir-71 , indicating that ALG-1/Argonaute-like protein is required for processing of mir-71 from premature form into the mature form [53] . alg-1 ( gk214 ) single mutants had wild-type str-2p::GFP expression . However , alg-1 ( gk214 ) significantly suppressed the 2AWCON phenotype of mir-71 ( OE ) and caused a weak 2AWCOFF phenotype in alg-1 ( gk214 ) ;mir-71 ( OE ) animals ( p<0 . 001; Figure 1E ) . In addition , alg-1 ( gk214 ) , like mir-71 ( n4115 ) mutants , also significantly suppressed the 2AWCON phenotype of tir-1 ( ky388ts ) mutants ( p<0 . 05; Figure 1E ) . These results suggest that alg-1 is required for mir-71 function in the AWCON cell . Consistent with previous northern blot analysis [53] , we found a significantly reduced level of mature mir-71 in alg-1 ( gk214 ) mutants ( p<0 . 001; Figure S5A ) using a stem-loop RT-PCR technique designed for specific quantification of mature miRNAs [54] . In addition , mature mir-71 was not detected in mir-71 ( n4115 ) mutants ( Figure S5B ) , suggesting that mir-71 ( n4115 ) is a null allele . Since mir-71 is expressed broadly in the animal [48] , [49] ( Figure 3A ) , we introduced the AWC-expressing transgene odr-3p::mir-71 in mir-71 ( n4115 ) mutants and used stem-loop RT-PCR to assay the level of mature mir-71 mainly in AWC cells ( Figure S5B ) . To determine if the maturation and/or the stability of mir-71 in AWCs is regulated by the signaling molecules that act upstream of tir-1 , we assayed the level of mature mir-71 in mir-71 ( n4115 ) ; nsy-4 ( ky627 ) double mutants , mir-71 ( n4115 ) ; nsy-5 ( ky634 ) double mutants , and mir-71 ( n4115 ) ; unc-36 ( e251 ) double mutants containing the AWC mir-71 ( OE ) transgene using stem-loop RT-PCR ( Text S1 ) . The level of mature mir-71 was significantly reduced in nsy-4 ( ky627 ) ( p = 0 . 015 ) and nsy-5 ( ky634 ) ( p<0 . 0001 ) mutants compared with control , but was not significantly different between control and unc-36 ( e251 ) mutants ( Figure S5B ) . The decreased level of mature mir-71 was not due to reduced transmission rates of the odr-3p::mir-71 transgene ( Figure S6A ) or downregulation of the odr-3 promoter in nsy-4 ( ky627 ) and nsy-5 ( ky634 ) mutants ( Figure S6B ) . These results suggest that nsy-4 and nsy-5 are required for the generation and/or the stability of mature mir-71 . To further determine whether nsy-4 and nsy-5 regulate the formation and/or the stability of mature mir-71 , we performed stem-loop RT-qPCR to quantify the level of premature and mature mir-71 in mir-71 ( n4115 ) mutants , mir-71 ( n4115 ) ; nsy-4 ( ky627 ) double mutants , and mir-71 ( n4115 ) ; nsy-5 ( ky634 ) double mutants containing the AWC mir-71 ( OE ) transgene . Consistent with stem-loop RT-PCR results ( Figure S5B ) , the abundance of mature mir-71 was significantly decreased in nsy-4 ( ky627 ) ( p<0 . 05 ) and nsy-5 ( ky634 ) ( p = 0 . 0003 ) mutants ( Figure 6 ) . However , the level of premature mir-71 was not significantly different between control and nsy-4 ( ky627 ) as well as nsy-5 ( ky634 ) mutants ( Figure 6 ) . These results suggest that the stability , but not the generation , of mature mir-71 is reduced in nsy-4 ( ky627 ) and nsy-5 ( ky634 ) mutants , and are consistent with a model in which nsy-4 and nsy-5 promotes the stability of mature mir-71 for downregulation of tir-1 in the future AWCON cell ( Figure 7 ) . mir-71 is expressed at a higher level in the AWCON cell than in the AWCOFF cell ( Figure 4A–4C ) , suggesting that mir-71 is differentially regulated at the transcriptional level in the two AWC cells . To determine if nsy-4 and nsy-5 also regulate differential expression levels of mir-71 between the two AWC cells , we crossed the transgene ( Figure 4A–4C ) containing mir-71p::GFP , ceh-36p::myr-TagRFP , and str-2p::2Xnls-TagRFP into nsy-4 ( ky627 ) and nsy-5 ( ky634 ) mutants . Since the AWCON marker str-2 is not expressed in nsy-4 ( ky627 ) or nsy-5 ( ky634 ) mutants , we analyzed and compared the expression levels of mir-71p::GFP between the two AWC cells in the mutants , instead of comparing the expression level between AWCON and AWCOFF ( Figure 4A–4C ) . We found that mir-71 was also differentially expressed between the two AWC cells in nsy-4 ( ky627 ) and nsy-5 ( ky634 ) mutants ( Figure S7 ) , like in wild-type animals . These results suggest that differential regulation of mir-71 transcription in the two AWC cells is not dependent on nsy-4 or nsy-5 .
Stochastic cell fate acquisition in the nervous system is a conserved but poorly understood phenomenon [1] . Here , we report that the miRNA mir-71 is part of the pathway that controls stochastic left-right asymmetric differentiation of the C . elegans AWC olfactory neurons through downregulating the expression of tir-1 , encoding the TIR-1/Sarm1 adaptor protein in a calcium signaling pathway . In addition , we have linked NSY-4/claudin- and NSY-5/innexin-dependent stability of mature mir-71 to downregulation of calcium signaling in stochastic AWC neuronal asymmetry . Previous studies have identified the role of miRNAs in reproducible , lineage-based asymmetry of the C . elegans ASE taste neuron pair , in which the miRNA expression pattern is largely fixed along the left-right axis [8] , [9] , [55] . This study provides one of the first insights into miRNA function in stochastic left-right asymmetric neuronal differentiation , in which the miRNA expression pattern is not fixed and is likely regulated by the stochastic signaling event driving random asymmetry . The seed match between mir-71 and the tir-1 3′ UTR is conserved between C . elegans and C . briggsae . However , the str-2 promoters share little sequence similarity between C . elegans and C . briggsae . The C . elegans str-2 promoter GFP reporter , when expressed in C . briggsae , does not show detectable GFP expression in AWC neurons in embryos , first stage larvae , or adults ( data not shown ) . This result suggests that the transcriptional regulation of str-2 has diverged in C . briggsae . mir-71 has been implicated in various cell biological and developmental processes including promotion of longevity , resistance to heat and oxidative stress , DNA damage response , control of developmental timing , dauer formation , and recovery from dauer [47] , [56]–[60] . However , it is largely unknown how mir-71 functions to regulate these biological processes . RNA interference ( RNAi ) of tir-1 did not affect C . elegans longevity [61] , suggesting that mir-71 may regulate distinct target genes for different functions . miRNAs are important post-transcriptional and translational regulators of gene expression during development and disease . Several miRNA target prediction algorithms such as MicroCosm Targets , TargetScan , PicTar , and mirWIP provide useful tools with which to identify potential target genes of miRNAs [62] . However , many miRNAs have redundant functions and therefore give subtle or no phenotypes when mutated [37]–[40] . Overexpression approach or phenotypic analysis of miRNA mutants in sensitized genetic backgrounds have been successful in elucidating the role of miRNAs for which null mutants are not available or functional redundancy is a potential problem [5] , [8] , [38]–[40] , [42] , [63]–[65] . Using miRNA target prediction programs , we identified mir-71 and five other miRNAs as potential regulators of the calcium-regulated UNC-43 ( CaMKII ) /TIR-1/NSY-1 ( MAPKKK ) signaling pathway . Through an overexpression approach and functional analysis of mir-71 ( n4115 ) mutants in sensitized genetic backgrounds , we revealed the role of mir-71 in genetic control of the AWCON identity . miRNAs that share the same sequence identity in their seed regions and could be potentially capable of downregulating the same set of target genes are grouped as members of a family [66]–[69] . Some miRNA family members have been shown to function redundantly and work together to regulate specific developmental processes [37] , [38] , [70]–[74] . However , many families of miRNAs did not show synthetic phenotypes , indicating that most miRNA families act redundantly with other miRNAs , miRNA families , or non-miRNA genes [38] . Since there is only one mir-71 family member identified , the absence of an AWC phenotype in mir-71 ( n4115 ) single mutants suggests that mir-71 may act redundantly with other miRNA family members or non-miRNA genes to regulate calcium signaling in AWC asymmetry . dcr-1 , encoding the ribonuclease III enzyme Dicer , is required for processing of premature miRNAs to mature miRNAs [28] . dcr-1 ( ok247 ) null mutants had wild-type AWC asymmetry ( data not shown ) . This result suggests that the dcr-1 mutation may cause simultaneous knockdown of several miRNAs ( including mir-71 ) with opposite functions in AWC asymmetry , thereby masking the role of mir-71 and its redundant miRNAs in AWC asymmetry . The UNC-76 axon guidance molecule and NSY-4 claudin-like protein act to antagonize the calcium-regulated signaling pathway to generate the AWCON identity [12] , [19] . We found that mir-71 ( n4115 ) mutants significantly suppressed the 2AWCON phenotype of nsy-4 ( OE ) and enhanced the 2AWCOFF phenotype of nsy-4 ( ky627 ) and unc-76 ( e911 ) mutants . These results suggest an alternative mechanism for functional redundancy of mir-71 in AWC asymmetry . mir-71 may act in parallel with other regulatory pathways downstream of unc-76 and nsy-4 to downregulate the calcium signaling pathway in the AWCON cell . Functional redundancy of miRNAs and other regulatory pathways has been demonstrated by a previous study suggesting that Drosophila miR-7 may act in parallel with a protein-turnover mechanism to downregulate the transcriptional repressor Yan in the fly eye [42] . Our results suggest that mir-71 is regulated at transcriptional and post-transcriptional levels in AWC . At the transcriptional level , mir-71 is expressed at a higher level in the AWCON cell than in the AWCOFF cell . This transcriptional bias of mir-71 is not dependent on NSY-4 claudin-like protein or NSY-5 innexin gap junction protein . The mechanisms that regulate differential expression of mir-71 in the two AWC cells are yet to be elucidated . At the post-transcriptional level , the stability of mature mir-71 is dependent on nsy-4 and nsy-5 . It is possible that nsy-4 and nsy-5 may antagonize the miRNA turnover pathway to increase the level of mature mir-71 . The C . elegans 5′→3′ exoribonuclease XRN-2 has been implicated in degradation of mature miRNAs released from Argonaute [75] . However , xrn-2 ( RNAi ) animals did not show AWC phenotypes ( data not shown ) , suggesting that the stability of mature mir-71 may be independent of xrn-2 . The TIR-1/Sarm1 adaptor protein assembles a calcium-regulated signaling complex at synaptic regions to regulate the default AWCOFF identity [16] . Downregulation of the TIR-1 adaptor protein by mir-71 and other parallel pathways may represent an efficient mechanism to inhibit calcium signaling in the cell becoming AWCON . Calcium signaling is one of the most common and conserved systems that control a wide range of processes including fertilization , embryonic pattern formation , cell proliferation , cell differentiation , learning and memory , and cell death during development and in adult life [76] . In addition , calcium signaling is implicated in left-right patterning in several tissues of different organisms [77] . It has been shown that negative regulation of calcium signaling by miRNAs is important for normal development and health [78]–[81] . In summary , our study and the studies from other labs demonstrate that downregulation of calcium signaling by miRNAs is one of the important mechanisms for cellular and developmental processes .
Wild-type strains were C . elegans variety Bristol , strain N2 . Worm strains were generated and maintained by standard methods [82] . Mutations and integrated transgenes used are as follows: kyIs140 [str-2p::GFP; lin-15 ( + ) ] I [12] , kyIs323 [str-2p::GFP; ofm-1p::GFP] II [22] , oyIs44 [odr-1p::DsRed] V [51] , kyIs136 [str-2p::GFP; lin-15 ( + ) ] X [12] , mir-71 ( n4115 ) I [40] , nsy-5 ( ky634 ) I [18] , ppfr-1 ( tm2180 ) unc-29 ( e1072 ) I ( gift from P . Mains , University of Calgary , Canada ) [43] , rol-6 ( e187 ) II , tir-1 ( ky388ts ) III [16] , tir-1 ( ky648gf ) III , tir-1 ( tm3036 ) III [22] , unc-36 ( e251 ) III , dcr-1 ( ok247 ) III; nsy-4 ( ky616 ) IV , nsy-4 ( ky627 ) IV [19] , unc-43 ( n498gf ) IV , eri-1 ( mg366 IV ) , unc-76 ( e911 ) V , lin-15b ( n744 ) X , and alg-1 ( gk214 ) X . Transgenes maintained as extrachromosomal arrays include kyEx1127 [odr-3p::nsy-4; myo-3p::DsRed] [18] , vyEx149 [odr-3p::mir-71 ( 25 ng/µl ) ; ofm-1p::DsRed ( 20 ng/µl ) ] , vyEx187 [mir-71p::YFP ( 50 ng/µl ) ; elt-2p::CFP ( 5 ng/µl ) ] , vyEx527 , 528 [odr-3p::mir-71 ( 50 ng/µl ) ; odr-1p::DsRed ( 12 ng/µl ) ; ofm-1p::DsRed ( 30 ng/µl ) ] , vyEx605 , 606 [odr-3p::GFP::tir-1 3′ UTR ( 7 . 5 ng/µl ) ; elt-2p::CFP ( 7 . 5 ng/µl ) ] , vyEx611 , 615 [odr-3p::GFP::unc-54 3′ UTR ( 7 . 5 ng/µl ) ; elt-2p::CFP ( 7 . 5 ng/µl ) ] , vyEx647 [odr-3p::GFP::nsy-1 3′ UTR ( 7 . 5 ng/µl ) ; elt-2p::CFP ( 7 . 5 ng/µl ) ] , vyEx649 , 651 [odr-3p::GFP::tir-1 3′ UTRmut ( 7 . 5 ng/µl ) ; elt-2p::CFP ( 7 . 5 ng/µl ) ] , vyEx835 , 836 , 838 [odr-3p::tir-1::tir-1 3′ UTRmut ( 70 ng/µl ) ; elt-2p::CFP ( 7 . 5 ng/µl ) ] , vyEx703 , 720 [odr-3p::tir-1::tir-1 3′ UTR ( 70 ng/µl ) ; elt-2p::CFP ( 7 . 5 ng/µl ) ] , vyEx905 , 907 [odr-3p::mir-74 ( 50 ng/µl ) ; ofm-1p::DsRed ( 30 ng/µl ) ] , vyEx914 , 917 [odr-3p::mir-248 ( 50 ng/µl ) ; ofm-1p::DsRed ( 30 ng/µl ) ] , vyEx915 , 918 [odr-3p::mir-72 ( 50 ng/µl ) ; ofm-1::DsRed ( 30 ng/µl ) ] , vyEx916 , 920 , 921 [odr-3p::mir-228 ( 50 ng/µl ) ; ofm-1p::DsRed ( 30 ng/µl ) ] , vyEx922 , 923 , 924 [odr-3p::mir-255 ( 50 ng/µl ) ; ofm-1::DsRed ( 30 ng/µl ) ] , vyEx927 , 931 [mir-71p::GFP ( 10 ng/µl ) ; ceh-36p::myr-TagRFP ( 5 ng/µl ) ; str-2p::2Xnls-TagRFP ( 25 ng/µl ) ; ofm-1p::DsRed ( 30 ng/µl ) ] , vyEx1316 , 1317 [mir-71p::NZGFP ( 30 ng/µl ) ; odr-3p::CZGFP ( 15 ng/µl ) ; str-2p::2Xnls-TagRFP ( 25 ng/µl ) ; ofm-1p::DsRed ( 30 ng/µl ) ) , vyEx1318 , 1319 [nsy-5p::mir-248IR ( 100 ng/µl ) ; odr-1p::DsRed ( 15 ng/µl ) ; ofm-1p::DsRed ( 30 ng/µl ) ] , vyEx1065 [str-2p::myr-mCherry ( 100 ng/µl ) ; ofm-1p::DsRed ( 30 ng/µl ) ] , vyEx1097 [odr-3p::2Xnls-TagRFP ( 40 ng/µl ) ; pRF4 ( rol-6 ( su1006 ) ( 50 ng/µl ) ] , vyEx1351 , 1352 [odr-3p::tir-1 5′ UTR::GFP::tir-1 3′ UTR ( 15 ng/µl ) ; odr 3p::TagRFP::unc-54 3′UTR ( 15 ng/µl ) ; elt-2p::CFP ( 7 . 5 ng/µl ) ] , and vyEx1353 , 1375 [odr-3p::tir-1 5′UTR::GFP::tir-1 3′ UTRmut ( 15 ng/µl ) ; odr 3p::TagRFP::unc-54 3′ UTR ( 15 ng/µl ) ; elt-2p::CFP ( 7 . 5 ng/µl ) ] . A 2476 bp PCR fragment of mir-71 promoter was subcloned to make mir-71p::YFP and mir-71p::GFP . mir-71p::NZGFP was made by replacing GFP in mir-71p::GFP with a NZGFP fragment from TU#710 ( Addgene ) [52] . odr-3p::CZGFP was made by cloning a CZGFP fragment from TU#711 ( Addgene ) [52] into an odr-3p vector . ceh-36p::myr-TagRFP , in which the 1852 bp ceh-36 promoter drives expression of myristoylated TagRFP , was generated by replacing TagRFP in ceh-36p::TagRFP [83] with myr-TagRFP . odr-3p::2Xnls-TagRFP was made by replacing the str-2 promoter in str-2p::2Xnls-TagRFP [22] with the odr-3 promoter [41] . str-2p::myr-mCherry was generated by replacing GFP in str-2p::GFP [12] with a myr-mCherry fragment . A 94 bp mir-71 PCR fragment was subcloned to make odr-3p::mir-71 . A 561 bp PCR fragment of the tir-1 3′ UTR , which represents the average length of the 3′ UTR in the majority of identified tir-1 cDNA clones such as yk1473h08 ( www . wormbase . org ) , was subcloned to make odr-3p::tir-1::tir-1 3′ UTR and odr-3p::GFP::tir-1 3′ UTR . miRNA target prediction algorithms including MicroCosm Targets , PicTar , and mirWIP use 300–590 bp of tir-1 3′ UTR for analysis . The predicted mir-71 binding site , TCTTTC , in the tir-1 3′ UTR was mutated into CAGGCA using QuikChange II XL Site-Directed Mutagenesis Kit ( Stratagene ) to make odr-3p::GFP::tir-1 3′ UTRmut . tir-1a splice form was used for all tir-1 constructs . odr-3p::tir-1 5′ UTR::GFP::tir-1 3′ UTR was made by cloning a 150 bp PCR fragment of tir-1 5′ UTR , amplified from wild-type embryo cDNA , into the odr-3p::GFP::tir-1 3′ UTR . odr-3p::tir-1 5′ UTR::GFP::tir-1 3′ UTRmut was made by replacing GFP::tir-1 3′ UTR in the odr-3p::tir-1 5′ UTR::GFP::tir-1 3′ UTR with GFP::tir-1 3′ UTRmut . To make shRNA anti-mir-248 ( mir-248IR ) , the sense and antisense oligos , each consisting of mir-248 sense ( 24 nt ) and antisense ( 24 nt ) sequences that flank a 12 nt linker ( loop ) sequence , were designed ( SBI System Biosciences ) and annealed ( IDT ) as described . This hairpin construct was subcloned to make nsy-5p::mir-248IR . To generate transgenic strains , DNA constructs were injected into the syncytial gonad of adult worms as previously described [84] . Z-stack images of transgenic animals expressing fluorescent markers were acquired using a Zeiss Axio Imager Z1 microscope equipped with a motorized focus drive and a Zeiss AxioCam MRm CCD digital camera . All animals of each set of experiments had the same exposure time for comparison of fluorescence intensity . The single focal plane with the brightest fluorescence in each AWC cell was selected from the acquired image stack and measured for fluorescence intensity . To measure fluorescence intensity , the outline spline tool in the Zeiss AxioVision Rel 4 . 7 image analysis software was used to draw around the AWC cell body ( Figure 2B; Figure 4A , 4B , 4D , 4E; Figure S4A , S4B; and Figure S6B ) or nucleus ( Figure 2B , Figure 4D′ and 4E′ ) from captured images . To measure fluorescence intensity in dim GFP-expressing cells ( Figure 4B and Figure S4B ) , the display contrast and brightness were adjusted to visualize and outline the cells . For each category of animals , images from a minimum of 10 animals were collected and analyzed . Mosaic analysis was performed as previously described [13] , [18] , [19] , [25] . Transgenic lines expressing the odr-3p::mir-71; odr-1p::DsRed transgene were passed for minimum of six generations before scoring for mosaic animals . The same transgenic lines were crossed into nsy-4 ( ky627 ) and nsy-5 ( ky634 ) mutants for the analysis . Three adult hermaphrodites from each tir-1 ( OE ) transgene line were collected in 25 µl of worm lysis buffer ( 50 mM KCl , 0 . 01% gelatin , 10 mM Tris-HCl pH 8 . 3 , 0 . 45% Tween 20 , 0 . 45% NP-40 , 2 . 5 mM MgCl2 , 100 µg/ml Proteinase K ) . Collected worms were then incubated at −80°C for minimum of one hour , 65°C for one hour , and 95°C for 15 minutes . 5 µl of the worm lysate was used for subsequent qPCR with Fast SYBR Green Master Mix ( Invitrogen ) . qPCR reactions were run in triplicate at 95°C for 3 minutes , followed by 45 cycles of 95°C for 30 seconds , 57°C for 30 seconds , and 72°C for 30 seconds on the CFX96 Real-Time PCR Detection System ( Bio-Rad ) . PCR product was scanned for fluorescent signal at the end of each cycle and the C ( T ) values were obtained using the CFX Manager Software ( Bio-Rad ) . The relative tir-1 ( OE ) transgene copy number was determined using the 2[−Delta Delta C ( T ) ] method as previously described [85] with the actin-related gene , arx-1 , as internal control . Stem-loop RT-qPCR was performed as described [54] to detect and quantify relative expression levels of premature and mature mir-71 . The odr-3p::mir-71 transgenes used in genetic mosaic analysis were crossed into various genetic backgrounds . Total RNA samples were isolated from first stage larvae using RNeasy Mini kit ( QIAGEN ) . Reverse transcription ( RT ) reactions were performed with 1 µg of total RNA , SuperScript III reverse transcriptase ( Invitrogen ) , and RT primer ( oligo d ( T ) 18 , premature mir-71 stem-loop RT primer , or mature mir-71 stem-loop RT primer ) . 1 µl of 1∶35 diluted reverse transcription product was used as template for subsequent qPCR reactions with Fast SYBR Green Master Mix ( Invitrogen ) . All PCR reactions were run in triplicate at 95°C for 3 minutes , followed by 45 cycles of 95°C for 30 seconds , 51°C for 30 seconds , and 72°C for 30 seconds on the CFX96 Real-Time PCR Detection System ( Bio-Rad ) . PCR product was scanned for fluorescent signal at the end of each cycle and the C ( T ) values were obtained using the CFX Manager Software ( Bio-Rad ) . The actin-related gene , arx-1 , was used as internal control to normalize variation between samples . Relative expression of premature and mature mir-71 was analyzed using the 2[−Delta Delta C ( T ) ] method as previously described [85] . Relative expression was set to one for mir-71 ( n4115 ) ; odr3p::mir-71 and was normalized accordingly for other samples . Student's t-test was used to calculate statistical significance . | Cell identity determination requires a competition between the induction of cell type–specific genes and the suppression of genes that promote an alternative cell type . In the nematode C . elegans , a specific sensory neuron pair communicates to establish stochastic asymmetric identities by inhibiting a calcium signaling pathway in the neuron that becomes an induced identity . However , it is not understood how cell–cell communication inhibits the calcium signaling pathway in the induced neuronal identity . In this study , we identify a microRNA that represses the expression of a key molecule in the calcium signaling pathway to promote the induced neuronal identity . Overexpression of the microRNA causes both neurons of the pair to become the induced identity , similar to the mutants that lose function in the calcium signaling pathway . In addition , the stability of the mature microRNA is dependent on a claudin-like protein and a gap junction protein , the two parallel signals that mediate communication of the neuron pair to promote the induced neuronal identity . Our results provide insight into the mechanism by which cell–cell communication inhibits calcium signaling to establish stochastic asymmetric neuronal differentiation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"animal",
"models",
"developmental",
"biology",
"caenorhabditis",
"elegans",
"developmental",
"neuroscience",
"model",
"organisms",
"biology",
"neuroscience",
"cell",
"fate",
"determination"
]
| 2012 | The MicroRNA mir-71 Inhibits Calcium Signaling by Targeting the TIR-1/Sarm1 Adaptor Protein to Control Stochastic L/R Neuronal Asymmetry in C. elegans |
The grey mould fungus Botrytis cinerea causes losses of commercially important fruits , vegetables and ornamentals worldwide . Fungicide treatments are effective for disease control , but bear the risk of resistance development . The major resistance mechanism in fungi is target protein modification resulting in reduced drug binding . Multiple drug resistance ( MDR ) caused by increased efflux activity is common in human pathogenic microbes , but rarely described for plant pathogens . Annual monitoring for fungicide resistance in field isolates from fungicide-treated vineyards in France and Germany revealed a rapidly increasing appearance of B . cinerea field populations with three distinct MDR phenotypes . All MDR strains showed increased fungicide efflux activity and overexpression of efflux transporter genes . Similar to clinical MDR isolates of Candida yeasts that are due to transcription factor mutations , all MDR1 strains were shown to harbor activating mutations in a transcription factor ( Mrr1 ) that controls the gene encoding ABC transporter AtrB . MDR2 strains had undergone a unique rearrangement in the promoter region of the major facilitator superfamily transporter gene mfsM2 , induced by insertion of a retrotransposon-derived sequence . MDR2 strains carrying the same rearranged mfsM2 allele have probably migrated from French to German wine-growing regions . The roles of atrB , mrr1 and mfsM2 were proven by the phenotypes of knock-out and overexpression mutants . As confirmed by sexual crosses , combinations of mrr1 and mfsM2 mutations lead to MDR3 strains with higher broad-spectrum resistance . An MDR3 strain was shown in field experiments to be selected against sensitive strains by fungicide treatments . Our data document for the first time the rising prevalence , spread and molecular basis of MDR populations in a major plant pathogen in agricultural environments . These populations will increase the risk of grey mould rot and hamper the effectiveness of current strategies for fungicide resistance management .
Synthetic fungicides are used worldwide that provide protection of major crops from destruction by fungal plant pathogens [1] . As a result of repeated fungicide treatments , however , resistant strains of the pathogens are being selected [2] . Different resistance mechanisms have been reported that reduce fungicide effectiveness in field and greenhouse environments . Mutations leading to changes in the target proteins that are still functional but less sensitive to the drugs are most common in plant pathogenic fungi . For example , rapid accumulation of mutations in the gene encoding ß-tubulin have been observed in a variety of plant pathogens after introduction of the benzimidazole fungicides , leading to resistance against these fungicides [3] . Other mechanisms , such as overexpression of the gene encoding the target site or increased fungicide metabolism have also been described [3]–[6] . Multidrug resistance ( MDR ) , an important resistance mechanism in human pathogenic microbes and cancer cells , has often been correlated with the activity of energy dependent plasma membrane efflux transporters with low substrate specificity . Mutations leading to overexpression of individual transporters can result in increased export and thereby reduced sensitivity to a variety of drug molecules [7]–[9] . In fungi , the major types of drug efflux proteins are ATP binding cassette ( ABC ) and major facilitator superfamily ( MFS ) transporters [10]–[12] . Constitutive overexpression of the ABC transporters CDR1 and CDR2 , or the MFS transporter MDR1 has been observed in Candida spp . with MDR phenotypes that have been selected by prolonged fluconazole treatments in humans [13] . In filamentous fungi , the role of ABC and MFS transporters in the efflux of natural and synthetic toxicants is well known [14] . For example , analysis of knock-out mutants revealed that the ABC transporter AtrB from Aspergillus nidulans and its orthologue AtrB from Botrytis cinerea transport a wide variety of fungicides as well as toxins of plant and microbial origins [14]–[16] . Similarly , the MFS transporter Mfs1 of B . cinerea was found to mediate efflux of several fungicides as well as plant derived and microbial toxins [17] . Several efflux transporter genes have been shown to be rapidly induced by fungicides or natural toxins , such as Mycosphaerella graminicola Atr1 and Atr2 , or B . cinerea atrD and atrB [16] , [18] . Some ABC transporters have been shown to be involved in plant pathogenesis [19]–[22] . This is probably mainly due to the export of plant defence compounds , for example the Arabidopsis phytoalexin camalexin in the case of B . cinerea AtrB [22] . For the ABC1 transporter of Magnaporthe grisea , evidence was provided that it is required for tolerance to oxidative stress during appressorial penetration [19] . Despite some reports of MDR phenotypes in laboratory mutants of B . cinerea [16] , [18] and field strains of Penicillium digitatum and Mycosphaerella graminicola [23] , [24] , a significant role of MDR in agricultural environments has not yet been described for plant pathogens . However , a long term monitoring for fungicide resistance of B . cinerea initiated in French wine-growing regions has revealed , in addition to drug specific resistance mechanisms , the appearance of strains with cross resistance to chemically unrelated fungicides in the Champagne [4] , [25] , [26] . In this report , we have investigated these strains in further detail and describe the increasing prevalence of three different MDR populations in commercial vineyards . We show that their phenotypes are caused by mutations leading to overexpression of efflux transporters , and present evidence for long-distance migration of MDR strains from France to German wine-growing regions .
In 1994 , strains with two different MDR phenotypes , formerly designated AniR2 ( here MDR1 ) and AniR3 ( here MDR2 ) , because of their reduced sensitivity to anilinopyrimidine fungicides , have been identified for the first time in the Champagne ( Fig . 1A ) [26] . A third MDR phenotype ( MDR3 ) was first detected in 2001 . Since then , the frequency of MDR strains in the Champagne steadily increased until 2008 , when the three MDR phenotypes together represented 55% of the total population ( Fig . 1A ) . In vineyards of the German Wine Road region , a similar survey of B . cinerea isolates for fungicide sensitivity was performed for three years . Between 2006 and 2008 , increasing MDR populations were also observed , but in contrast to the Champagne the MDR1 phenotype was clearly dominating ( Fig . 1B ) . As previously described [26] , MDR1 and MDR2 strains had overlapping but distinct profiles of increased tolerance to a number of different classes of fungicides and other drugs ( Table 1 ) . Although the levels of tolerance observed were not as high as specific resistance mechanisms ( e . g . target site mutations ) , they were clearly genetically based and heritable and therefore called resistance throughout this paper . While MDR1 strains showed considerable resistance levels mainly towards fludioxonil , cyprodinil and tolnaftate , MDR2 strains were characterized by increased resistance to fenhexamid , tolnaftate , cycloheximide and cyprodinil . MDR3 strains showed the highest levels and broadest spectrum of resistance against most fungicides tested ( Table 1 ) . MDR phenotypes in fungi are usually correlated with increased drug efflux [8] . When the B . cinerea MDR strains were tested , they showed indeed lower fungicide accumulation than sensitive strains , indicating increased efflux activity . As previously reported [16] , sensitive strains show a transient accumulation of 14C-fludioxonil , followed by efflux of the drug after approximately 30 min , due to activation of efflux transporters ( Fig . 2A ) . In contrast , two MDR1 strains showed only low initial fludioxonil accumulation . After addition of the uncoupler CCCP , rapid influx was observed for all strains indicating the presence of energy-dependent efflux systems ( Fig . 2A ) [16] . Similarly , MDR3 strains accumulated little fludioxonil , while MDR2 strains behaved similar to sensitive strains ( Fig . 2B ) . These data indicated the presence of a constitutive efflux system in MDR1 and MDR3 strains . With 14C-bitertanol , all MDR strains showed reduced initial accumulation levels compared to sensitive strains , although this effect was less pronounced in MDR1 than in MDR2 and MDR3 strains ( Fig . 2B ) . In accordance with these data , lower accumulation of tebuconazole and triadimenol by a B . cinerea strain with MDR2 phenotype has been described previously [25] . The specificity of the uptake experiments was confirmed by experiments with heat-inactivated germlings , which displayed very high , non-transient fungicide accumulation ( Fig . 2C ) . The phenotype of MDR1 strains , including the low initial fludioxonil accumulation , was similar to that of a B . cinerea laboratory mutant which showed overexpression of the ABC transporter AtrB [16] . AtrB and its orthologs in other filamentous fungi , including Aspergillus nidulans and Penicillium digitatum , is a conserved efflux pump that contributes tolerance to various fungicides and natural antifungal compounds [14]–[16] , [21] , [22] , [27] , [28] . Indeed , atrB was constitutively upregulated in MDR1 and MDR3 strains , but not in MDR2 strains , showing 50–150 fold overexpression relative to sensitive strains . As previously reported , high levels of atrB expression were observed in sensitive strains after 30 min treatment with fludioxonil ( Fig . 3 ) [16] . Two other genes encoding ABC transporters , atrK and BMR3 , were also upregulated in the absence of drug induction in MDR1 and MDR3 strains when compared to sensitive strains , but only 2 . 5–5 fold ( Fig . 3B ) . To identify efflux transporters that are specifically upregulated in MDR2 strains , microarray hybridizations with B . cinerea whole genome chips were performed ( data not shown ) . These experiments revealed that mfsM2 ( Major facilitator superfamily transporter involved in MDR2 ) , which showed very weak expression in sensitive and MDR1 strains , was more than 600 fold overexpressed in MDR2 and MDR3 strains ( Fig . 3 ) . To confirm a causal relationship between overexpression of efflux transporters and MDR phenotypes , atrB and mfsM2 mutants were generated . As described further below , MDR1 strains with atrB deletions had lost the MDR phenotype . Two MDR2 strains with mfsM2 deletions had lost increased efflux activity for 14C-bitertanol ( Fig . 4A ) . Furthermore , the mfsM2 deletion mutants had lost the reduced sensitivity to various fungicides , showing levels similar to sensitive strains ( Fig . 4B , C ) . In contrast , when a sensitive strain was transformed with a construct providing constitutive overexpression of mfsM2 ( mfsM2ox ) , which led to 1481 ( ±309 ) -fold upregulation of mfsM2 relative to the parent strain , it acquired drug resistance levels similar to MDR2 strains ( Fig . 4B , C ) . Overexpression of atrB and mfsM2 is therefore necessary and probably sufficient to generate MDR1 and MDR2 phenotypes , respectively , in B . cinerea field strains . Apart from its role in MDR , we found only slight growth differences of the mfsM2 deletion mutants . When tested for pathogenicity , the mfsM2 deletion mutants showed no significant differences compared to their parent strains ( data not shown ) . Mutations leading to changes in gene expression are often located either in the promoters of these genes , or in regulatory genes . Sequencing of the atrB promoter regions from several sensitive , MDR1 and MDR3 strains did not reveal any MDR1-specific mutations ( not shown ) . Since other ABC-transporter genes besides atrB were also found to be upregulated in MDR1 and MDR3 strains ( Fig . 3B ) , we assumed that the MDR1 phenotype might have been generated by mutations in a regulatory gene . In order to locate the suspected MDR1-specific regulator gene , a map-based cloning approach was performed . When F1 progeny isolates of several crosses with MDR strains were analyzed , the segregation data confirmed that MDR1 and MDR2 phenotypes are determined by just one genetic locus each , and that strains with MDR3 phenotype can originate from recombination between MDR1 and MDR2 strains ( Table S1 ) [26] . By identifying polymorphic molecular markers that cosegregate with MDR1 and MDR3 phenotypes in the F1 progeny , it was possible to localize and identify mrr1 ( multidrug resistance regulator 1 ) encoding a putative Zn ( II ) 2Cys6 zinc cluster transcription factor ( TF; Table S2 ) [29] . The genetic marker closest to mrr1 , BC63-17 , located 1 . 8 kb away from mrr1 , showed 100 percent cosegregation with MDR1 phenotypes . Similarly , using crosses with MDR2 strains , a marker located just 1 . 6 kb away from the efflux transporter gene mfsM2 was found completely cosegregate with MDR2 phenotypes , indicating that they are caused by mutations in mfsM2 ( Table S3 ) . Sequencing of mrr1 from eight sensitive field strains revealed no or only silent nucleotide changes when compared to mrr1 of the sequenced reference strains T4 and B05 . 10 ( data not shown ) . In contrast , all MDR1 ( n = 15 ) and MDR3 ( n = 5 ) strains analyzed showed at least one point mutation leading to amino acid changes in Mrr1 . In total , at least eight different MDR1-related mutations were identified ( Fig . 5; Table S4 ) . Their role in generation of the MDR1 phenotype was supported by the observation that five of these mutations had occurred in more than one MDR1 or MDR3 strain . To confirm that mrr1 encodes the TF responsible for MDR1-related atrB overexpression , mrr1 and atrB deletions were generated in MDR1 strains and a sensitive strain . Consistent with the expected phenotype of a regulatory mutant , mrr1 mutants showed very low levels of atrB expression , not inducible by fludioxonil . Furthermore , they showed reduced expression of the ABC transporter genes BMR3 and atrK that are also upregulated in MDR1 strains ( Fig . 6A ) . Both the mrr1 and atrB mutants of the MDR1 strain showed increased fludioxonil uptake , indicating loss of AtrB-mediated efflux activity ( Fig . 6B ) . With regard to their drug sensitivity , the mrr1 and atrB mutants of MDR1 strains had completely lost their MDR1 phenotypes and were slightly hypersensitive to fludioxonil , similar to previously described atrB mutants ( Fig . 6C , D ) [16] . The MDR1 strain D06 . 7-27 showed an unusually high cyprodinil resistance and a rather low tolnaftate resistance , compared to other MDR1 strains , for unknown reasons . In the D06 . 7-27 ( Δmrr1 ) mutant , the cyprodinil resistance was significantly reduced , but still higher than in sensitive strains ( Fig . 6C ) . To confirm that single mrr1 mutations are sufficient for generation of the MDR1 phenotype , a sensitive strain was transformed with the mrr1V575M allele from MDR1 strain D06 . 7-27 . The transformants , expressing both wild type mrr1 and mrr1V575M , showed constitutive upregulation of atrB and , to a lower extent , atrK and BMR3 ( Fig . 6A ) , as well as a drug resistance phenotype similar to MDR1 strains ( Fig . 6C , D ) . The atrB and mrr1 mutants showed only little changes in sensitivity to fenhexamid , a substrate for MfsM2 but not AtrB , which confirmed that the functions of AtrB and Mrr1 are fungicide specific . These data confirmed that Mrr1 is the main transcriptional activator of atrB , and that activating mutations of mrr1 lead to overexpression of atrB and thus to MDR1 phenotypes . There are interesting parallels to the yeasts S . cerevisiae and C . albicans , in which MDR-related efflux transporter genes are regulated by Zn ( II ) 2Cys6 TFs as well . C . albicans MDR strains selected by fluconazole treatments in humans also resulted from gain-of-function TF mutations , leading to overexpression of ABC- and MFS-type MDR transporters [13] , [30] . Focusing on the search for mutations that are responsible for mfsM2 upregulation in MDR2 and MDR3 strains , we found a rearrangement in the mfsM2 upstream region , caused by insertion of a foreign gene fragment and concurrent deletion of a portion of the putative mfsM2 promoter . The inserted DNA , 1326 bp in length , is probably derived from an as yet unknown fungal long-terminal-repeat ( LTR ) retrotransposon ( Fig . 7A ) [31] . Surprisingly , this sequence is not present in the published genome sequences of the two B . cinerea strains B05 . 10 and T4 . It encodes truncated portions of a putative enzyme with domains of reverse transcriptase and RNase H , with closest homologs to sequences in the REAL retrotransposon of the plant pathogenic fungus Alternaria alternata [32] and in the Boty retroelement of B . cinerea ( Fig . S1 ) [33] . Out of 17 MDR2 and MDR3 strains analyzed from the Champagne ( 9 strains ) and from the German Wine Road ( 8 strains ) by sequencing or PCR analysis , all revealed the identical mfsM2 promoter rearrangement . In contrast , in 8 sensitive and 15 MDR1 strains no such rearrangement was found ( Table S5; data not shown ) . This observation strongly suggests that the mfsM2 alleles of these MDR2 and MDR3 strains have a common progenitor . That the promoter rearrangement is responsible for mfsM2 overexpression , was supported by the already described MDR2-like phenotype of a sensitive strain transformed with an mfsM2 overexpression construct . This was further confirmed by creating B . cinerea strains expressing mfsM2::uidA reporter gene fusion constructs . Only with the mfsM2 promoter fragment from an MDR2 strain , but not with a fragment from a sensitive strain , strong expression of ß-glucuronidase was observed ( Fig . 7B ) . The rapidly increasing MDR populations in French and German wine-growing regions indicate that strong selection for MDR phenotypes occurs by fungicide treatments . This was confirmed by two field experiments , in which mixtures of an MDR3 strain and a sensitive strain were introduced into two vineyards . A single treatment with a commercial fungicide mixture ( fludioxonil and cyprodinil ) during early berry development led to a significantly increased recovery of the MDR3 strain relative to the sensitive strain during grape harvest ( Fig . 8 ) . The recovery rates of the introduced MDR3 strain were 16% ( untreated vineyard ) and 62% ( treated vineyard ) in 2007 , and 34% ( untreated ) and 55% ( treated ) in 2008 , while the recovery rates of the introduced sensitive strain were 26% ( untreated ) and 14% ( treated ) in 2007 , and 28% ( untreated ) and 15% ( treated ) in 2008 . In addition , the MDR3 strain showed high survival rates after the following winter periods in the absence of fungicide treatments ( 40% in spring 2008 , 62% in spring 2009 ) . These data indicate also that the mutations in mrr1 and mfsM2 do not impair the fitness of MDR strains to a major extent .
Since the introduction of modern fungicides with specific modes of action , resistance development in fungal field populations has been observed , notably in B . cinerea which is considered to be a high risk pathogen [2] . Therefore , rules for fungicide resistance management have been established that include a recommendation to avoid the repetitive use of fungicides with similar targets within one growing season [2] , [5] . For Botrytis control in commercial European vineyards , two or three treatments with different mode-of-action fungicides are common . The increasing and widespread prevalence of MDR strains indicates that they have been selected under these conditions within the last decade . The three MDR phenotypes found in French and German vineyards were clearly correlated with increased drug efflux activity . Increased fludioxonil efflux was observed for MDR1 and MDR3 strains but not for MDR2 strains , while bitertanol efflux was observed for all MDR phenotypes , although more weakly for MDR1 . Furthermore , all MDR strains showed strong constitutive overexpression of one ( MDR1 , MDR2 ) or two ( MDR3 ) drug efflux transporter genes . In addition , weak overexpression of two other ABC transporter genes was observed in MDR1 and MDR3 strains . A causal correlation of MDR1 phenotypes with overexpression of the ABC transporter atrB , and of MDR2 phenotypes with MFS transporter mfsM2 overexpression was confirmed by the analysis of deletion and overexpression mutants . Two MDR1 strains with an atrB mutation had lost the MDR1 phenotype , and two MDR2 strains with an mfsM2 mutation had lost the MDR2 phenotypes . In addition , a sensitive strain which artificially overexpressed mfsM2 showed an MDR2-like phenotype . Thus , overexpression of atrB and mfsM2 are likely to be sufficient for the observed MDR1 and MDR2 phenotypes in B . cinerea field strains . While the role of atrB in the export of multiple natural and synthetic toxicants has been described in detail [14] , [28] , the function of mfsM2 in sensitive strains remains unknown . The protein is not highly conserved in other fungi , and even in the genome sequence of the closely related Sclerotinia sclerotiorum , no apparent orthologue to B . cinerea mfsM2 could be identified . In sensitive strains mfsM2 expression is very low , and up to now we did not find any fungicide or other compounds that induce mfsM2 ( data not shown ) . In MDR1 strains , the mutations leading to atrB overexpression were found to be located not in atrB itself , but in the transcription factor gene mrr1 . Out of 20 MDR1 and MDR3 strains analyzed , all carried mutations in the coding region of mrr1 , and several strains with different geographical origin or collected in different years showed identical mutations . Because mrr1 mutants failed to express atrB to significant levels , and because sensitive strains expressing an activated version of Mrr1 ( Mrr1V575M ) showed an MDR1-like phenotype , Mrr1 was confirmed to be a transcriptional activator of atrB . The mrr1 mutants also showed reduced expression of atrK and BMR3 , when compared to MDR1 strains , indicating that Mrr1 also plays a role in activation of other efflux transporter genes . However , since the drug sensitivity phenotypes of MDR1 strains with either atrB or mrr1 mutations were indistinguishable from each other , and since MDR1 atrB mutants have completely lost their MDR1 phenotypes , the weak overexpression of atrK and BMR3 does not seem to contribute significantly to the phenotype of MDR1 strains . While our data indicate that the main physiological role of Mrr1 is regulation of atrB , the whole set of genes controlled by Mrr1 in the genome of B . cinerea remains to be determined . Interestingly , while structurally and also functionally conserved orthologues of B . cinerea AtrB occur in other ascomycetous fungi , e . g . in A . nidulans , P . digitatum [15] , [22] , no clear orthologues ( best bidirectional hits ) of B . cinerea Mrr1 could be identified in other fungi . This indicates that the regulation of ABC transporters in filamentous fungi might be not highly conserved . Similar observations have been made for yeasts , because the major efflux transporter in S . cerevisiae , PDR5 , and its orthologue pair CDR1/CDR2 in C . albicans , are under control of different transcription factors [34] . With regard to fungicide resistance , efflux activities , efflux transporter gene overexpression , and genetic data , MDR3 strains were clearly identified as recombinants carrying both MDR1-specific mutations in mrr1 and MDR2-specific mutations in mfsM2 . Collectively , all data indicate that the three MDR phenotypes in B . cinerea have originated by mutations in just two genes . Obviously , the different mrr1 point mutations leading to MDR1 have occurred repeatedly . Thus MDR1 phenotypes could appear ( and might have already appeared ) in different agricultural environments in which selective conditions for these phenotypes prevail . Similarly , a variety of gain-of-function mutations have been found in transcription factor genes TAC1 and MRR1 of clinical MDR isolates of C . albicans , leading to overexpression of the efflux transporter genes MDR1 and CDR1 or CDR2 , respectively [13] , [30] , [35] . In contrast , the rearrangement in the mfsM2 promoter appeared to be a unique event , found in all MDR2 and MDR3 strains analyzed so far . Based on the time course of appearance of these strains in the Champagne , and their lower frequency compared to MDR1 strains in Germany , we assume that the rearrangement in mfsM2 originated once in the Champagne , possibly in the early 1990s . The rearranged mfsM2 allele later spread into the German Wine Road region , 250 km east of the Champagne , possibly by air currents . Because of the small size of the retrotransposon-derived gene fragment and of the lack of any remaining LTR sequences , the origin of the sequence inserted into mfsM2 and the mechanism of the insertion-deletion rearrangement remain obscure . The absence of the integrated sequence in the published genomes of strains B05 . 10 and T4 indicates that it occurs only in subpopulations of B . cinerea . A search for the presence of the sequence in a variety of field strains by using PCR , Southern hybridization and sequencing revealed that similar but non-identical sequences are present in some sensitive strains ( data not shown ) . The mfsM2 mutation is reminiscent of a transposable element insertion into the promoter of a gene for a cytochrome P450 monooxygenase involved in insecticide detoxification and resistance in Drosophila , leading to overexpression and global spread of the mutated gene [36] . A summarizing model of the data in this paper is shown in Fig . 9 . It is assumed that MDR strains in the Champagne have appeared due to selection pressure in fungicide treated vineyards , and to mutations leading to the appearance of MDR1 and MDR2 strains , and a few years later also to the appearance of MDR3 strains . Because a repeated occurrence of the unusual rearrangement found in the mfsM2 promoter appears to be highly unlikely , we assume that MDR2 strains carrying this rearrangement have migrated from France to Germany , probably in the last decade . To support this hypothesis , population genetic studies with MDR strains from different geographical origins are currently performed . In addition , we are searching for MDR strains in other regions with different crop cultures and different fungicide treatment schedules , in order to achieve a better understanding of the distribution of MDR strains , and the factors leading to their appearance and selection . A field experiment has clearly demonstrated selection of an artificially introduced MDR3 strain by a standard fungicide treatment . This confirms that the MDR3 phenotype confers selective advantage to B . cinerea in fungicide-treated vineyards , and that this advantage outweighs possible fitness defects . Furthermore , the MDR3 strain was recovered with high albeit varying frequencies after overwintering periods , in the absence of fungicide selection pressure . These data indicate that the general fitness of strains showing atrB and/or mfsM2 overexpression can be rather high in field environments , but this needs further studies . We are currently testing various fitness parameters in isogenic atrB , mrr1 and mfsM2 knock-out and overexpression strains in order to estimate the performance of these strains . Because the natural role of MDR-related efflux transporters seems to be the protection against various biotic toxic compounds [14] , [26] , [28] , it is possible that the MDR strains have acquired properties that increase their fitness in natural environments even in the absence of fungicides . Our work is the first documented case of a massive appearance of MDR populations in a major plant pathogen in fungicide-treated agricultural environments . To what extent the effectiveness of fungicide treatments against MDR strains is reduced in comparison to sensitive strains needs to be investigated . Nevertheless , the possibilities of a further rise of MDR3 strains and of additional mutations leading to higher levels of broad-spectrum fungicide resistance are expected to be a major threat for chemical control of grey mould disease in the near future .
Strains were isolated from commercial vineyards in the Champagne and the German Wine Road ( Palatinate ) . In the Champagne , samples were collected from vineyards located around Moulins , Hautvillers , Vandières and Courteron , Moulins being the northernmost ( 49°34′N/03°28′E ) and Courteron the southernmost ( 48°01′N/04°26′E ) town , 167 km apart from each other . Samples were collected from approximately 200 locations each year . Each sample represented a bulk population consisting of spores of at least 20 infected berries within the chosen plot . The spores from each bulk sample were spread onto agar media containing different fungicides , and analyzed for different phenotypes as described [25] , [37] . In the German Palatinate , the same six vineyards were used for sampling each year . The plots are located along the German Wine Road , between Dackenheim ( northernmost: 49°53′N/8°19′E ) and Walsheim ( southernmost: 49°23′N/8°13′E ) , 32 km apart from each other . Thirty isolates were obtained per vineyard from single infected berries , resulting in about 180 isolates per year . From each sample , HA ( 1% ( w/v ) malt extract , 0 . 4% ( w/v ) yeast extract , 0 . 4% ( w/v ) glucose , pH 5 . 5 ) cultures were grown , and terminal mycelial fragments cut off for subculture of isolates . Conidia were used for fungicide tests . A list of B . cinerea strains is shown in Table S5 . Fludioxonil , cyprodinil ( Syngenta-Agro , Maintal , Germany ) , fenhexamid , tebuconazole ( Bayer Crop Sciences , Monheim , Germany ) , boscalid , iprodione ( BASF , Ludwigshafen , Germany ) , were kindly provided by the companies , carbendazim , tolnaftate and cycloheximide were purchased from Sigma-Aldrich ( St . Louis , USA ) . The drugs were dissolved in 100% ethanol or 100% DMSO ( carbendazim ) , and added to the required concentrations to the assays . For dilution series , fungicide stock solutions were adjusted to keep the final solvent concentrations between 0 . 2 and 1 . 5% ( v/v ) for ethanol and between 0 . 3 and 1 . 0% ( v/v ) for DMSO . Control assays revealed no significant differences in growth of the strains at these concentrations relative to no-solvent controls ( not shown ) . For each isolate tested , 2×105 conidia were pre-incubated for 1 . 5 hours in 1 ml malt extract broth ( pH 5 . 5; Difco ) before use . Effective inhibitory drug concentrations ( EC50; mg/l ) were determined with 1000 spores in 0 . 1 ml 96-microplate cultures , using threefold drug dilution series . Tests were performed in malt extract broth , except for cyprodinil ( Gamborg B5 minimal medium supplemented with 10mM KH2PO4 , 50mM glucose; pH 5 . 5 ) , and boscalid [38] . After 48 h ( boscalid: 96 h ) incubation at 20°C , A600 was determined . The assays were repeated at least 3 times . Mean data , with standard deviations are presented . For calculation of EC50 values , the Origin6 . 0 software package ( Origin Lab Cooperation , USA ) was used . Fungicide accumulation assays with 14C-labeled fludioxonil and bitertanol were performed with 14 h old germlings germinated as described previously [27] . Experiments were initiated by adding the labeled fungicide to final concentrations of 6 µM ( 10 Bq/nmol ) fludioxonil , or 10 µM ( 10 Bq/nmol ) bitertanol . The uncoupler CCCP was added at a final concentration of 10 µM . Three 5 ml samples each were taken 10 and 60 min after adding the fungicide . Heat inactivation of germlings for control experiments was performed for 10 min at 60°C . Experiments were done in triplicates and repeated at least three times . DNA isolation and manipulation was performed according to established protocols . For transcript studies , B . cinerea conidia ( 2×106 ) were germinated for 15 h in polystyrene Petri dishes coated with apple wax ( 0 . 01 mg/cm2 ) using Gamborg B5 medium supplemented with 10 mM fructose and 10 mM KH2PO4 ( pH 5 . 5 ) . The germlings were incubated for further 30 min either without or with 1 mg/l fludioxonil . For RNA isolation , the wax with the embedded germlings was scraped from the surfaces with a tissue cell scraper ( TPP AG , Trasadingen , Switzerland ) , centrifuged for 5 min at 4000 rpm at 4°C , washed with 20 ml of ice-cold water and centrifuged once more . The pellet was transferred into a mortar containing liquid nitrogen and sea sand for grinding . Total fungal RNA was isolated using the RNeasy Plant Mini Kit ( Qiagen , Hilden , Germany ) , and reverse transcribed into cDNA with oligo ( dT ) primers ( Verso cDNA Kit; Thermo Fisher Scientific , Surrey , United Kingdom ) . Northern hybridization and quantitative RT-PCR were performed according to standard protocols . Expression of the genes was calculated according to Pfaffel [39] . Transcript levels were normalised against the expression levels of housekeeping genes encoding elongation factor 1α ( BC1G_09492 . 1 ) and actin ( BC1G_08198 . 1 ) , and shown as normalized fold-expression relative to expression levels of non-induced germlings from sensitive strains . Means of at least two biological replicates , with three strains of each phenotype , are shown . The following efflux transporter genes were analyzed: ( Bc ) atrB [16] , ( Bc ) atrD [18] , ( Bc ) atrA [40] , ( Bc ) atrF ( BC1G_01454 . 1 ) , ( Bc ) atrK/BMR1 [16] , BMR3 ( BAC67160; BC1G_02799 ) [41] , ( Bc ) mfsM2 ( BofuT4_P024110 . 1 ) . For identification of mutations in the mrr1 alleles of sensitive , MDR1 and MDR3 strains , mrr1 fragments were amplified from total DNA by PCR , using primers mrr1_TF1-1 and mrr1_TF1-4 , and sequenced . For identification of the mfsM2 alleles in sensitive and MDR2/MDR3 strains , the mfsM2 upstream region was amplified from genomic DNA by primers mfsM2-pfor/mfsM2-prev , yielding a 1625 bp fragment with sensitive strains , and a 2273 bp fragment with MDR2 and MDR3 strains . For confirmation of their identity , the insertions of 6 MDR2 strains were sequenced . For atrB mutagenesis , the construct described by Vermeulen et al . [16] was used . For mrr1 deletion , a genomic B . cinerea mrr1 fragment was amplified ( primers mrr1-for1/mrr1-rev2 ) , digested ( ApaI/SacII ) and cloned into pBSKS ( + ) . Inverse PCR was performed ( primers mrr1-rev1/mrr1-for2 ) , the product digested ( EcoRV/XmaI ) and ligated with a hygromycin cassette [42] . After transformation into B . cinerea [43] , mutants were identified with primers mrr1-for1/tubB-inv ( yielding a 1304 bp product in k . o . mutants ) . For mfsM2 deletion , two mfsM2 flanking fragments were amplified ( 1: mfsM2-KO1/mfsM2-KO2; 2: mfsM2-KO3/mfsM2-KO4 ) , digested ( 1: XbaI/EcoRI; 2: KpnI/XhoI ) , and successively cloned into pBSKS ( + ) . The hygromycin cassette from pLOB1 ( AJ439603 ) was inserted between the fragments via EcoRI and XhoI . The construct was amplified ( primers mfsM2KO/mfsM2-KO4 ) and transformed into MDR2 strains D06 . 6-5 and D06 . 2-6 . The ΔmfsM2 mutants were confirmed by PCR ( primers mfsM2-KO1/oliC-Sma-Rev ) , yielding a 1605 bp product in k . o . mutants . To construct strains expressing an mrr1 allele conferring MDR1 , mrr1V575M of strain D06 . 7-27 was amplified ( primers mrr1-atg/mrr1-uaa ) , digested ( XmaI/PvuII ) and cloned into pBSKS ( + ) carrying a 5′-fragment of the hygromycin resistance gene driven by the oliC promoter [44] . Additional 1578 bp of the mrr1 upstream region were amplified ( primers mrr1-pro1/mrr1-rev1 ) , digested ( BamHI/XmaI ) and ligated next to mrr1V575M . The resulting plasmid was linearized ( KpnI ) and transformed into B . cinerea B05 . Hyg-3 [44] . To generate mfsM2 overexpression strains ( mfsM2ox ) , mfsM2 was fused to the oliC promoter . The mfsM2 coding sequence was amplified ( primers mfsM2-ATG-SmaI/mfsM2-TAG-EcoRI ) , digested ( SmaI/EcoRI ) and ligated into poliGUS-Hyg5 , replacing uidA . poliGUS-Hyg5 was constructed by fusing an oliC promoter fragment from pLOB1 ( primers KO-Hyg1-BamHI/oliC-Sma-Rev ) , an uidA coding sequence from p35S-GUS [45] ( primers 35S-gus-for-Sma/35S-gus-rev-Eco ) , the B . cinerea niaD terminator ( primers niaDTerm-for-Eco/niaDTerm-rev-Hind ) and the 5′ part of a splitted hygromycin resistance cassette from pBS . Hyg-5 [44] ( HindIII/XhoI ) into pBSKS ( + ) . The resulting plasmid was linearized ( KpnI ) and transformed into strain B05 . Hyg-3 , yielding B05 . Hyg-3 ( mfsM2ox ) . To construct mfsM2 promoter-reporter fusions , the oliC promoter fragment in poliGUS-Hyg5 was replaced either by a 1501 bp mfsM2 upstream fragment from strain B05 . 10 , or by a 2149 bp fragment from MDR2 strain D08 . 2-12 including the 1326 bp retrotransposon-derived fragment and the remaining mfsM2 upstream region ( primers mfsM2-pfor-Not/mfsM2-prev-Sma ) , before transformation into B05 . Hyg-3 . B . cinerea B05 . Hyg-3 transformants were analyzed for correct genomic integration of the constructs by PCR . GUS staining of transformants was performed as described [46] . MDR3 strain D06 . 7-33 and the sensitive ( BenR ) strain D06 . 5-25 were chosen for a mixed-inoculation experiment , performed twice in 2007 and 2008 in experimental vineyards at the German Wine Road ( Neustadt an der Weinstrasse ) . In June ( berry stage BBCH-77 ) the vineyards received a standard treatment with Teldor . In summer , immature grapes ( stage BBCH-81 ) were inoculated , using hand spraying bottles , with a 1∶1 strain mixture ( 2×104 conidia/ml per strain ) in water until runoff . The vineyards were randomly divided into three fungicide-treated and three untreated plots . The first inoculation was one day before , the second one day after standard treatment with Switch in the treated plots , and in the same way in non-treated plots . Before grape harvest at late September , 50 ( 2007 ) and 90 ( 2008 ) isolates per plot were recovered from moulded berries from the inoculated plots and from a non-inoculated , untreated control plot nearby . The introduced isolates were identified by fungicide tests , using HA plates containing 0 . 2 mg/l fludioxonil , 5 mg/l iprodione , or 5 mg/l carbendazim . In the control plot , MDR3 strains were never detected , while BenR strains were found with frequencies of 12% ( 2007 ) and 6% ( 2008 ) . Their genetic identity was further confirmed by IGS-AFLP markers [31] . For analysis of the overwintered populations , B . cinerea isolates were recovered in the following spring from bark fragments of inoculated grapevines ( 2008: 137; 2009: 313 isolates ) , by incubation on a selection medium for B . cinerea [47] , followed by fungicide tests . Experiments were performed at least three times , unless indicated otherwise . Statistical differences of data were checked by unpaired , two-tailed t tests , and labeled as follows: n . s . : not significant; * p<0 . 05; ** p<0 . 01; *** p<0 . 001 . In the graphs , standard deviations are indicated . The DNA sequence reported in this paper has been deposited in GenBank , under accession number GQ292709 ( RE-like gene fragment inserted in mfsM2 ) . Further accession numbers: mfsM2: BofuT4_P024110 . 1; mrr1: BofuT4_P063510 . 1 ( http://urgi . versailles . inra . fr/projects/Botrytis/ ) . | Bacterial and fungal pathogens cause diseases in humans and plants alike . Antibiotics and fungicides are used for disease control , but the microbes are able to adapt quickly to these drugs by mutation . Multiple drug resistance ( MDR ) is well investigated in human pathogens and causes increasing problems with antibiotic therapy . Driven by the continuous use of fungicides in commercial vineyards , three types of rapidly increasing multidrug resistant populations of the grey mould fungus Botrytis cinerea have appeared in French vineyards since the mid 1990s . Using a combination of physiological , molecular and genetic techniques , we demonstrate that these MDR phenotypes are correlated with increased drug efflux activity and overexpression of two efflux transporters . Just two types of mutations , one in a regulatory protein that controls drug efflux , and the other in the gene for an efflux transporter itself , are sufficient to explain the three MDR phenotypes . We also provide evidence that a subpopulation of the French MDR strains has migrated eastward into German wine-growing regions . We anticipate that by continuous selection of multi-resistant strains , chemical control of grey mould in the field will become increasingly difficult . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"microbiology/plant-biotic",
"interactions",
"microbiology/microbial",
"growth",
"and",
"development",
"microbiology/microbial",
"evolution",
"and",
"genomics"
]
| 2009 | Fungicide-Driven Evolution and Molecular Basis of Multidrug Resistance in Field Populations of the Grey Mould Fungus Botrytis cinerea |
The immune state of wild animals is largely unknown . Knowing this and what affects it is important in understanding how infection and disease affects wild animals . The immune state of wild animals is also important in understanding the biology of their pathogens , which is directly relevant to explaining pathogen spillover among species , including to humans . The paucity of knowledge about wild animals' immune state is in stark contrast to our exquisitely detailed understanding of the immunobiology of laboratory animals . Making an immune response is costly , and many factors ( such as age , sex , infection status , and body condition ) have individually been shown to constrain or promote immune responses . But , whether or not these factors affect immune responses and immune state in wild animals , their relative importance , and how they interact ( or do not ) are unknown . Here , we have investigated the immune ecology of wild house mice—the same species as the laboratory mouse—as an example of a wild mammal , characterising their adaptive humoral , adaptive cellular , and innate immune state . Firstly , we show how immune variation is structured among mouse populations , finding that there can be extensive immune discordance among neighbouring populations . Secondly , we identify the principal factors that underlie the immunological differences among mice , showing that body condition promotes and age constrains individuals’ immune state , while factors such as microparasite infection and season are comparatively unimportant . By applying a multifactorial analysis to an immune system-wide analysis , our results bring a new and unified understanding of the immunobiology of a wild mammal .
Immune systems protect animals from infection and disease . The mammalian immune system reacts to antigenic exposure with a multitude of cellular and humoral responses , including cytotoxicity , phagocytosis , and the production of cytokines and antibodies . The ultimate result of these responses—which we call immune function—is to reduce the effects of infection on an individual . While it is possible to measure many aspects of the immune response ( such as the number and state of cells and the concentrations of antibodies and cytokines , both as observed in an animal and after experimental perturbation ) , the relationship between these measures , their effect on infections , and the consequences for an animal’s health , and ultimately its fitness , is much less clear . Although the terminology is debatable , for clarity we will refer to the observed state of an animal’s immune system as its immune state and so differentiate this from immune response , which we reserve to refer to specific responses to a known antigenic challenge . A wild animal’s immune state is therefore a consequence of its immune responses to the myriad , largely unknown , antigenic challenges that it receives during its daily life . The immune state and immune responses of laboratory animals , particularly mice ( Mus musculus domesticus ) , are extremely well known , which contrasts starkly with what is known about the immune state of wild animals [1] . Because the behaviour of the immune system is highly context dependent , the a priori expectation is that the immune state of wild animals will differ from that of laboratory animals . This assumption is supported by the relatively few studies that have compared wild and laboratory mice [1–7] . For example , 2 studies have shown that wild mice respond more strongly to immunization than laboratory mice [1 , 4]; comparison of wild and laboratory mice has shown that the wild mice have higher proportions of activated CD4+ T cells [2] . A detailed analysis of the adaptive humoral , adaptive cellular , and innate immune state of wild mice found that , compared with laboratory mice , the immune systems of wild mice were in a highly activated state , seen as elevated serum antibody and acute phase protein concentrations; proportionately greater CD4+ and CD8+ effector T cell populations; highly activated natural killer ( NK ) cells; and a myeloid cell population hitherto unknown from laboratory mice [1] . This highly activated state is likely due to the many infections that wild mice have experienced ( and the consequent intense antigenic challenge to which they are subject ) compared to laboratory mice [1] . In contrast to this comparatively activated state , wild mouse cytokine responses to pathogen-associated molecular patterns ( PAMPs ) were similar to , or lower than , those of laboratory mice , possibly representing the impact of homeostatic mechanisms among wild mice to avoid immunopathology in the face of sustained antigenic challenge . These results also highlighted the very extensive interindividual variation among wild mice in their immune state [1] . The importance of the environment in driving differences between wild and laboratory animals' immune systems has also been shown by experiments that cohoused wild and laboratory mice ( thus exposing the laboratory mice to a range of infections ) , resulting in the laboratory mice acquiring more effector memory cells as well as more tissue-resident memory T cells , compared with control laboratory mice [2] . These trends also appear to be consistent in other wild rodents , though there are probably interspecific differences too [7] . These immunological differences between wild and laboratory mice therefore imply that laboratory-based models cannot fully inform us about the immune state or function of wild animal populations and , rather , that direct study of wild populations is required . Understanding the immune function of wild populations is necessary to understand their biology , but also necessary to understand the biology of the populations of pathogens with which they are naturally infected . Animals' immune responses are the environment in which pathogens live and evolve , and this immunological environment exerts a strong selective force on these pathogens . The zoonotic spillover of infections , particularly from wild animals to humans , is currently of enormous international importance [8 , 9] . An important , understudied aspect of this is the immunological state of wild animals that likely modulates the zoonotic transmission risk . Analysis of wild animal populations also emphasizes that individuals differ in their immune state [1 , 3 , 10] . Such interindividual heterogeneity could be due to genetic and/or environmental effects and/or their interaction , but the nature of the specific environmental factors that drive and constrain wild animals' immune systems is not known . Identifying these factors and how they modulate the immune responses ( and so the immune state ) of wild mice is the purpose of the work we present here . Among wild animals ( principally mammals and birds ) there are well-described seasonal effects on measures of individuals' immune state , but these patterns differ among studies . For example , some measures of immune state increase in winter , whereas others decline [11–13] . Notwithstanding this variation , these seasonal patterns may be driven by a number of different ( and potentially interacting ) factors , including hormonal differences ( for example , due to photoperiod-driven effects per se , and/or seasonal reproduction ) , resource availability , and the force of infections [11–14] . Animals' resource availability has been shown to affect immune responses in both laboratory and wild animals , consistent with the idea that these responses are costly and become compromised when resources are limiting [15 , 16] . Age has a significant deleterious effect on immune responses and immune state ( described as immunosenescence ) in laboratory animals and in humans [17 , 18] , though for wild mammals this has only been shown in wild Soay sheep [13 , 19–21] . For wild mammals more generally , both the difficulty of determining wild animals' age and that wild animals may lead shorter lives than laboratory animals [1] may make detecting age-related effects challenging . Collectively , these studies point to the effects that a range of different factors can have on animals’ immune responses and on immune state . However , what these previous studies have not considered is the relative magnitude , importance , and interaction of these different factors acting on wild animals’ immune state . Moreover , our immune system-wide analysis of laboratory and wild mice demonstrated that the various compartments of the immune system differ in very different ways [1] , meaning that analysis of just one or a few immune parameters will underrepresent an individual animal’s immune state and underestimate differences among animals or between populations . By necessity , many previous studies of wild animals have used just one , or a few , immune measures , with these sampled in different ways ( for example , peripherally or systemically ) ; the disparate conclusions drawn from different studies of wild animal immune systems is likely due , at least in part , to the use of different immune parameters . Here we analyse in detail the immune state of populations of wild M . musculus domesticus in the southern United Kingdom . We have studied M . musculus domesticus as an example of a wild mammal population , and one that is ideal for this because it is the same subspecies as the laboratory mouse [22] , making the plethora of laboratory mouse immunological tools available . M . musculus domesticus live commensally , typically in farm outbuildings , within which there is deme-based breeding centred around a dominant male [23 , 24] . These populations are regulated by the same top-down and bottom-up ecological processes that regulate populations of fully free-living species , and so , critically , the immune state of these commensal animals will be subject to similar drivers and constraints as are other wild species . Commensal mice have limited migration between sites , with dispersal typically undertaken by young males , together making populations relatively stable , which facilitates sampling [25] . We find that local M . musculus domesticus populations differ immunologically and that these populations are genetically distinct , but that genetic differences among populations do not contribute to the populations' immunological differences . We then identify the principal factors that underlie the immunological differences among mice , showing that body condition promotes and age constrains individuals’ immune state . In female mice , these factors interact such that although age constrains immune state , improving body condition with age leads to an overall improvement in immune state with increasing age . In male mice , these 2 parameters act independently . Thus , this work demonstrates that detailed immunological analyses of wild populations can reveal extensive heterogeneity of immune state among wild animals and can identify the key factors affecting this . Importantly , our study reveals that host intrinsic factors such as body condition and age are much more important drivers of immune state than genetic background , microparasite infection , or extrinsic factors such as season . These results emphasise that understanding wild animals' immune systems is necessary to fully understand the biology of wild animal populations , as well as that of their pathogens .
We sampled 460 house mice , M . musculus domesticus , from 12 sites in the southern UK ( Fig 1; S1 Table ) . The sex ratio was 1 female to 1 . 18 male , and the median age was 7 . 4 weeks , with 75% of mice being 12 weeks of age or younger ( S1 Fig ) . The mice had a limited macroparasite fauna , consisting primarily of Syphacia sp . pinworm nematodes ( prevalence 79% ) and Myocoptes musculinus mites ( prevalence 67%; S1 Data ) . We tested mice for seropositivity to 7 microbial infections ( Noro , Minute , Parvo , Sendai , Corona , and Mouse Hepatitis viruses , and Mycoplasma pulmonis ) , finding a high seroprevalence ( 95% ) of one or more of these microparasite infections , consistent with previous reports [26 , 27] . We detected antibodies to Sendai virus in these populations , which , to our knowledge , is its first report in wild mice ( S1 Data ) . The number of these microbial infections accumulated as mice aged ( number of infections and age r = 0 . 27 and 0 . 3 , p < 0 . 0001 , n = 214 and 209 for males and females , respectively , S2 Data ) . For each mouse , we made a large number of immune measures , using a cross-sectional study design ( S1 Text ) [1]; here we focus on the ( 1 ) serum concentration of immunoglobulin ( Ig ) G and IgE and acute phase proteins ( serum amyloid P and haptoglobin ) , ( 2 ) faecal concentrations of IgA , and ( 3 ) numbers , proportions , and ex vivo activation status of splenic T and B cells , regulatory T cells ( Tregs ) , NK cells , dendritic cells ( DCs ) , and myeloid cells . Although there may be short-term perturbations to individuals’ immune state that would not be captured by this cross-sectional approach , studies in voles , Microtus agrestis , and humans have shown that over the longer term , immune parameters are largely temporally stable [28 , 29] , suggesting that our data are generally representative of an individual’s immune state . Because we sampled mice from 12 sites , many of which were geographically clustered ( Fig 1 ) , we first wanted to understand the extent of immunological diversity among these sites . We did this by first using a principal component analysis of our immunological data , revealing 3 components: component 1 ( accounting for 53% of the variation ) comprised 7 cell populations ( scaled number of CD4+ and CD8+ T cells , B cells , NK cells , neutrophils , DCs , and macrophages ) , component 2 ( 13% of the variation ) comprised the serum concentration of IgG and IgE , and component 3 ( 10% of the variation ) was the faecal concentration of IgA ( S3 Data ) . We then calculated the immunological distance among all individual mice [28] as the pairwise distance among mice of these 3 principal components , and then compared these distances within sample sites and among sample sites . The results showed variation in immune state of mice both within and among sites . Overall , the mean among-site immunological distance was greater than the mean within-site distance ( 4 . 7 ± 0 . 197 versus 3 . 8 ± 0 . 096 , mean ± SE; Fig 2 , S2 Table ) , meaning that mice at one site are immunologically more similar to each other than they are to mice from other sites . The immune similarity of mice within each site differs among sites . Within some sites , mice are immunologically very similar , whereas at other sites they are comparatively diverse . For example , mice within sites PF and PH are very homogeneous ( immune distance 1 . 21 ± 0 . 14 , 1 . 43 ± 0 . 019 , mean ± SE ) , while mice within sites ST and HW are more diverse ( immune distance 9 . 64 ± 2 . 3 and 5 . 68 ± 0 . 21 , respectively ) . The immunological differences among sites can operate on a very local geographical scale . For example , mice at site ST are comparatively highly immunologically diverse ( immune distance 9 . 67 ± 2 . 33 ) , and consistently distinct from mice at 5 neighbouring sites ( BM , HW , PF , SP , and JB; Fig 1 , S1 and S2 Tables ) that were 0 . 2–8 ( mean 4 . 3 ) km distant . While there have been some observations of differences in specific immunological parameters among rodent populations ( e . g . [29 , 30] ) , this is the first demonstration , of which we are aware , of the structure of immune variation within and among different wild animal populations . In view of the immunological structure of these wild mouse populations , we next investigated their genetic structure . To do this , we genotyped the mice at 1 , 183 neutral autosomal loci and used this to calculate the genetic distance among individuals ( Fig 3 , S4 Data ) , the fixation index , FST , and the inbreeding coefficient , FIS . This showed that there was strong genetic differentiation among populations from the 12 sites , seen as mice from each site being genetically more similar to each other than to mice from other sites ( Fig 3 ) . This was confirmed by STRUCTURE analyses showing that the most likely number of genetic clusters was 9 , due to mice at some of the sites being genetically more closely related ( S2 Fig ) . The FST values for the wild mice are among the highest described for a wild mammal , with the average FST = 0 . 48 , while the average FIS = 0 . 00052 ( Fig 3 , S3 Table , S2 Fig ) , suggesting that there is limited effective movement of individuals among the sites . The local population genetic structure is not driven by geographical distance ( FST and geographical distance among sites do not correlate; Fig 4 , S1 and S3 Tables ) , as previously seen for this species [23] . For example , sites separated by only 2 . 3 km ( sites BM and HW ) have FST values equal to sites separated by more than 50 km ( sites GL and HW ) ( FST = 0 . 47 and 0 . 45 , respectively ) . Because it is an island , mice on Skokholm are , presumably , a relatively isolated population . Commensal mice have limited migration [25] , and not all individuals breed [24 , 25] , probably contributing to the population genetic structure we observe [23] . These effects are likely exacerbated by episodes of pest control resulting in bottlenecks or local extinction , which is followed by recolonization ( often anthropogenically [25] ) , with consequent genetic founder effects . Together , these processes will contribute to the generation of these highly genetically structured populations of M . musculus domesticus . Importantly , there is no relationship between immunological distance among sites and genetic ( Fig 4 , r = 0 . 087 , p = 0 . 382 ) or geographical distance ( Fig 4 , r = 0 . 192 , p = 0 . 264 and r = 0 . 082 , p = 0 . 339 for distance and log ( distance + 1 ) , respectively ) ( S2 and S3 Tables ) . This means that immunological differences among populations must be driven by other factors . While genotyping at this number of loci provides good coverage of the mouse genome , not all genetic variation will be captured and , for example , effects of loci of strong immunological effect that are unlinked to any of these markers will not be accounted for . With this caveat , these results strongly suggest that the immunological structure of wild mouse populations is driven by factors other than the genetic background or geographical structure of the mouse populations . To identify the other factors that drive , or constrain , immune state in these populations and to explain the differences among individual mice in their immune state , we used structural equation modelling ( SEM ) , in which causal relationships among potential factors are specified and then tested against empirical data to generate quantitative , causal conclusions . We explored a range of different models containing a wide range of factors ( S1 Text , S4 and S5 Tables , S3 and S4 Figs ) . Here we present the model that was the most parsimonious , inclusive , and robust among those that we tested . This model assigns causal relationships among a number of factors acting on the latent variable of Immune State ( S5 Fig ) . We considered factors previously implicated in affecting immune state in wild populations [11 , 12 , 13 , 16–18 , 31–33] , focusing particularly on Season ( measured as day length ) , Age ( calculated from eye lens mass [34] ) , Body Condition ( measured as the scaled mass index [SMI] [35] ) , and microparasite Infection ( measured as the number of microbial infections ) . Each of these factors have been shown individually to affect the immune systems of animals , but how they act and interact in wild , free-living populations is unknown . Body condition has repeatedly been implicated in affecting immune responses and state , but it can be measured in various ways [33] . We used SMI that calculates the scaling exponent of body mass and body size for the population and then used this to scale each animal’s body mass to that of the average body length of the population [35] . In wild mice , SMI correlates strongly with body mass index ( BMI ) ( males r = 0 . 88 , p < 0 . 0001 , n = 247; females r = 0 . 83 , p < 0 . 0001 , n = 211 , respectively , at all sites; S2 Data ) and also correlates with abdominal fat mass ( males r = 0 . 13 , p = 0 . 038 , n = 244; females r = 0 . 25 , p < 0 . 0001 , n = 188; S2 Data ) , although correlations with serum leptin concentration are only significant in females ( r = 0 . 24 , p = 0 . 01 , n = 117; S2 Data ) . This is consistent with studies in laboratory mice and humans , finding a relationship between BMI and the concentration of leptin , but a poor correlation between leptin and the mass of fat [36] . Despite these correlations , we find that different measures of body condition are not all always equivalent in their relationship to various immune parameters ( S2 and S5 Data ) , emphasizing the importance of comprehensively assessing and understanding animals' body condition [33 , 35] . In the SEM analysis , because of the heterogeneity in immune state among the different sites , we focused on the HW site , which was deeply sampled and where the results were broadly representative of all sampled mice ( Fig 2 , S1 Text ) ; SEM results for the full data set are shown in S6 Fig and S6 Table . The SEM analyses considered different immune components in turn , because in preliminary analyses conglomeration of all these immune measures was not tractable within the structural equation models ( S1 Text ) . The 3 immune compartments we used were the latent variables ( 1 ) adaptive cellular immune state ( scaled numbers of CD4+ and CD8+ T cells , and CD19+ B cells ) , ( 2 ) innate cellular immune state ( scaled numbers of NKp46+ NK cells , Ly6G+ neutrophils , CD11c+ DCs , and F4/80+ macrophages ) , and ( 3 ) humoral immune state ( serum IgG and IgE , and faecal IgA concentration ) . Importantly , these 3 immune compartments were each immunologically coherent . The structural equation models revealed that the adaptive and innate cellular compartments of the immune system are principally affected by body condition ( Fig 5 , S7 Table; and correlation of SMI and 7 cell populations in males , r = 0 . 26–0 . 64 , p ≤ 0 . 035 , n ≥ 65; correlation of SMI and 6 cell populations in females , r = 0 . 30–0 . 49 , p ≤ 0 . 033 , n > 49; S5 Data ) and are constrained by age ( Fig 5 , S7 Table ) . In females , these effects are linked because although age has a direct negative effect on the immune system ( Fig 5 , S7 Table; and correlation of age and 4 cellular populations r = −0 . 28 to −0 . 32 , p ≤ 0 . 049 , n ≥ 48; S5 Data ) , as females age their body condition improves , which enhances their overall immune state ( Fig 5; and correlation of SMI and age in females r = 0 . 46 , p < 0 . 001 , n = 79; S5 Data ) . In contrast , in males , age and body condition operate more independently on adaptive and innate cellular compartments ( Fig 5 ) , since the correlation of SMI and age is weaker in males than in females ( correlation of SMI and age in males r = 0 . 24 , p = 0 . 019 , n = 99 , S5 Data ) . In males , for adaptive cellular immune state there is a marginally nonsignificant effect of age on condition ( 0 . 19 , 0 . 099 [Estimate , SE] , p = 0 . 051; S7 Table ) . The age-related decline in immune state is suggestive of immunosenescence [17 , 18] , and this is particularly notable given the short lives ( median age 7 . 4 weeks , S1 Fig ) of these wild animals . Adaptive cellular state and innate cellular state were negatively affected by season in both male and female mice , with the effects stronger in females than in males ( Fig 5 ) . For innate cellular state in males , this effect is marginally nonsignificant ( −0 . 19 , 0 . 1 [Estimate , SE] , p = 0 . 064; S7 Table ) . Although the immune state of these wild mice is very likely driven—at least in part—by immune responses to microparasite infection [1] , the infection parameters assessed here explained very little of the difference among individuals in their immune state . This may be because our primary measure of microparasite infection was serological and thus reflective of historical infection , which would tend to minimize effects on current cellular immune state or , for example , an animal’s condition . Interestingly , the humoral immune compartment is affected very differently by these factors . Antibody concentrations increase with age in both sexes ( Fig 5 , S7 Table; and correlation of age and IgG , E , and A in females r = 0 . 35–0 . 49 , p ≤ 0 . 002 , n ≥ 44; correlation of age and IgE in males r = 0 . 32 , p = 0 . 002 , n = 95 , S5 Data ) . Because antibodies persist , their increasing concentration with age is likely to reflect age-associated accumulated exposure to microparasite infection . Body condition has no effect on the humoral immune compartment , but microparasite infection does ( Fig 5 ) ; microparasite infection is significant in females and shows a similar trend in males ( 0 . 39 , 0 . 21 [Estimate , SE] , p = 0 . 055 , S7 Table ) . Comparing the results for the HW site alone with the results for all sites together shows some coherence in the results , but not identity . Thus , for both innate and adaptive cellular immune compartments , the positive effects of body condition and the negative effects of age are qualitatively consistent; effects of season on immune state are largely consistent ( but differ between these compartments ) , while effects of microparasite infection are more variable . For humoral immune responses , the effects of age and microparasite infection are also qualitatively consistent between the HW site and all sites together . Differences among our sample sites in how these extrinsic and intrinsic factors can affect the immune state of wild mice are , in principle , consistent with our observations of immune variation among different populations ( Fig 2 ) .
We have undertaken an in-depth , immune system-wide analysis of the immune state of wild house mice and used a multifactorial analysis of factors potentially affecting this to understand what actually drives and constrains immune state in the wild , and so why animals differ in their immune state . We have , firstly , shown how immune variation is structured among different populations of wild mice , which is the first time that the structure of immune variation in a wild mammal has been described . We find that mice within a sample site are more immunologically similar to each other than they are to mice from other sites , which is broadly equivalent to the recently reported situation in humans where cohabitation of unrelated individuals explains a very significant amount of the interindividual immune variation [28] . Individual wild field voles , M . agrestis , have been shown to consistently differ in the expression of genes whose products have immunological function , suggesting that there are long-term , individual-specific immune responses [29] and that such effects may contribute to population-level patterns such as we observe for M . musculus domesticus . If this pattern of local , population-specific immune differences is generalised to other wild animal populations this could have very important consequences . For example , for pathogens that are subject to immune-dependent selection , this immune heterogeneity may drive different evolutionary trajectories of the pathogens in different populations of hosts . The potential heterogeneity of the immune landscape in which pathogens live and evolve is not usually considered in studies of pathogen evolution . However , the temporal stability of the immunological patterns we observe , both within and among sites , remains to be investigated in longitudinal studies . The possibility of rapid changes in the immune landscape could bring further complexity to the environment and the selection pressures that pathogens experience . More broadly , our results suggest that the health and disease status of wild animals , as well as their potential resistance to new infections , need to be considered at a local population scale . The causes of this immunological structure—which are likely to be multifactorial—are not yet known , and so investigating this is a priority . Work with human populations has shown that carefully designed studies for this specific purpose are required to unpick these effects [37] . Our second key finding is that cellular immune state ( both innate and adaptive ) is principally affected ( positively ) by animals' body condition and ( negatively ) by their age . In females these factors interact so that as females age , their body condition improves , improving their immune state . In males these 2 factors are less interconnected . Together , this means that the highly activated state of wild mouse immune systems [1] is underpinned by good body condition , but this is also a state that can rapidly senesce . Body condition is also likely to underpin many other important life history traits , such as reproductive success , meaning that there will be competing demands for resources among these different traits . The concept of body condition is widely used , aiming to capture an idea of an animal’s ‘plumpness’ [38] . However , there are different definitions and measures of body condition , and these different measures do not always correlate [39] . We used SMI as our principal measure of body condition and found that it correlated with BMI and with the mass of abdominal fat ( but not with the serum concentration of leptin ) , thus suggesting that in this study at least , body condition represents an animal’s overall body reserves . Given that wild animals are likely to have temporally heterogeneous access to food , it will be important to understand the temporal variability in individuals’ body condition and its relationship to the dynamics with which fat deposits can be converted into available energy . Season and microparasite infection explained rather little of the variation in immune state among individual animals . The absence of any substantial effect of season may be due , at least in part , to the commensal lifestyle of these mice , which might buffer them from typical natural environmental fluctuations . Additionally , complex effects of season may not be adequately captured by day length , the measure of season that we used . Similarly , while infection likely explains the highly activated immune state of wild mice compared with laboratory mice [1] , our measures of infection do not explain the immunological heterogeneity among wild mice . This may be because our measures of microparasite infection were mostly serological and therefore represent historical infection; this is also why our SEM did not allow an interaction from immune state to infection . In undertaking these SEM analyses we explored a large range of potential models , which included diverse immunological and other physiological parameters . Our modelling approach also considered data obtained from site HW alone and from all sample sites combined . While the results from the 2 approaches were broadly equivalent , they are not identical . This discordance likely points to different local ecologies , population demographies , and infections at the different sites , all of which will affect the qualitative and quantitative nature of the factors that both drive and constrain animals’ immune state . Our observation of how immune variation is structured among the populations that we have sampled is also consistent with this view . We explored many ways in which measures of animals’ immune state could be incorporated into our models . The approach we favoured ( modelling 3 immune compartments separately ) is probably not the only approach that could be used . However , our extended analyses of alternative approaches points to the difficulty of capturing ‘immune state’ in a meaningful way in the face of multiple humoral and cellular measures . Overall , the results of our SEM analyses show the differing effects of intrinsic and extrinsic factors on the different immune compartments , suggesting the likely diversity of underlying selective pressures acting on these animals in optimizing their immune responses [40] . The analyses of the effects of these extrinsic and intrinsic factors also emphasize that it is critical to give careful thought to which immune compartments are sampled and how individual parameters are assayed in any given situation to properly and fully understand immune state . While we have made very extensive measurements of the immune state of splenocytes , together with analysis of antibodies and other serum proteins [1] , there are numerous additional measurements that could be made , including from other immune compartments ( in particular from mucosal sites that are major portals of pathogen entry ) , which may be particularly relevant to wild animals . Nevertheless , by using a multifactorial analysis we have been able to understand how different factors act and interact in affecting immune state , and for the first time been able to provide a unified understanding of what drives and constrains this in wild animals . While the mice we studied live commensally in farm buildings , they are subject to the same range of ecological processes as entirely wild animals living independently of humans—for example , top-down and bottom-up regulation , due to density-dependent limitation of resources and predation , and disease-related processes—and we therefore expect that the immune systems of such free-living species will be affected similarly . The commensal mice in this study had a very much more limited macroparasite fauna than has been reported for other species of wild rodents , indicating that the dominant pathogen challenge of wild M . musculus domesticus is provided by bacteria and viruses . However , bacteria and viruses probably provide the bulk of the antigenic challenge in most animals , irrespective of their macroparasite fauna . Overall , therefore , we believe that house mice represent both a valid and a tractable example of wild mammal populations and reveal important , likely generalizable , interactions between environment , physiology , and immunology . However , the key next step is to extend this approach to other species , to understand the common factors that support and constrain immune systems and those that differ among species . Trapping , handling , and sampling wild animals is unavoidably highly stressful for them , and will undoubtedly affect some ( possibly many ) of the immune parameters that are measured; this needs to be borne in mind when interpreting such data and poses considerable challenges for the prosecution of longitudinal studies . Technological developments that allow remote monitoring of key physiological parameters in free-living animals may help to overcome this constraint . Nevertheless , understanding the extrinsic and intrinsic factors that affect wild animals' immune systems gives us a rational way to understand how such perturbations may affect their immune responses and immune state and thus their health and disease , and also what can be done to ameliorate these effects . This study shows the feasibility of analysing in detail the immune state of wild animals ( if suitable immunological tools are available ) so as to begin to understand how the immune system contributes to individuals’ fitness in the wild . The immune heterogeneity that we have observed among different populations begs the question of the relative roles of host biology , antigenic challenge , and other environmental factors in supporting or constraining immune state at the local level .
This work was reviewed and approved by the University of Bristol's Animal Welfare and Ethical Review Board as UB/15/055 . Trapping and killing nonprotected wild animals is not encompassed by the Animals ( Scientific Procedures ) Act 1986 . The mice used in this study were trapped between March 2012 and April 2014 at 12 locations ( S1 Table , S6 Data ) , principally within a 30-mile radius of Bristol , UK; the 2 exceptions were the London Underground system and Skokholm island , southwest Wales . We surveyed many sites for the presence of mice and trapped at each site where mice were found . We visited each site , in succession , every season , although our trapping success was dependent both on site access on the day and the continuing presence of mice , which was variable . Inevitably , therefore , dates when mice were trapped are confounded with site . Mice were trapped with Longworth traps ( Pennon , UK ) baited with oats , raw carrot , and hay bedding and checked at least once every 24 hours . Once caught , mice were transferred to a conventional animal house , where they were weighed before being individually housed and fed on a commercially available rodent diet ad libitum ( EURodent diet 22%; PMI Nutrition International , Brentwood , Missouri , United States ) for an average of 5 . 5 ( SD = 3 . 8 ) days before they were killed ( S6 Data ) [41] . This allowed us to process mice in batches with relevant internal controls , which were C57/BL6 laboratory mice [1] . There were no correlations between the number of days a mouse was housed in the lab and the immune measures , the concentration of IgG , IgA , IgE , the proportion of splenocytes that were CD3+ T cells , CD4+ T helper cells , CD49+CD4+ activated T helper cells , CD8+ T cytotoxic cells , CD49+CD8+ activated T cytotoxic cells , KLRG1+CD8+ terminally differentiated T cytotoxic cells , NKp46+ NK cells , CD19+ B cells , mean fluorescence intensity of PNA+CD19+ B cells , CD11c+ DCs , and F4/80+ macrophage cells . Mice caught on the London Underground were taken to the lab , where they were killed immediately and processed as described below . A total of 460 mice were captured and processed ( S6 Data ) . Mice were killed by intraperitoneal injection with an overdose of sodium pentobarbital , followed by cervical dislocation . Mice were weighed after death; female mice that were later found to be pregnant had the mass of the foetuses subtracted from their final mass , and these values were used in all subsequent analyses . Blood samples were taken by cardiac puncture , and the blood was stored in heparinized containers for no more than 1 hour prior to being centrifuged at 13 , 000 g at 18°C for 10 minutes . The resulting plasma was divided into aliquots and stored at −20°C . Spleens were removed aseptically , weighed , and transferred to 5 mL of cell culture media for further processing as previously described [3] to determine a viable spleen cell number . For each mouse , the gut was removed and stored at −20°C , while the kidneys , liver , heart , and lungs were all removed and weighed prior to storage in 5 mL of formalin ( 10% v/v formal saline ) . The eyes were removed using forceps and stored in 2 mL of formalin . The eye lenses were later removed , dried , and weighed as previously described [33] and used to calculate the age of the mice . Mouse carcasses were then stored at −20°C prior to further measurement and dissection . After defrosting , body length was measured from the tip of the snout to the base of the tail . Fat from within the abdominal cavity was removed and weighed . Skull length and width measurements were taken after the soft tissue had been mechanically removed . Skull length was measured from the tip of the nasal cavity to the base of the skull ( the longest point ) , and the width was taken at the widest point perpendicular to this line , using callipers . The blood haemoglobin concentration was measured using a HemoCue Hb 201 analyser ( HemoCue AB , Ängelholm , Sweden ) . The serum concentration of leptin was measured using a commercially available ELISA kit , following the manufacturer’s instructions ( Insight Biotechnology , UK ) . Mice were examined for evidence of microbial infection using 2 immunocomb kits ( Biogal Galed Labs , Israel ) , which detect antibodies to the Corona , Mouse Hepatitis , Sendai , Minute , Noro , and Parvo viruses and to M . pulmonis . These tests were scored on a 0 to 4 scale , where 0 is the absence of a response , and 1 , 2 , 3 , and 4 are all seropositive results in increasing degrees of positivity . These microbial infections are likely to only represent a subset of all the microbial infections to which wild mice are exposed , many of which may be unknown infections . A visual inspection of each mouse was made as they were removed from the trap to detect large ectoparasites such as ticks and fleas; smaller ectoparasites , such as mites , were counted by visual inspection of the dead mice . The trap contents were also inspected . For each mouse , the total pelt ( skin and fur ) was examined under a dissecting microscope . Ectoparasite samples were removed for identification , and the number of mites observed was classified into the categories: 0–10 recorded as the actual number , 11–20 , 21–30 , 31–40 , 41–50 , 51–100 , 101–200 , 201–300 , 301–400 , 401–500 , 501–1 , 000 , 1 , 001–1 , 500 , and 1 , 501–2 , 000 . A sample of mites were mounted on glass slides using Hoyer’s medium and identified as M . musculinus by morphological examination . To determine the intestinal nematode fauna , stored intestines ( see above ) were defrosted and slit longitudinally , the gut then examined using a dissecting microscope , and worms counted as described previously [41] . Detailed morphological examination of a sample of the worms identified them as nematode pinworms , Syphacia sp . The SMI was calculated using the equation previously described [35] , which determines the exponent of the relationship between body mass and body length in the population under study , and then uses this value to calculate the mass of each individual animal if it was average size . In this way , this identifies animals that have disproportionately large or small body masses . The BMI of each mouse was calculated as body mass in kg / ( body length in m ) 2 . For female mice that were found to be pregnant , the mass of their foetuses was subtracted from their final body mass . Counts of immune cell type were obtained using methods previously described [1] , and these counts then scaled analogously as for the calculation of SMI ( above ) , but replacing mouse mass with the count for each cell type measured . The relevant scaling component was calculated separately for each of the relevant cell types . This results in the count of each immune cell for each mouse scaled to the length of the average mouse . We used this scaling method for the following cell types: CD4+ T cells , CD8+ T cells , CD19+ B cells , NKp46+ NK cells , Ly6G+ neutrophils , CD11c+ DCs , and F4/80+ macrophages [1] . Scaling the immune cells in this manner , rather than using a proportion of each cell type , allowed us to use count data in our modelling while still controlling for effects of mouse body size . Mice were genotyped at 1 , 183 autosomal loci as described in [1] . In total , 445 mice were genotyped: 443 wild mice and 2 laboratory C57BL/6 mice ( L88 female and L90 male ) , as in [1] . The neutrality of the loci was tested using the Bayescan program [42] , which detected 15 loci that were unlikely to be neutral and these were removed from further analyses , leaving data for 1 , 168 loci . We used Arlequin to calculate Hardy-Weinberg expected measures of heterozygosity and tested for deviations from Hardy-Weinberg equilibrium using the Markov chain exact test [43] with 1 million iterations . Initial analysis identified many loci as significantly deviating from Hardy-Weinberg equilibrium , suggesting the existence of the Whalund effect , and therefore , a repeat of these analyses on a per sample-site basis was conducted , resulting in far fewer loci being found to deviate from Hardy-Weinberg equilibrium . We further tested for genetic subdivision among the mice using Wright’s FST , calculated in Arlequin , using a pairwise method , where FST significance was determined through 100 repeated measures; FIS was calculated similarly . We calculated the genetic distance among individual mice as the number of nucleotide differences and used this to construct a nearest-neighbour joining tree using 1 , 000 bootstraps , as in [1] . We also used STRUCTURE [44] to further investigate the population structure of the mice , which we did using the admixture model with correlated allele frequencies . This procedure assigns proportions of an individual’s genome to defined genetic clusters , K , and we used K values from 1–12 ( with 12 being the number of sample sites , Fig 1 ) , with 15 iterations ( with 100 , 000 Markov Chain Monte Carlo iterations and a burn-in length of 50 , 000 ) . Because of the unequal sample sizes among sites , we repeated these analyses using 6 randomly selected mice from each site , with 15 iterations , as above . We favoured the admixture model , recognizing that individual mice likely had some mixed ancestry . We also used Structure Harvester [45] to determine the posterior probability for the true value of K , and CLUMPPAK [46] was used to compile consensus results and to generate figures . This was calculated as described by [28] . Briefly , we performed a principal component analysis of 10 immune measures: the scaled number of CD4+ T cells , CD8+ T cells , CD19+ B cells , NKp46+ NK cells , Ly6G+ neutrophils , CD11c+ DCs , F4/80+ macrophages , the serum concentration of IgG and IgE , and the faecal concentration of IgA . This generated 3 principal components representing all 7 cell populations; serum IgG , serum IgE; and faecal IgA ( accounting for 53% , 13% , and 10% of the total variance , respectively ) and from these we calculated the 3-principal component Euclidean pairwise distance among mice , which is the immunological distance [28] , and multidimensional scaling of these data was used to plot them using R 3 . 2 . 2 . We constructed tanglegrams [47] of the immunological , genetic , and geographical distances ( S1 , S2 , and S3 Tables ) among sample sites ( Fig 1 ) from UPGMA trees of each measure , which were calculated using MEGA6 [48] , where the crossovers between trees were minimized . We used Mantel tests to test whether there were significant correlations between immunological distance and geographical distance and between immunological distance and genetic distance , using 999 randomisations to assess significance . Because geographical distance is skewed , we used both the actual geographical distance and log ( distance + 1 ) . We used an SEM approach to seek to understand how a range of extrinsic and intrinsic factors constrained and drove the immune state of wild mice . We used an iterative approach with respect both to the factors included in the models and to the model structures . This is explained more fully in S1 Text , S4 and S5 Tables , and S3 and S4 Figs . Based on these preliminary analyses , we divided our data by sex , and we also analysed all mice from all sites together , and for site HW only ( Fig 1 ) . The available sample size prevented random division of the data ( were it possible , one-half of the data would be used for model construction while the other half of the data would be used for model validation ) . The results of these preliminary analyses led to the causal diagram shown in S5 Fig . In this way , we were able to assign directionality ( and thus causality ) to each pathway among the factors . This led to the structural equation model of the observed variables of Season , Age , Condition , and microparasite Infection , all in relation to the latent variable of Immune State . The model included all possible directional pathways ( though because infection was measured serologically , thus representing historical infection , whereas condition was a contemporary measure , we only considered microparasite Infection affecting Body Condition and not vice versa ) . Season was the number of minutes of daylight , between sunrise and sunset , on the day that the mouse was caught , with these data obtained from the UK Meteorological Office ( www . metoffice . gov . uk ) ; we note that this minutes-of-daylight measure may not fully capture all seasonal aspects , for example , because days of equal length can occur in different seasons . Age was calculated from the eye lens mass as in [34] . Body condition was the SMI , calculated as in [35] . Infection was the number of microbial infections that were serologically detected . We used separate models for the latent variable of Immune State as ( 1 ) the adaptive cellular ( scaled number of CD4+ and CD8+ T cells , and CD19+ B cells ) , ( 2 ) the innate cellular ( scaled number of NKp46+ NK cells , Ly6G+ neutrophils , CD11c+ DCs , and F4/80+ macrophages ) , and ( 3 ) the humoral immune ( concentration of serum IgG and IgE , and faecal IgA ) state . All variables were scaled by factors of 10 so that the majority of data were within the range of 1–10 , as required for SEM analysis . Data do not need to be normally distributed for SEM analyses , but they do need to be linear , as our data were . Each model was run separately for female and male mice . The goodness of fit to the data for each model was assessed by the chi-square test of model fit ( χ2 ) , the root mean square error of approximation ( RMSEA ) , the standardized root mean square residual ( SRMR ) , and the comparative fit index ( CFI ) . The results shown are of the standardized covariances since most variables were measured using different scales . Nonsignificance ( p > 0 . 05 ) for the χ2 goodness-of-fit test was used to indicate an acceptable fit of the model to the data , but our sample sizes make χ2 unreliable as a measure of goodness of fit . RMSEA and SRMR values less than 0 . 05 were accepted as a good fit of the model to the data; however , RMSEA values greater than 0 . 05 were accepted where the lower 90% confidence limit was 0 . 000 . CFI values greater than 0 . 95 were accepted as a very good fit , and anything greater than 0 . 80 was considered as acceptable [49 , 50] . Occasionally , models ran , but with warnings about the fit of the model or the appropriateness of the data used . If these problems could not be resolved , the models were either discarded as unreliable or have been presented with the relevant warning message . All structural equation models were constructed using Mplus version 7 . 2 [49–51] . | The immune state of wild animals—and the factors that affect this—is largely unknown . Knowing this is important in understanding how infection and disease affects wild animals , and the biology of their pathogens . The paucity of knowledge about wild animals' immune state is in stark contrast to our exquisitely detailed understanding of the immunobiology of laboratory animals . Here , we have investigated the immune ecology of wild house mice—the same species as the laboratory mouse—characterising their adaptive humoral , adaptive cellular , and innate immune state . Firstly , we investigated how immune variation is structured among mouse populations , finding that there can be extensive immune discordance among neighbouring populations . Secondly , we identified the principal factors that underlie the immunological differences among mice , showing that body condition promotes and age constrains individuals’ immune state . By applying a multifactorial analysis to an immune system-wide analysis , our results bring a new and unified understanding of the immunobiology of a wild mammal . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"blood",
"cells",
"animal",
"types",
"medicine",
"and",
"health",
"sciences",
"animal",
"pathogens",
"immune",
"cells",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"population",
"genetics",
"animals",
"clinical",
"medicine",
"population",
"biology",
"antibodies",
"zoology",
"immune",
"system",
"proteins",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"t",
"cells",
"immune",
"response",
"immune",
"system",
"biochemistry",
"eukaryota",
"wildlife",
"cell",
"biology",
"clinical",
"immunology",
"genetics",
"of",
"the",
"immune",
"system",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"evolutionary",
"biology",
"organisms"
]
| 2018 | The ecology of immune state in a wild mammal, Mus musculus domesticus |
Viral immune evasion by sequence variation is a major hindrance to HIV-1 vaccine design . To address this challenge , our group has developed a computational model , rooted in physics , that aims to predict the fitness landscape of HIV-1 proteins in order to design vaccine immunogens that lead to impaired viral fitness , thus blocking viable escape routes . Here , we advance the computational models to address previous limitations , and directly test model predictions against in vitro fitness measurements of HIV-1 strains containing multiple Gag mutations . We incorporated regularization into the model fitting procedure to address finite sampling . Further , we developed a model that accounts for the specific identity of mutant amino acids ( Potts model ) , generalizing our previous approach ( Ising model ) that is unable to distinguish between different mutant amino acids . Gag mutation combinations ( 17 pairs , 1 triple and 25 single mutations within these ) predicted to be either harmful to HIV-1 viability or fitness-neutral were introduced into HIV-1 NL4-3 by site-directed mutagenesis and replication capacities of these mutants were assayed in vitro . The predicted and measured fitness of the corresponding mutants for the original Ising model ( r = −0 . 74 , p = 3 . 6×10−6 ) are strongly correlated , and this was further strengthened in the regularized Ising model ( r = −0 . 83 , p = 3 . 7×10−12 ) . Performance of the Potts model ( r = −0 . 73 , p = 9 . 7×10−9 ) was similar to that of the Ising model , indicating that the binary approximation is sufficient for capturing fitness effects of common mutants at sites of low amino acid diversity . However , we show that the Potts model is expected to improve predictive power for more variable proteins . Overall , our results support the ability of the computational models to robustly predict the relative fitness of mutant viral strains , and indicate the potential value of this approach for understanding viral immune evasion , and harnessing this knowledge for immunogen design .
The ideal way to combat the spread of HIV-1 is with an effective prophylactic or therapeutic vaccine [1] , [2] . One of the greatest challenges hindering the achievement of this goal is the incredible sequence diversity and mutability of HIV-1 [3] , which can limit the effectiveness of the immune response [2] , [4] . CD8+ T cells are instrumental in reducing viral load in HIV-1 acute infection [5] and in maintaining the viral set point during chronic HIV/SIV infection [6] , [7] . However , HIV-1 is able to escape the CD8+ T cell response through mutations in or adjacent to HIV-1 epitopes that are presented by HLA class I molecules on the surface of the infected cells [7] . One proposed strategy for realizing a potent prophylactic or therapeutic vaccine is to target CD8+ T cell responses to conserved regions of HIV-1 , aiming to reduce incidences of immune escape or , if escape occurs , to reduce viral fitness and lower the viral set point , thereby slowing disease course and reducing transmission at the population level [8] , [9] . While escape mutations at highly conserved sites often damage the viability of virus [10] , this approach is confounded by the development of compensatory mutations which restore or partially restore viral fitness [9] . Thus , to maximize the effectiveness of a vaccine-induced immune response one must look beyond conservation of single residues to identify regions where mutations are not only highly deleterious , but where further mutations elsewhere in the proteome are unlikely to restore lost fitness , but rather , lead to additional fitness costs due to deleterious synergistic effects . Our group has developed computational models to identify such vulnerable regions of the HIV-1 proteome and to predict the fitness landscape of HIV-1 proteins , providing tools for designing vaccine immunogens that may limit both HIV-1 evasion of CD8+ T cell responses and the development of compensatory mutations [11] , [12] . In an early qualitative study we identified groups of amino acids in HIV-1 Gag coupled by structural and functional constraints that cause these residues to co-evolve with each other , but evolve nearly independently of the other residues in the protein [12] . In analogy with past studies on the economic markets and enzymes [13]–[15] , we termed these groups of residues “sectors” . This analysis and human clinical data revealed one sector in Gag , which we termed sector 3 , where multiple mutations were more likely to be deleterious . This group of residues is naturally targeted more by elite controllers [12] . It is expected to be particularly vulnerable to CD8+ T cell responses that target multiple residues in it since multiple mutations within this sector are likely to significantly diminish viral fitness , thereby restricting available escape and compensatory paths [12] . This approach , however , does not allow us to determine precisely which residues should be targeted , as it does not quantify the relative replicative viability of viral strains bearing specific mutations . Nor does it identify viable escape routes that remain upon targeting residues in the vulnerable regions , or inform how best to block them . To begin to address these issues , we developed a computational model , rooted in statistical physics , which aims to predict the viral fitness landscape ( viral fitness as a function of amino acid sequence ) from sequence data alone and applied it to HIV-1 Gag [11] . Similar methods have previously been employed to study other complex biological systems , from describing the activity patterns of neuronal networks [16]–[19] to the prediction of contact residues in protein families [17] , [20] , [21] . The idea underlying our approach is to first characterize the distribution of sequences in the population , which we expect to be correlated with fitness ( see below ) . Due to the small number of available sequences compared to the size of the sequence space , direct estimation of the probability distribution characterizing the available sequences is precluded . Thus , we instead aim to infer the least biased probability distribution of sequences that fits the observed frequency of mutations at each site , and all correlations between pairs of mutations ( the one- and two-point mutational probabilities ) . Mathematically , “least biased” implies the distribution that has maximum entropy in the information-theoretic sense [22] . The maximum entropy distribution that fits the one- and two-point mutational probabilities has a form reminiscent of that describing equilibrium configurations of an Ising model in statistical mechanics . We generated such models using multiple sequence alignments ( MSA ) for the four subunit proteins of Gag in HIV-1 clade B [11] ( described in Supporting Information Text S1 , Section 1 ) . This model assigns to each viral strain an “energy” ( E ) , which is inversely related to the probability of observing this sequence . We expect more prevalent sequences to be more fit , consistent with expectations from simple models of evolution [23] though the precise correspondence between fitness and prevalence may have a more complicated dependence on factors such as the shape of the fitness landscape , as predicted by quasispecies theory [24] . Furthermore , this expectation could be confounded by immune responses in the patients from whom the virus samples were collected , and phylogeny . Recent analyses suggest ( described more fully in the discussion ) that in spite of these effects , at least for Gag proteins , the rank order of prevalence and in vitro replicative fitness should be similar [25] . Strains with high E values are predicted to be less fit than strains with low E values . Predictions of the model seemed to be in good agreement with experimental data on in vitro replicative fitness , as well as clinical observations on the frequency and impact of viral escape mutations [11] . Our aim in the current work is twofold . First , we present new advances in the inference and modeling of viral fitness landscapes that address previous theoretical and computational limitations . Second , we describe new in vitro fitness measurements for viruses containing multiple Gag mutations , performed to further test fitness predictions using the improved computational methods . To give a broad test of the predictive power of the fitness models , we have performed comparisons for HIV-1 strains containing multiple mutations predicted to harm HIV-1 viability as well as combinations predicted to be relatively fitness neutral . We find that fitness measurements of these mutant strains are in good agreement with model predictions .
Our key hypothesis in formulating models of HIV fitness is that the prevalence of viruses with a given sequence , that is , how often the sequence is observed , is related to its fitness . Simply , fitter viruses should be more frequent in the population than those that are unfit . This hypothesis can be proven for some idealized evolutionary models [23] , but cannot be made exact for the complicated nonequilibrium host-pathogen riposte between humans and HIV . However , our theoretical work , backed by extensive computational studies , suggests that the rank order of fitness and prevalence of strains should be strongly monotonically correlated , provided we compare sequences that are phylogenetically close [25] . Thus , if we construct a model to predict the likelihood of observing different viral strains with given sequences , it can predict the relative fitness of the strains . We achieved this goal by constructing a maximum entropy model for the probability of observing sequences in the MSA [26] . The simplest model in this class is an Ising model , a simple model of interacting binary variables from statistical physics which has been widely applied to study collective behavior in complex systems . The parameters of this Ising model are obtained by imposing the constraint that it reproduce the pattern of correlated mutations ( relative to the consensus sequence ) observed in a multiple sequence alignment ( MSA ) of HIV-1 amino acid sequences extracted from infected hosts . Specifically , the parameters were chosen such that the frequency of mutations at each single residue and the frequency of simultaneous mutations at each pair of residues were the same in both the Ising model and the MSA . Importantly , the model also reproduced higher order mutational correlations accurately , even though these mutational frequencies were not directly fitted [11] . As described in our previous publication [11] , in the Ising model amino acid sequences in the MSA are compressed into binary strings by assigning a 0 to each position where the amino acid matches the consensus sequence ( “wild-type” ) , and a 1 to each position with a mismatch ( “mutant” ) . While this binary approximation greatly simplified our modeling approach , the reduction in complexity has several drawbacks . Firstly , there is a loss of residue-specific resolution . The fitness predictions of our model are insensitive to the precise identity of mutant amino acids , and thus the model cannot resolve fitness differences between proteins containing different mutant amino acid residues in a particular position . Secondly , for relatively conserved proteins such as HIV-1 Gag , where the number of viable amino acids at each position is rather limited , this binary simplification represents a reasonable approximation . However , it is less justified for highly mutable proteins where the wild-type residue in each position is not the overwhelmingly most probable amino acid , as is the case for the HIV-1 Env protein . In our original approach , we fit the Ising model parameters to precisely reproduce the observed one and two-residue mutational correlations within the MSA . However , simultaneous mutations at certain pairs of residues were never observed . This led to another deficiency in our original modeling approach in that pairs of mutations not observed in the MSA were predicted to be completely unviable ( E = ∞ ) . While it is possible that such mutant viral strains have exactly zero replicative fitness , it is more likely that they are highly unfit strains ( possessing non-zero replicative fitness ) that simply arise too seldom to be observed within our finite-sized MSA . In this work , we present three significant advances of our original model to predict viral fitness , which also the aforementioned limitations . First , we incorporate Bayesian regularization into our fitting procedure to eliminate the prediction of zero replicative fitnesses for mutations not present within our MSA . Second , we implement a new algorithm for inferring an Ising model from sequence data , which dramatically accelerates the computation of model parameters . Third , we relax the binary approximation to infer viral fitness landscapes that explicitly retain the amino acid identities at each position . We achieve this by describing the viral fitness landscape using a multistate generalization of the Ising model known as the Potts model , another established and well-studied model in statistical physics [27] . We also implement Bayesian regularization into the fitting of the Potts model parameters . Inference of the parameters of the Ising models , commonly referred to as the inverse Ising problem , is a canonical inverse problem lacking an analytical solution that may be tackled in many ways [16] , [17] , [19]–[21] , [28] , [29] . We improve upon our previous techniques described in [11] by incorporating regularization and implementing new inference algorithms , which greatly decrease the computational burden and accelerate model fitting . To control the effects of undersampling and to improve the predictive power of the inferred fitness models , we incorporate Bayesian regularization into our inference algorithm [18] , [19] , [30] , [31] in the form of a Gaussian prior distribution for the model parameters describing pairwise couplings between residues ( see Text S1 , Sections 1 . 3 and 2 . 5 ) . Regularization of this form is also known as Tikhonov regularization or ridge regression [32] . With this addition , the probability of observing any sequence , including those containing pairs of mutations not observed in the MSA , is nonzero . We have also computed a correction to the energy of each sequence to account for the possible bias that strains near fitness peaks are more likely to be observed than would be expected from their intrinsic fitness when sampled from a finite distribution ( see Text S1 , Section 3 . 2 ) . In an algorithmic advance over our previous fitting procedure , we fit the parameters of our regularized Ising model using the selective cluster expansion algorithm of Cocco and Monasson [18] , [19] which identifies clusters of strongly interacting sites and iteratively builds a solution for the whole system by solving the inverse Ising problem for each cluster . With this approach , we cut the CPU time necessary to infer the parameters of the Ising model from roughly 12 years [11] to 5 hours for p24 , an improvement by four orders of magnitude . Roughly , we expect algorithm run-time to scale as O ( Nn exp ( n ) ) , where N is the system size ( number of amino acids ) and n is the size of a typical “neighborhood” of strongly interacting sites . For a review and applications of this method see [18] , [30] . Complete details of our modeling approach and numerical fitting procedures are provided in the Text S1 , Section 1 . An ideal model of viral fitness would be able to capture the full ( unknown ) distribution of correlated mutations throughout the sequence , and thus reproduce the prevalence of every viral strain . Sequences in the MSA represent a sample of the possible strains of the virus , providing information about the distribution of point mutations , pairs of simultaneous mutations , triplets of simultaneous mutations , and all higher orders . However , since the number of available sequences in the MSA is very small compared to the size of the accessible sequence space , and because mutations at most sites are rare , higher order mutations will be severely undersampled . Thus , following our previous approach we appeal to the maximum entropy principle to seek the simplest possible model capable of reproducing the single site and pair amino acid frequencies [11] , [22] , for which the problem of undersampling is less severe . From this analysis , the Potts model is the least structured model capable of reproducing the one and two-position frequencies of amino acids observed within the MSA [21] . To introduce the Potts model , we represent the sequence of a particular m-residue protein as a vector , , where the elements Ak can take on the q = 21 integer values [1 , 2 , … , 21] denoting an arbitrary encoding of the 20 natural amino acids , plus a gap [21] . In the Potts model the probability of observing a particular sequence is given by ( 1 ) In analogy with the statistical physics literature , we refer to E as a dimensionless “energy , ” the function as the Hamiltonian , and the normalizing factor Z as the partition function [27] . The model is parameterized by a set of m q-dimensional vectors , , and a set of m ( m−1 ) /2 q-by-q matrices , . The hi vectors give the contribution of the identity of each amino acid in each position to the overall sequence energy , and the Jij matrices give the contribution to the energy of pairwise interactions between amino acids in different positions . To fit the Potts model , we implemented a generalization of the semi-analytical extension of the iterative gradient descent implemented by Mora and Bialek [11] , [17] . This approach implements a multi-dimensional Newton search to iteratively adjust the model parameters until the predictions of the model for the one and two-position frequencies of amino acids reproduce those observed within the MSA . In an advance over the original incarnation of this algorithm , we have derived closed form expressions for the gradients required by the Newton search , thereby obviating the need for their numerical estimation by finite differences ( which would result in a more computationally expensive and less numerically stable secant search procedure ) . Our approach is semi-analytical in the sense that while we have analytical expressions for the Newton search gradients , we use a Monte Carlo procedure to numerically estimate the one and two-position amino acid frequencies predicted by the model at each stage of parameter refinement . We are currently developing a Potts generalization of the cluster expansion algorithm [19] to accelerate fitting . We incorporate Bayesian regularization into our fitting procedure in a precisely analogous manner to that described above for the regularized Ising model by introducing a Gaussian prior distribution over the Jij parameters . Inference of the Potts model parameters for p24 required approximately 1 . 4 years of CPU time using a generalization of the gradient descent approach described in Ref . [11] . Fitting the model parameters by gradient descent is expected to scale as O ( ( Np ) 2 ) , where N is the number of amino acids in the protein , and p is the characteristic number of mutant residues observed at each position . Full details of the fitting and regularization procedures are provided in Text S1 , Section 2 . The code implementing the inverse Potts inference algorithm is also provided in Supporting Information Code S1 . To test the accuracy of these models in predicting the fitness landscape of HIV-1 Gag , we performed in vitro experiments to measure the fitness of various Gag mutants . Previously we had measured the in vitro replication capacities of 19 Gag p24 mutants , 16 of which contained single mutations in Gag p24 , and compared these with fitness predictions of our original Ising model [11] . Here , we extend this work to measure the replication capacities of HIV-1 strains containing various combinations of mutations , predicted to be either harmful to HIV-1 viability or fitness-neutral , in Gag p24 and p17 and we compare measurements not only to the original Ising model described in Ref . [11] , but also to regularized versions of Ising and Potts models that we have developed here . Specifically we considered 17 mutations pairs , one triple , and 25 single mutations within these combinations , as listed in Table 1 . These mutations were introduced into the widely used laboratory-adapted HIV-1 clade B reference strain NL4-3 . The tested mutants can be divided into 4 categories , viz . ( i ) Gag p24 pairs with high E values located within a group of co-evolving amino acids termed sector 3 ( cf . Ref [12] ) , ( ii ) HLA-associated Gag p24 pairs with high E values , ( iii ) Gag p24 pairs/triple with low E values , and ( iv ) Gag p17 pairs ( Table 1 ) . These mutation combinations were chosen according to E values predicted by the published Ising model [11] , where E>90 or E = ∞ were considered high E values and E<15 were considered low E values . Note that , due to the couplings between mutations at different sites , parameterized by the Jij in equation 1 , the E values depend not only on the specific mutations introduced but also on the sequence background . The E values for mutations reported here are computed with the HIV-1 NL4-3 sequence background , which differs from the p17 and p24 MSA consensus sequences by 8 mutations ( R15K , K28Q , R30K , K76R , V82I , T84V , E93D , S125N ) and 2 mutations ( N252H , A340G ) , respectively . The p24 region of Gag was focused on since this is the most conserved region of the protein . First , we selected six mutation pairs , predicted to be unfavorable in combination , in sector 3 of Gag p24 since we previously found this to be an immunologically vulnerable group of co-evolving residues in which multiple mutations are not well-tolerated [12] . Since it is desirable to identify low fitness/non-viable combinations of escape mutations for vaccine immunogen design aimed at reducing viral fitness or blocking viable escape pathways , we aimed to identify pairs of likely escape mutations with high E values . Virus mutations that are statistically associated with the expression of specific host HLA class I alleles , which also restrict the same epitopes in which the mutations are found , are likely to be CD8+ T cell-driven escape mutations [33] . We therefore tested five high E pairs of mutations located at HLA-associated Gag p24 codons ( HLA-associated variants defined in [34] , [35] ) in or next to optimal CD8+ T cell epitopes ( A-list epitopes from the Los Alamos HIV sequence database [36] ) that were restricted by the same HLA . For comparison with high E mutation pairs , mutation combinations with low predicted E values were included in testing , comprising known favorable compensatory pairs in Gag p24 where 219Q compensates for the 242N escape mutant [37] and 147L compensates for the 146P escape mutant [10] , as well as one pair in sector 3 of Gag p24 and a Gag p24 triple mutant . Additionally , for broader testing , two mutation pairs in Gag p17 were selected . We note that the most commonly observed mutant amino acid at each codon was tested . We introduced these mutation combinations into the HIV-1 NL4-3 plasmid by site-directed mutagenesis and their presence was confirmed by sequencing , as described previously [38] . Generation of mutant viruses from mutated plasmids and the measurement of their replication capacities were performed as previously [11] , [38] . Briefly , mutated plasmids were electroporated into an HIV-1-inducible green fluorescent protein reporter T cell line , harvested at ≈30% infection of cells , and the replication capacities of the resulting mutant viruses were assayed by flow cytometry using the same cell line . Replication capacities were calculated as the exponential slope of increase in percentage infected cells from days 3–6 following infection at a MOI of 0 . 003 , normalized to the growth of wild-type NL4-3 ( RC = 1 ) . Three independent measurements were taken and averaged . Mutant viruses were re-sequenced to confirm the presence of introduced mutations .
The values of E predicted by our original and new modeling approaches for the 43 HIV-1 NL4-3 Gag mutants tested here are shown in Table 2 . Absolute comparison of the E values between the models are not meaningful , but the relative E values of mutants are generally in excellent concordance between models ( Pearson's correlation , r≥0 . 85 and p≤5 . 3×10−11 , two-tailed test ) . The in vitro fitness measurements for all mutants , grouped according to categories , are shown in Figure 1 . We initially compared our model predictions and fitness measurements for each category of mutant pairs to evaluate whether mutant combinations with high and low predicted E values corresponded to substantial fitness cost or little/no fitness cost , respectively . Briefly , all Gag p24 sector 3 mutation pairs with high E values were not viable in our assay system , and were assigned a replication capacity of zero ( Figure 1A ) . Similarly , with the exception of 315G331R , the five high E HLA-associated mutation pairs showed substantial reduction in replication capacity , to between 0–56% of wild-type levels ( Figure 1B ) . Non-viable mutants ( RC = 0 ) were those for which the generation of virus stocks from plasmids encoding these mutation pairs failed , or , in two instances – mutants 186I295E and 186I331R – were not viable unless further mutations developed , confirming unfavorability of the mutation combination . Briefly , concentrated virus stocks for mutants 186I295E and 186I331R were harvested at >22 days post-electroporation compared with the median harvesting time of 6 days post-electroporation for all mutants ( at which time the 186I295E and 186I331R mutants had infected ≈1% cells ) . Sequencing of these viruses revealed the presence of additional mutations and/or reversion of introduced mutations . For mutant 186I295E , amino acid mixtures were detected at codons 63 ( Q/R ) , 177 ( D/E ) and 186 ( I/V ) , and for mutant 186I331R , mixtures were detected at codons 168 ( I/V ) and 331 ( K/R ) , as well as reversion of 186I to 186T . On repeating virus generation for these mutants , additional mutations similarly developed – mixtures were observed at codons 214 ( R/K ) and 271 ( N/S ) for mutant 186I295E , and 232 ( R/M ) and 260 ( D/E ) for mutant 186I331R . With the exception of 186I295E and 186I331R , sequencing confirmed that all mutant viruses had only the specific mutations introduced . The spontaneous mutations 186V , 271S and 232M were not observed in the MSA and the new mutation combinations did not have lowered E values in any of the models , with the exception of the incomplete 186I331R260D combination ( complete observed combination 186I , 331R , 232R/M , 260D/E ) which displayed a slightly lower energy than 186I331R in the regularized Ising model only ( 11 . 5 vs . 13 . 7 ) ( data not shown ) . Nevertheless , these observations confirm that 186I295E and 186I331R are unfit mutation combinations requiring compensatory paths to restore viability . Taken together , the data on high E p24 mutants confirm mutation combinations predicted to be unfit , and also identify combinations of HLA-associated mutations in/next to optimal CD8+ T cell epitopes ( mutations likely to result in CD8+ T cell escape [33] ) that carry substantial fitness costs . Those p24 mutation combinations , including known compensatory pairs , that were predicted to have low E values displayed replication capacities similar to that of wild-type NL4-3 , indicating that these combinations had little or no cost to HIV-1 replication capacity in accordance with predictions ( Figure 1C ) . Similarly , all p17 mutants tested had replication capacities close to that of the wild-type NL4-3 virus , consistent with the predicted E values of all mutants except 86F92M ( Figure 1D ) . Overall , for only two ( 86F92M and 315G331R ) of the 17 mutant pairs the fitness measurement did not correspond to the E value prediction of high or low fitness cost . It should however be noted that the disparity between E values and measured replication capacities for these mutant pairs is somewhat mitigated in the regularized models . The E values for the regularized Ising model for these mutants ( which were assigned an E value of infinity by the original Ising model ) are lower than those of other mutants previously assigned infinite energies , and the same is true for mutant 86F92M in the regularized Potts model . Next , we assessed the relationship between fitness measurements and E values predicted by our original Ising , regularized Ising and regularized Potts models using Pearson's correlation tests . There is a strong correlation between the metric of fitness ( values of E , Table 1 ) predicted by the original unregularized Ising model and our experimental measurements ( Pearson's correlation , r = −0 . 74 and p = 3 . 6×10−6 , two-tailed ) ( Figure 2A ) , however this correlation out of necessity excludes mutants with E values equal to infinity ( n = 13 ) . The regularized Ising model allows for inclusion of these data points resulting in a stronger correlation between predictions and fitness measurements ( Pearson's correlation , r = −0 . 83 and p = 3 . 7×10−12 , two-tailed ) ( Figure 2B ) , which is slightly improved by focusing on Gag p24 mutants only ( Pearson's correlation , r = −0 . 85 and p = 1 . 4×10−11 , two-tailed ) . There is also a strong agreement between the residue-specific Potts model energies and replication capacity ( Pearson's correlation , r = −0 . 73 and p = 9 . 7×10−9 , two-tailed ) ( Figure 2C ) . In practice , one may be concerned with a more coarse-grained measure of viral fitness: will a virus with a given sequence be able to replicate with similar efficiency to the wild-type , or will it be significantly impaired ? To explore this point , we grouped the experimentally tested mutants into two categories , “fit” ( RC≥0 . 5 ) and unfit ( RC<0 . 5 ) , and tested the ability of the fitness landscape models to predict which class each sequence would belong to based on their E values . This was accomplished by fitting a linear classifier to the data using logistic regression ( Text S1 , Section 3 . 1 ) . The regularized Ising model E classifier is highly accurate ( 91% accuracy at optimal threshold , AUROC = 0 . 93 ) – we observed a strong , significant difference in replication capacities between the mutants classified as unfit and those classified as fit ( Mann-Whitney U = 32 , ) ( Figure 3A ) . Specifically , four mutants ( 86F92M , 190I , 190I302R and 243P ) were not classified correctly . However , 190I302R , which was classified as unfit ( E = 8 . 6 ) , exhibited a fitness close to that of the 0 . 5 cutoff ( RC = 0 . 56 ) and 243P , which displayed low fitness ( RC = 0 . 36 ) , had a predicted E value ( E = 7 . 4 ) bordering on the classifier E value . The Potts model classifier also performs well ( 81% accuracy at optimal threshold , AUROC = 0 . 80 ) , but provides a slightly weaker difference between the fit and unfit classes ( Mann-Whitney U = 70 , ) ( Figure 3B ) . Here , seven mutants were not classified correctly , including the same four not classified correctly by the regularized Ising model as well as mutants 174G , 181R , 269E and 315G331R . Similar to mutant 243P , mutants 174G , 181R and 269E were unfit ( RC = 0 ) but had a predicted E values ranging from 7 . 3 to 7 . 7 , fairly close to that of the classifier E value .
In this study , we have substantially advanced our modeling approaches and tested the predictive power of these models by in vitro fitness measurements of HIV encoding various mutation combinations in the Gag protein . The in vitro functional data are overall in strong agreement with the viral fitness landscape models and support the capacity of these models to robustly predict both continuous and “coarse-grained” measures of HIV-1 in vitro replicative fitness . Performance of the regularized Potts and regularized Ising models here is similar , which is not unexpected as Gag in general is not highly mutable and the mutants tested here were the most common ones , making the binary approximation a fairly good assumption . Indeed , in instances where the binary approximation is valid , we might encounter poorer performance from the Potts model relative to the Ising due to a diminished ratio of samples ( i . e . , sequences in the MSA ) to parameters ( i . e . , h and J values ) making robust numerical fitting of the former more challenging than the latter . It is nevertheless encouraging that we are capable of fitting a significantly more complicated Potts model that retains residue-specific resolution without compromising the fidelity of our predictions . Improved inverse Potts inference methods which better meet these numerical challenges may also improve performance of the Potts model with respect to the Ising model results . Simple theoretical analysis suggests that models which differentiate between different mutant amino acids at the same site , like the Potts model employed here , will be necessary to make fitness predictions for highly mutable proteins such as Env and Nef , or to predict the fitness of sequences containing sites with mutations to less frequently observed amino acids . Using a simple toy model , we show in Text S1 , Section 3 . 4 that the binary approximation ( Ising model ) has several potential deficiencies compared to a Potts model . In particular , the Ising model generically overestimates the fitness of mutant sequences , particularly for sequences containing uncommon mutations . Also , in the Ising case the inferred interaction between mutations at different sites is dominated by the interaction between the most common mutants , while the Potts model is able to accurately capture interactions between rare mutants . Future work will involve testing Ising and Potts model predictions for more highly variable proteins and for mutations to uncommon amino acids . While this study confirms the usefulness of this method for predicting HIV-1 replicative fitness , at least for closely related sequences , caution will be necessary in applying this method to predict the relative fitness of multiple strains separated by a large number of mutations . In the measure of prevalence used to infer the Ising and Potts model fitness landscapes , factors such as phylogeny are implicitly included . Analysis conducted in [25] suggests that phylogenetic effects influence the value of the inferred fields hi , and that a correction should be included for predictions of energy or fitness . This form of a correction is sensible , as phylogenetic effects should make mutations at individual residues less frequent , leading to larger inferred fields . For closely related strains such as those studied here experimentally , any systematic inaccuracies in the energy due to phylogeny should be similar in magnitude , and thus differences in energy should predict relative replicative fitness fairly accurately . This would not necessarily be true , however , for sequences separated by many mutations . Further theoretical developments may be needed to separate out the contributions of phylogeny and intrinsic fitness from the Ising and Potts model landscapes presented here [25] to predict the relative fitness of strains that differ by many mutations . In addition to phylogeny , other factors such as host-pathogen interactions and pure stochastic fluctuations affect the observed distribution of sequences , and could complicate fitness predictions . In another work [25] we have investigated these issues by carrying out stochastic simulations that aim to mimic the way the samples were collected and host-pathogen dynamics . In this paper , for the p17 protein , we found that the fitness and prevalence were not the same . However , the rank order of fitness and prevalence were the same as long as the strains being compared were not very far apart in sequence space . This is largely because of the diversity of immune responses due to diverse HLA types in the human population . Additionally , the number of virus particles in single infected individuals from whom the virus sequences were extracted is large , as is the number of patients from whom the virus samples were taken . We find that the one and two-point mutational probabilities in the sequence databases have converged [11]; i . e . , these correlations do not change upon removal of some sequences , suppressing the effects fluctuations on the inferred model . We also note that some caution should be taken in comparing E values for sequences belonging to different proteins . The fitness predictions of the Ising and Potts models are unchanged by a constant shift in energy for all sequences , thus comparisons of absolute energy values are not physically meaningful . Differences in energy between two sequences in the same protein , however , can be unambiguously interpreted as the fitness ratio of those sequences . This is the approach we have taken when examining E values from sequences with mutations in p17 and p24 together: rather than comparing the absolute energies , we compare the differences in energy between the mutant and the NL4-3 reference sequence in each protein , which reflect the fitness of the mutant relative to the NL4-3 reference sequence . Finally , translation from differences in energy to differences in fitness might depend on the specific protein that is being considered . While comparisons of energy differences and relative fitnesses of p17 and p24 mutants performed here exhibit no obvious incongruences , further study is needed to confirm the generality of fitness predictions across proteins . In the case of two mutant pairs ( 186I295E and 186I331R ) that were predicted by the models to have very low fitness , partial reversions and/or additional mutations spontaneously arose in culture that restored virus viability . However , with the exception of one of the spontaneous mutations ( 260D ) observed in combination with 186I331R that modestly decreased the predicted energy ( increased fitness ) in the regularized Ising model but not the Potts or original Ising models , the models do not predict lower energies ( increased fitness ) for these mutant pairs in combination with the additional mutations arising in vitro . Further , three of the spontaneous mutations – 186V , 271S and 232M – were not observed in the MSA and therefore could not be assessed by the Potts model . As a possible interpretation of these findings , we suggest that it may be the case that these mutation patterns observed in vitro are not typically observed in vivo , perhaps since these are infrequently explored mutational routes . As a corollary , this could indicate an inherent limitation of computational models derived from clinical sequence data to identify all possible escape-compensatory pathways , and the importance of in vitro and in vivo experiments to validate and complement model predictions . A mitigating factor , of course , is that mutational pathways observed in vitro but not in vivo may be of less direct clinical relevance . In future work , the model predictions will be further validated in animal models by testing the viable escape pathways predicted to emerge following immunization with immunogens containing vulnerable HIV-1 regions only . The validated fitness landscape could then be used to design vaccine immunogens containing epitopes from the vulnerable regions that could be presented by people with diverse HLAs and that target residues particularly harmful to HIV-1 when mutated simultaneously , thereby substantially diminishing viral fitness and/or blocking viable mutational escape [11] , [12] . Such immunogens potentially represent good therapeutic vaccine candidates to overcome the challenge of HIV-1 evasion of CD8+ T cell responses . However , further work will also be required to optimize design of such immunogens to ensure that epitopes included are processed effectively and that they are sufficiently immunogenic , as well as to test their immunogenicity , optimal delivery methods and protection efficacy in animal models . Furthermore , fitness landscapes of HIV-1 proteins may also be more widely applied to identify effective antibody targets , and help design potent combinations of neutralizing antibodies for passive immunization as well as small molecule inhibitors for therapy . | At least 70 million people have been infected with HIV since the beginning of the epidemic and an effective vaccine remains elusive . The high mutation rate and diversity of HIV strains enables the virus to effectively evade host immune responses , presenting a significant challenge for HIV vaccine design . We have developed an approach to translate clinical databases of HIV sequences into mathematical models quantifying the capacity of the virus to replicate as a function of mutations within its genome . We have previously shown how such “fitness landscapes” can be used to guide the design of vaccines to attack vulnerable regions from which it is difficult for the virus to escape by mutation . Here , using new modeling approaches , we have improved on our previous models of HIV fitness landscape by accounting for undersampling of HIV sequences and the specific identity of mutant amino acids . We experimentally tested the accuracy of the improved models to predict the fitness of HIV with multiple mutations in the Gag protein . The experimental data are in strong agreement with model predictions , supporting the value of these models as a novel approach for determining mutational vulnerabilities of HIV-1 , which , in turn , can inform vaccine design . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"immunodeficiency",
"viruses",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"medical",
"microbiology",
"hiv",
"viral",
"pathogens",
"genetics",
"microbial",
"pathogens",
"biology",
"and",
"life",
"sciences",
"immunology",
"microbiology",
"computational",
"biology",
"viral",
"diseases"
]
| 2014 | The Fitness Landscape of HIV-1 Gag: Advanced Modeling Approaches and Validation of Model Predictions by In Vitro Testing |
The process of fertilization is critically dependent on the mutual recognition of gametes and in Plasmodium , the male gamete surface protein P48/45 is vital to this process . This protein belongs to a family of 10 structurally related proteins , the so called 6-cys family . To identify the role of additional members of this family in Plasmodium fertilisation , we performed genetic and functional analysis on the five members of the 6-cys family that are transcribed during the gametocyte stage of P . berghei . This analysis revealed that in addition to P48/45 , two members ( P230 and P47 ) also play an essential role in the process of parasite fertilization . Mating studies between parasites lacking P230 , P48/45 or P47 demonstrate that P230 , like P48/45 , is a male fertility factor , consistent with the previous demonstration of a protein complex containing both P48/45 and P230 . In contrast , disruption of P47 results in a strong reduction of female fertility , while males remain unaffected . Further analysis revealed that gametes of mutants lacking expression of p48/45 or p230 or p47 are unable to either recognise or attach to each other . Disruption of the paralog of p230 , p230p , also specifically expressed in gametocytes , had no observable effect on fertilization . These results indicate that the P . berghei 6-cys family contains a number of proteins that are either male or female specific ligands that play an important role in gamete recognition and/or attachment . The implications of low levels of fertilisation that exist even in the absence of these proteins , indicating alternative pathways of fertilisation , as well as positive selection acting on these proteins , are discussed in the context of targeting these proteins as transmission blocking vaccine candidates .
Sexual reproduction is an obligate process in the Plasmodium life cycle and is required for transmission of the parasites between the vertebrate and mosquito hosts . The sexual phase is initiated by the formation of male and female cells ( gametocytes ) in the blood of the vertebrate host . Gametocytes are the precursors to the haploid male and female gametes that are produced in the mosquito midgut where fertilisation takes place . Successful fertilisation requires an ordered series of gamete-gamete interactions , specifically , the recognition of and adhesion to the female gamete by the motile male gamete , followed by a cascade of signalling events resulting from the fusion of the two gametes . Despite their fundamental importance , relatively little is known about gamete receptors/ligands and their involvement in the process of gamete interactions of eukaryotes [1] , [2] , which is partly due to their rapid evolution and species-specific characteristics [3] . In Plasmodium the involvement of two gamete specific surface proteins P48/45 and HAP2/GCS1 has been demonstrated in male fertility and these proteins are to date the only known proteins with a demonstrable role in gamete-gamete interaction [4] , [5] , [6] . Parasites lacking P48/45 produce male gametes that fail to attach to fertile female gametes [4] while male gametes lacking of HAP2/GCS1 do attach to females , but they do not fuse due to an absence of membrane fusion between the two gametes [5] . P48/45 is one member of a family of proteins encoded within the genome of Plasmodium and this family is characterised by domains of roughly 120 amino acids in size that contain six positionally conserved cysteines ( 6-cys ) . The 6-cys family of proteins appears to be Apicomplexan specific and has a predicted relationship to the SAG proteins in Toxoplasma gondii [7] , [8] , [9] , [10] , [11] . Ten members of the 6-cys family have been identified . Most members are expressed in a discrete stage-specific manner in gametocytes , sporozoites or merozoites [8] , [12] , [13] , [14] , [15] , [16] . The surface location of members of this family and their expression in gametes or in invasive stages ( sporozoites and merozoites ) suggests that they function in cell-cell interactions as has been shown for P48/45 in gamete adhesion . In addition to P48/45 , five other 6-cys genes are transcribed in gametocytes , three of which ( p230 , p230p and p47 ) are exclusively expressed in the gamete stages of the malaria parasite [4] , [8] , [10] , [12] , [16] , [17] , [18] , [19] , indicating that these members of the gene family may also play a role in the process of gamete recognition and fertilisation . Indeed specific antibodies against the sexual stages of the human parasite Plasmodium falciparum , P48/45 and P230 can prevent zygote formation and thus block transmission of the parasite [19] , [20] , [21] , [22] , [23] , [24] , [25] , [26] . Interestingly , P . falciparum mutants lacking P230 expression produce male gametes that fail to attach to erythrocytes resulting in a reduced formation of the characteristic ‘exflagellation centres’ and reduced oocyst formation in mosquitoes [27] . In order to investigate the role of the 6-cys proteins in parasite fertilisation we performed genetic and functional analysis on the five 6-cys proteins that are expressed in gametocytes . In this paper , we present evidence that in addition to P48/45 , two 6-cys members ( P230 and P47 ) also have an essential role in parasite fertilization . Interestingly , in P . falciparum evidence has been published that P48/45 , P47 and P230 are under positive selection resulting in non-neutral sequence polymorphisms [28] , [29] , [30] , [31] . By sequence analysis , we provide evidence that these three 6-cys proteins are undergoing strong but different rates of positive selection , either as a consequence sexual-selection driven by the competition between gametes or from natural selection exerted by the adaptive immune system of the host on proteins expressed in gametocytes .
The gametocyte-producer clone cl15cy1 ( HP ) of P . berghei ANKA was used as the reference parasite line [32] . In addition , the following mutant lines of the ANKA strain were used: 2 . 33 , a non-gametocyte producer ( NP ) line [33] and 137cl8 ( RMgm-15 , www . pberghei . eu ) , a mutant lacking expression of P48/45 [4] . To disrupt genes encoding different members of the 6-cys family , we constructed a number replacement constructs using plasmid pL0001 ( www . mr4 . com ) which contains the pyrimethamine resistant Toxoplasma gondii ( tg ) dhfr/ts as a selectable-marker cassette ( SC ) . Target sequences for homologous recombination were PCR amplified from P . berghei genomic DNA ( ANKA , cl15cy1 ) using primers specific for the 5′ or 3′ end of the different 6-cys genes ( see Table S1 for the sequence of the different primers ) . The PCR–amplified target sequences were cloned in plasmid pL0001 either upstream or downstream of the SC to allow for integration of the construct into the genomic target sequence by homologous recombination . DNA constructs used for transfection were obtained after digestion of the replacement constructs with the appropriate restriction enzymes ( Table S1 ) . Replacement constructs pL1138 ( p47 ) and pL0123 ( p36 ) , were constructed using replacement plasmid pDB . DT∧H . DB [34] and plasmid pL0121 ( p47&48/45 ) was constructed in the previously described replacement plasmid for disruption of pb48/45 ( plasmid p54 is renamed here to pL1137; [4] ) . This plasmid was made by exchanging the 5′ pb48/45 targeting sequence with the 5′ targeting sequence of pb47 . The p230pII replacement construct pL0120 is a derivative of plasmid pL0016 [35] containing the tgdhfr-ts SC , gfp ( under control of the pbeef1aa promoter and 3′UTR of pbdhfr/ts ) and p230p 5′ and 3′ targeting sequences [36] . Transfection , selection and cloning of mutant parasite lines were performed as described [32] , [37] using P . berghei ANKA cl15cy1 as the parent reference line . For all mutants with an observable phenotype , mutants were generated and selected in two independent transfection experiments ( Table S1 ) . Of each transfection experiment we selected one cloned line for further genotype and phenotype analysis . Correct integration of the construct into the genome of mutant parasites was analysed by standard PCR analysis and Southern blot analysis of digested genomic DNA or of FIGE separated chromosomes [32] . PCR analysis on genomic DNA was performed using specific primers to amplify either part of the wild type locus ( primers WT1 and 2 ) or the disrupted locus ( primers INT1 and 2 ) . See Table S2 for the sequence of these primers . Total RNA was isolated from the different blood stage parasites of the gametocyte-producer clone cl15cy1 of P . berghei ANKA ( HP ) , the non-gametocyte producer line 2 . 33 ( NP ) and the different mutant lines according to standard methods . To determine stage-specific transcription of the 6-cys family members , Northern blots containing RNA from different blood stages were hybridised with different gene specific probes , which were PCR-amplified using the primers shown in Table S2 ( primer pairs WT1+ 2 ) . To detect expression of the P48/45 protein we used polyclonal antiserum raised against recombinant P . berghei P48/45 as described [4] . For detection of P47 we generated the following polyclonal antiserum; a fragment of the Pb47 ORF ( encoding amino acids 80–411 ) was PCR-amplified using primers L964 and L965 ( Table S2 ) and cloned into the NdeI/BamHI sites of the expression vector pET-15b ( Novagen ) providing an N-terminal 6-Histidine tag . Polyclonal antiserum was raised in New Zealand rabbits by injection of 200 µg of gel-purified recombinant protein . Boosting was carried out subcutaneously with 3-weeks intervals using 200 µg protein in incomplete Freund's adjuvant . Serum ( P47 ) obtained 2 weeks after the third boost was immuno-purified on immobilised purified recombinant P47 . To detect P48/45 and P47 in the different mutant lines , total protein samples of purified gametocytes were fractionated on non-reducing 10% SDS polyacrylamide gels . The fertility of wild type and mutant gamete populations was analysed by standard in vitro fertilisation and ookinete maturation assays [4] , [17] from highly pure gametocyte populations [38] . The fertilisation rate of gametes is defined as the percentage of female gametes that develop into mature ookinetes determined by counting female gametes and mature ookinetes in Giemsa stained blood smears 16–18 hours after in vitro induction of gamete formation . Fertility of individual sexes ( macro- and micro-gametes ) was determined by in vitro cross-fertilisation studies in which gametes are cross-fertilised with gametes of lines that produce only fertile male ( Δp47; 270cl1 ) or only fertile female gametes ( Δp48/45; 137cl1 [4] , [17] , [39] . All fertilisation and ookinete maturation assays were done in triplicate on multiple occasions in independent experiments . In vivo ookinete , oocyst and salivary gland sporozoite production of the mutant parasites were determined by performing standard mosquito infections by feeding of Anopheles stephensi mosquitoes on infected mice [40] . Oocyst numbers and salivary gland sporozoites were counted at 7–10 days and 21–22 days respectively after mosquito infection . For counting sporozoites , salivary glands from 10 mosquitoes were dissected and homogenized in a homemade glass grinder in 1000µl of PBS pH 7 . 2 and sporozoites were counted in a Bürker-Türk counting chamber using phase-contrast microscopy [41] . Infectivity of sporozoites was determined by infecting mice through bites of 25–30 infected mosquitoes at day 21–25 after mosquito infection . The formation of exflagellation centres ( i . e . male gamete interactions with red blood cells ) was determined by adding 10µl of infected tail blood to 100–300 µl of standard ookinete culture medium pH 8 . 2 to induce gamete formation . Ten minutes after induction of gamete formation a droplet of 5–10 µl was placed on a cover slip and analysed under a standard light microscope ( 40× magnification ) as a hanging-drop using a well slide . When red blood cells were settled in a monolayer , the number of exflagellating male gametocytes was counted that form or did not form exflagellation centres . An exflagellation centre is defined as an exflaggelating male gametocyte with more than four tightly associated red blood cells [27] . The formation of exflagellation centres was performed using tail blood collected at day 6 or 7 from mice that were infected with 105 parasites without treatment with phenylhydrazine . For quantification of male-female interactions tail blood was collected from phenylhydrazine-treated mice with high numbers of gametocytes [42] . Tail blood ( 10µl ) was collected at gametocytemias ranging between 4–8% and added to 100µl of standard ookinete culture medium pH 8 . 2 to induce gamete formation . Ten minutes after induction of gamete formation , the cell suspension was placed in a Bürker-Türk counting chamber and during a period of twenty minutes the male-female interactions were scored using a phase-contrast light microscope at a 40× magnification . Attachments of males to females were scored if the male had active ( attachment- ) interactions with the female for more than 3 seconds . Penetration of a female by the male gamete was scored as a fertilisation event . Pairwise alignments were generated between the orthologous sequences of p48/45 , p47 and p230 genes in P . berghei , P . yoelii and P . chabaudi; sequences were obtained from PlasmoDB ( http://www . plasmodb . org version 6 . 1; see Table S3 for the accession numbers of the 6-cys gene family members ) . Complete gene sequences for a number of these genes were obtained from the Sanger Institute ( A . Pain , personal communication ) . Maximum-likelihood estimates of rates of non-synonymous substitution ( dN ) and synonymous substitution ( dS ) between pairwise alignments were generated using the PAML algorithm ( version 3 . 14; [43] , [44] ) using a codon-based model of sequence evolution [45] , [46] , with dN and dS as free parameters and average nucleotide frequencies estimated from the data at each codon position ( F3×4 MG model [47] ) . For this analysis we assumed a transition/transversion bias ( i . e . kappa value ) that had been estimated previously and found to be similar in case of P . falciparum and P . yoelii , i . e . 1 . 53 [48] . A sliding window analysis of dN/dS ratios was performed of p230 , p47 and p48/45 from the three rodent parasites . We analysed the dN/dS values of these genes across their length by analysing sequentially 300bp of the gene in 150bp steps . This analysis is essentially the same as the calculation of π ( i . e . the number segregating or polymorphic sites ) described for p48/45 in distinct P . falciparum isolates described by Escalante et al . [29] . We obtained the single nucleotide polymorphisms ( SNPs ) data identified from field and laboratory isolates of P . falciparum ( excluding all P . reichenowi SNPs ) from PlasmoDB ( www . PlasmoDB . org ) . The alignment of these SNPs along the different genes ( to scale ) was extracted from the Genome Browser page of PlasmoDB . The locations of the SNPs were aligned onto the schematic representation of the 6-cys genes of the rodent parasites . It should be noted that the alignment of the p230 gene of the different Plasmodium species was only possible around 1008bp after the putative start site . In order to determine which residues of p230 , p47 and p48/45 genes were under positive selection in the rodent malaria parasites , a Bayes Empirical Bayes ( BEB ) analysis was performed using sequences from the 3 rodent genomes and was calculated as described in Yang et al . [49] . To test which genes were undergoing positive selection the likelihood ratio test ( LRT ) was performed using a comparison of site specific models of evolution [50] , [51] . This test compares a ‘nearly neutral’ model ( without any residues under positive selection ) and a ‘positive selection’ model ( with residues under positive selection and therefore under adaptive evolution ) . Both models assume that there are different categories of codons , which evolve with different speeds . The ‘nearly neutral’ model assumes two categories of sites at which amino acid replacements are either neutral ( dN/dS = 1 ) or deleterious ( dN/dS<1 ) . The ‘positive-selection’ model assumes an additional category of positively selected sites at which non-synonymous substitutions occur at a higher rate than synonymous ones ( dN/dS>1 ) . Likelihood values indicate how well a model fits to the analyzed alignment and answers the question if the ‘positive selection’ model fits better to the analyzed alignment than the ‘nearly neutral’ model . All animal experiments were performed after a positive recommendation of the Animal Experiments Committee of the LUMC ( ADEC ) was issued to the licensee . The Animal Experiment Committees are governed by section 18 of the Experiments on Animals Act and are registered by the Dutch Inspectorate for Health , Protection and Veterinary Public Health , which is part of the Ministry of Health , Welfare and Sport . The Dutch Experiments on Animal Act is established under European guidelines ( EU directive no . 86/609/EEC regarding the Protection of Animals used for Experimental and Other Scientific Purposes ) .
Ten members of the 6-cys family have been identified in Plasmodium and are found in all Plasmodium species ( Table S3 ) . We analysed the transcription profile of the 10 members during blood stage development of P . berghei by Northern blot analysis and combined this analysis with a search of publicly available literature , transcriptome and proteome datasets . This method established that multiple members are transcribed in gametocytes of which four members , p48/45 , p47 , p230 , p230p , are transcribed exclusively in the gametocyte stage ( Fig . 1A ) . The gametocyte specific expression of p48 and p230p has been shown before [4] , [8] . Transcription of p38 occurs both in gametocytes and in asexual blood stages as has also been reported [8] , whereas p12 is transcribed in all blood stages . The relative weak band observed in gametocytes might be due to low contamination of the gametocyte preparation with asexual blood stages ( gametocyte samples always contain a small degree of contamination with schizonts when density gradients are used for gametocyte purification ) . Transcription of p41 and p12p show a complex pattern of multiple transcripts in all blood stages . The close paralogue pair p36 and p36p have quite different transcriptional profiles: p36p is not transcribed in blood stages but transcription is exclusive to sporozoites [14] , [15] whereas p36 is transcribed both in gametocytes ( Fig . 1B; [8] , [52] ) and in sporozoites [14] , [15] . Since no polyclonal or monoclonal antibodies exist for most of the 6-cys family members of P . berghei , except for P48/45 [4] , P47 ( this study ) , P36 and P36p [14] , data on expression of these proteins in different life cycle stages mainly comes from large-scale proteome analyses . For most members of the 6-cys family which have been detected by proteome analysis , the presence of the protein coincides with transcription of its gene ( Fig . 1B ) . The exclusive presence of P48/45 , P47 , P230 and p230p in the proteomes of gametocytes corresponds to the transcription pattern of their respective genes . The presence of P48 and P47 in P . berghei gametocytes has been confirmed using polyclonal antibodies against these proteins ( Fig . S1; [4] ) . P12 , P38 and P41 have been detected in the proteome of merozoites which agrees with their transcription in the asexual blood stages and with their identification in the raft-like membrane proteome of the P . falciparum merozoite surface [13] . Also the presence of P36 in proteomes of both gametocytes and sporozoites [41] , [52] and P36p in sporozoites [14] , [41] fits with the transcription profile of these genes . Up to now only P12p has not been detected in any proteome of Plasmodium . Comparison of the transcription and expression patterns of the 10 conserved members of the 6-cys family of P . berghei with those of P . falciparum from large scale transcriptome and proteome analyses demonstrates that the expression patterns are conserved between the rodent and human parasite ( Fig . 1B ) and also confirms that four out of the 10 members are specific to the gametocyte stage . We previously reported the functional analysis of mutant P . berghei parasites that were deficient in expressing P48/45 , generated by targeted disruption of p48/45 through a double crossover homologous recombination event [4] . Here we have used the same approach , schematically shown in Fig . 2A , to disrupt 5 other members of the 6-cys family that are transcribed in gametocytes . We excluded p12 , p12p , p41 and p36p from this analysis since the results obtained from transcriptome and proteome analyses indicate a role for the first three of these genes during the asexual blood stage development ( Fig . 1B ) . We have previously demonstrated in both , P . berghei and P . falciparum , that P36p is involved in liver-cell infection and disruption of its gene had no effect on development of gametes and fertilisation [15] , [53] . Mutant parasite lines have been generated deficient in P47 ( Δp47 ) , P230 ( Δp230 ) , P230p ( Δp230p ) , P38 ( Δp38 ) or P36 ( Δp36 ) and for each gene , mutants were selected from two independent transfection experiments ( Table S1 ) . Two different Δp230p mutant lines were generated , Δp230p-I and Δp230p-II , differing in which regions of 230p have been disrupted . In mutant Δp230-I a fragment is deleted from the second 6-cys domain ( i . e . first 894aa still present ) onwards whereas in mutant Δp230-II the deleted fragment includes part of the first 6-cys domain ( i . e . first 492 amino acids still present ) . In addition we generated a mutant line deficient in the expression of both P48/45 and P47 ( Δp48/45&Δp47 ) . Correct disruption of the target-genes was verified by diagnostic PCR analysis ( Fig . 2B ) and Southern blot analysis of separated chromosomes and/or digested genomic DNA ( data not shown ) . To demonstrate that the mutant parasite lines were deficient in expression of the targeted gene we analysed transcription of the corresponding genes by Northern blot analysis using mRNA collected from purified gametocytes ( Fig . 2B ) . No transcripts of p47 and p38 could be detected in ΔP47 and Δp38 mutants , and no p48/45 and p47 transcripts are present in the DKO mutant Δp48/45&Δp47 . Only small , truncated transcripts were detected for p230 and p230p in gametocytes of the Δp230 and Δp230p lines and also in Δp36 a truncated p36 transcript was found . Full length transcripts of wt p230 and p230p are 8 . 5 and 9 . 5 kb respectively , whereas truncated transcripts are approximately 2 . 5 kb in size . Since several of the disrupted genes are organised as pairs within the genome ( i . e . p230&p230p and p48/45&p47 ) , we analysed whether disruption of one member of a pair affected transcription of the other gene . For Δp48/45 parasites it has been shown before that disruption of p48/45 had no effect on expression of its paralog P47 [4] . In this study we similarly show for p47 , p230 and p230p that disruption had no effect on transcription of its paralogous member ( Fig . S1 A&B ) . In addition to the transcription analysis of the disrupted genes , we analysed the presence or absence of the proteins P47 and P48/45 in the mutant parasites by Western analysis using polyclonal antiserum ( Fig . S1C ) . P47 is present in wt gametocytes and gametocytes of the Δp48/45 but is absent in Δp47 and Δp48/45&Δp47 gametocytes . P48/45 is present in wild type and absent in the Δp48/45&Δp47 gametocytes . We next analysed the phenotype of the different mutant lines during gametocyte and gamete development as well as during fertilisation , ookinete and oocyst formation using standard assays for phenotype analysis of the sexual- and mosquito stages of P . berghei . Surprisingly , three of the six mutants lacking expression of genes that are transcribed in gametocytes did not exhibit a phenotype that was different from wild type parasites during these stages of development . These mutants , Δp230p , Δp38 or Δp36 , showed a normal growth of the asexual blood stage ( data not shown ) , sexual development and development of the mosquito stages up to the mature oocysts ( Table 1 ) . All these mutant lines produced wild type numbers of gametocytes and gametes and showed normal fertilisation rates as measured by in vitro zygote/ookinete production ( Table 1; Fig . 3 ) . In contrast to the absence of a discernable fertilisation phenotype with the Δp230p , Δp38 and Δp36 mutants , we found that the capacity of fertilisation is severely affected in the other three mutants , ( Fig . 3A ) . Specifically , Δp47 , Δp230 and Δp48/45&Δp47 lines showed a fertilisation rate that was reduced by more than 99 . 9% compared to wt , as shown by the inhibition of zygote/ookinete production in vitro ( Table 1; Fig . 3A ) . These mutants produced normal numbers of mature gametocytes during blood stage development . The analysis of in vitro gamete formation ( exflagellation of males; emergence of female gametes from the erythrocyte ) by light-microscopy also revealed that the process of gametocyte and gamete formation was not affected , resulting in the production of motile male gametes and female gametes , emerged from the host erythrocyte by more than 80% of the mature gametocytes ( Table 1 ) . At 16–18h after activation of gamete formation , the in vitro cultures of Δp47 , Δp230 and Δp48/45&Δp47 lines contained many ( clusters of ) unfertilized , singly nucleated , female gametes . This phenotype of a strong reduction of fertilisation despite the formation of male and female gametes closely resembles the phenotype of Plasmodium parasites lacking P48/45 [4] . As had also been previously observed with the P48/45 deficient mutant , the fertilisation rate of gametes of the three mutant lines seems to be more efficient in the mosquito compared to in vitro fertilisation [4] . Compared to wild type parasites , the in vivo fertilisation of the mutants is reduced by 93–98% as calculated by ookinete and oocyst production in mosquitoes ( Table 1 ) , whereas the reduction of in vitro fertilisation rate is greater than 99 . 9% . Infections of naïve mice through bite of 20–30 mosquitoes infected with parasites of Δp47 , Δp48/45&Δp47DKO and Δp230 parasites , resulted in blood stage infections containing only gene disruption mutants ( i . e . mutant genotype and no ‘wild type’ parasites ) , as determined by PCR and Southern analysis of genomic DNA ( results not shown ) . These results show that gametes of all three mutant lines still have a low capacity to fertilise , resulting in the production of viable and infective ookinetes , oocysts and sporozoites . Moreover , the results obtained with the double knock-out mutant Δp48/45&Δp47 indicate that the few fertilisation events in single knock-out mutants deficient in expression of either P47 or P48/P45 ( this study and [4] ) cannot be explained by a compensation effect due to its paralogous protein because the Δp48/45&Δp47 mutant still shows a comparable , albeit greatly reduced , ability to fertilise and to pass through the mosquito . Fertility of the male and female gametes produced by the mutant lines can be determined by in vitro cross-fertilisation studies , where gametes are cross-fertilised with gametes of parasite lines that produce either only fertile male gametes or female gametes . Such an approach was used to establish that Δp48/45 parasites produced infertile male gametes , whereas the female gametes are completely fertile [4] . We performed different in vitro cross fertilisation experiments to determine whether the reduced fertilisation capacity of the Δp47 and Δp230 mutants was due to affected male gametes , female gametes or to both sexes . Gametes of both mutants were cross-fertilised with female gametes of Δp48/45 ( males are infertile ) to determine male fertility of Δp47 and Δp230 . Male gametes of Δp47 were able to fertilise Δp48/45 females ( at wild-type levels ) whereas the males of Δp230 were unable to fertilise the Δp48/45 females ( fertilisation rates <0 . 01%; Fig . 3B ) . These results demonstrate that male gametes of Δp47 are viable with wild type fertilisation capacity and therefore the fertilisation defect of Δp47 must be due to infertile females . The normal fertility of male gametes of Δp47 has also been shown in previous studies in which the males of this mutant have already been used in other cross-fertilisation studies [17] , [39] , [54] , [55] . The lack of fertilisation in the crossing experiments of gametes of Δp230 with Δp48/45 shows that P230 plays a role in male fertility . In order to test the fertility of Δp230 females we crossed the gametes of this line with the fertile male gametes of Δp47 ( as mentioned above the females are infertile ) . We find that Δp47 male gametes are able to fertilise Δp230 female gametes in a manner identical to their ability to fertilise Δp48/45 females ( Fig . 3B ) . This demonstrates that female gametes of Δp230 have a fertility that is comparable to wild type female gametes and that the fertilisation defect is the result of infertile males . Crossing experiments performed with gametes of the double knockout mutant , Δp48/45&Δp47 with gametes of either Δp230 , Δp47 or Δp48/45 did not result in increased fertilisation rates ( <0 . 01% ) , demonstrating that gametes of both sexes are infertile in the double knock-out mutant ( Fig . 3B ) . In P . falciparum it has been shown that male gametes lacking P230 expression have a reduced capacity to adhere to red blood cells , as measured by the formation of ‘exflagellation centres’ [27] . We therefore examined the ability of P . berghei male Δp230 gametes to attach to erythrocytes , by microscopic examination of exflagellation centre formation under standardized in vitro conditions . In these experiments 76–92% of exflagellating wt males and 72–90% exflagellating Δp230 male gametocytes , formed such centres ( Table 2 ) , indicating that in contrast to P . falciparum Δp230 in P . berghei both wt and Δp230 male gametes have a similar ability to interact with red blood cells . Gametocytes that did not form exflagellation centres were often floating on/above the red blood cell layer during exflagellation . Further analysis of single , free male gametes of Δp230 revealed that they were highly motile and often attach to red blood cells , producing characteristic red blood cell shape deformations due to the active interactions between the male gamete and the erythrocyte . Male gametes lacking expression of P48/45 do not attach to female gametes as has been previously shown by analysing male-female interactions by light microscopy [4] , [5] . We therefore analysed the interactions between male and female gametes of Δp230 or Δp47 , between 10 and 30 minutes after induction of gamete formation using phase-contrast microscopy . In wt parasites attachment of males to females was readily detected with a mean of over 25 attachments during a 20 minutes period of observation , with a mean of more than 6 confirmed fertilisations ( i . e . male gamete penetrations; Table 2 ) . In preparations of gametes of both Δp230 and Δp47 not a single fertilisation event was detected and the number of male and female gamete attachments was drastically reduced ( Table 2 ) . We observed that while male gametes of both mutants undergo active interactions with red blood cells and platelets , attachment of males to female gametes are hardly ever observed . These results show that P230 like P48/45 is a male fertility factor involved in recognition or attachment to females and that P47 is a female fertility factor involved in recognition or adherence by the male gamete . Whether P48/45 and P230 once on the surface of the male gamete directly interact with P47 on the surface of the female gamete is unknown . Unfortunately , repeated immuno-precipitation experiments with anti-P . berghei P48/45 antibodies and wt gamete preparations , in order to identify interacting partners , were unsuccessful ( data not shown ) . Analyses of sequence polymorphisms of p48/45 , p47 and p230 of laboratory and field isolates of P . falciparum has provided evidence that these proteins are under positive selection [28] , [29] , [30] , [31] . We analysed synonymous ( dN ) and non-synonymous ( dS ) polymorphisms of p48/45 , p47 and p230 by comparing these genes in three closely related rodent parasites P . berghei , P . yoelii and P . chabaudi by making use of the newly available gene sequences ( www . PlasmoDB . org version 6 . 1 ) . The updated dN/dS values for these genes obtained here , which is commonly used as an indicator of positive selection , were in all comparisons higher than the mean dN/dS value of all genes within the respective genomes ( Table S4 ) . However , only the dN/dS ratio of p47 in the P . berghei/P . yoelii comparison showed a significant difference with the mean dN/dS value ( 0 . 82 compared to the mean dN/dS of 0 . 26 ) . Overall , P47 is in the top 4–6% of fastest evolving proteins in the rodent parasite genomes as compared to top 10–16% for P230 and 15–50% for P48/45 ( Table S4 ) . In addition , we have used the likelihood ratio test ( LRT ) to analyse if these genes were undergoing neutral or positive selection ( see Materials and Methods ) . This test shows that p47 is indeed under positive selection ( P = 0 . 006 ) when comparing the site/residue specific models of evolution . We next examined sequence mutations in the same genes in more detail by performing a comparative dN/dS ratio analysis across these genes using small and corresponding regions of these genes using a ‘sliding window analysis’ ( i . e . 300bp in 150bp intervals; Fig . 4; Table S4 ) . This analysis showed that p47 has an exceptionally elevated dN/dS value ( i . e . 1–2 ) in one area corresponding to the truncated B-type domain II as defined by [7] . Interestingly , although P230 had a relatively low overall dN/dS value ( 0 . 33–0 . 44 ) , the sliding window analysis revealed that P230 contains several areas where the dN/dS ratio is higher than 1 . 0 with an increased ratio in all 3 species in particular around the B-type domain IV as defined by Gerloff et al . ( 2005 ) . In order to analyse similarities in the location of sequence polymorphism between P . falciparum and the three rodent parasites , we aligned all known single nucleotide polymorphisms ( SNPs ) described for P230 , P47 and P48/45 in P . falciparum ( i . e . www . PlasmoDB . org; [56] , [57] , [58] ) with the dN/dS ratios determined by the ‘sliding window analysis’ ( for details see Materials and Methods; Fig . 4 ) . Interestingly , the elevated dN/dS ratios of p47 domain II and domain IV of P230 , both correspond with the location of high SNP densities in the orthologous P . falciparum genes . These findings would suggest that similar regions in the p47 and p230 genes of rodent parasites and P . falciparum are subject to positive selection . To predict which residues of the three P . berghei genes are under positive selection we performed a Bayes Empirical Bayes analysis ( BEB; [49] ) . This analysis calculates dN/dS values ( ω values ) on each residue of a particular protein when the genes encoding these proteins are compared in least 3 similar species and an ω>1 indicates positive selection on a residue . For P47 ten residues were identified undergoing positive selection with ω values ranging between 4 and 7 ( Table S5 ) . Nine of these 10 residues are confined to the first two domains of P47 including the region B-type domain II . In P48/45 four residues were identified ( ω values ranging between 1 and 2 ) and for P230 only one amino acid ( ω = 1 . 3 ) . Interestingly , this one residue in P230 ( i . e . residue 845V ) maps to the corresponding region of the P . falciparum P230 , domain IV , where 6 of the 27 non-synonymous polymorphisms described by Gerloff et al . map ( Table S4 ) .
Until recently the only protein proven to play a direct role in merging of the male and female gamete of Plasmodium gametes in Plasmodium was P48/45 , a surface protein principally of male gametes shown to play an essential role in recognition of and attachment to females [4] , [5] . Recently , two studies have identified a second protein , HAP2/GCS1 with a role early in fertilisation [5] , [6] . Male gametes of mutant parasites lacking this protein can attach to female gametes but the subsequent fusion of the gametes is absent [5] , a process which is clearly after the mutual recognition and attachment of gametes . Our studies provide evidence for the direct involvement of two additional proteins , P47 and P230 , which like P48/45 play a key role in the initial phase of gamete-gamete recognition and attachment . The phenotype of mutants lacking P230 expression is identical to the phenotype of mutants lacking P48/45 , i . e . male gametes do not recognize and attach to female gametes whereas the female gametes are fertile . These results show that the P230 protein , like P48/45 , is a male fertility factor . A similar role of P48/45 and P230 in male fertility is perhaps not surprising since evidence has been reported that both proteins interact with each other . Unlike P48/45 , P230 does not contain a glycosylphosphatidylinositol ( GPI ) anchor and in P . falciparum evidence has been found that P230 forms a complex with P48/45 at the surface of gametocytes and gametes [18] , [27] , [59] , [60] . Indeed , analysis of P . falciparum mutants has shown that in the absence of P48/45 the P230 protein is not retained on the surface of gametes , a result which may indicate that tethering of P230 to the surface of the male gamete is mediated by P48/45 [27] . In contrast , in the absence of P230 the surface location of P48/45 is not affected in P . falciparum [27] , [61] . If in P . berghei the same interaction occurs , and Δp48/45 gametes also lack surface expression of P230 , then the failure of Δp48/45 and Δp230 males to attach to females might be solely due to the absence of P230 on the male gamete surface . This would imply that P230 and not P48/45 is the major male protein that is responsible for recognition of and attachment to the female . However , it has been shown that antibodies directed against P48/45 strongly reduce oocyst formation [19] , [20] , [24] , [25] , [26] , indicating that either P48/45 antibodies disrupt the attachment of the translocated P230 to P48/45 after gamete formation or it may play a more direct role in fertilisation and that its function is not exclusively as a membrane anchor for P230 . Interestingly , in P . falciparum it has been shown that male gametes with a disrupted p230 gene are incapable of interacting with erythrocytes and do not form the characteristic exflagellation centres and these mutants show a strong reduction in oocyst formation [27] . These observations , in P . falciparum , indicate that P230 not only plays a role in gamete-gamete interactions but male gamete interactions with erythrocytes may be required for gamete maturation resulting in an optimal fertilisation capacity [27] , [62] . Our analyses of Δp230 P . berghei male gametes in live preparations did not reveal any difference in their capacity to interact with red blood cells , suggesting that there are functional differences between P230 of P . berghei and P . falciparum . As the interaction between male and female gametes has not been analysed in the P . falciparum Δp230 mutants it is unknown whether the decreased oocyst formation results from the reduced gamete-erythrocyte interactions or is due to the lack of gamete recognition and attachment , as we have observed in P . berghei . Therefore , further research is needed to unravel whether P . falciparum P230 is also involved in gamete-gamete interactions like P . berghei P230 . Moreover , additional research is required to identify the proteins at the surface of P . berghei male gametes that are responsible for the adherence of the male gametes to erythrocytes . Disruption of the close paralogue of p230 , p230p , did not have any effect on fertilisation or on red blood cell attachment . The distinct phenotypes of Δp230 and Δp230p gametes demonstrate that the proteins encoded by these genes are not functional paralogues that are able to complement each others function as has been demonstrated for the paralogous protein pair P28 and P25 on the surface of zygotes [63] . The same is true for the paralogous proteins P48/45 and P47 ( see below ) or P36 and P36p [15] , [64] . In addition to the important role of P230 in male fertility , our studies demonstrate that P47 plays a key role in P . berghei female gamete fertility . Both proteome analyses of P . berghei gametocytes [17] and IFA analysis of P . falciparum gametocytes using anti-P47 antibodies [12] have shown the female-specific expression of P47 . In P . falciparum , P47 is located on the surface of the female gametes following emergence from the host erythrocyte . Our studies demonstrate that P . berghei females lacking P47 are not recognized by wild type males . These observations may suggest that P48/45 or P230 on the male gamete directly interact with P47 on the female for recognition and attachment . However , P48/45 and P230 may alternatively interact with additional , as yet unknown protein/s on the surface of the female that are dependent on the presence of P47 , in an analogous manner to the interaction between P230 and P48/45 on the surface of the male gamete . Both P48/45 and P230 are also expressed in the female gametes of P . berghei and P . falciparum [17] , [27] . The presence of these proteins on the female gamete surface does not result from male proteins that are released by the male during activation and subsequent binding to the female since ‘pre-activated’ female gametocytes also express these proteins ( B van Schaijk , personal communication and [65] . However , an essential role for P48/45 and P230 in female gametocytes is not implicated in P . berghei since both Δp230 and Δp48/45 females demonstrate normal fertilisation , i . e . to wild-type levels , when incubated with wild type males . Unexpectedly , the lack of expression of P47 in P . falciparum mutants appears not to have a role in fertilisation as determined by oocyst formation in mosquitoes [12] . This difference between P . berghei and P . falciparum suggests that the proposed model of the interactions between male P48/45 and/or P230 with female P47 ( and/or P47-interacting proteins ) being key for the recognition and attachment of gametes does not hold true for all Plasmodium species . However , these differences between P . falciparum and P . berghei might also be explained by the presence of an additional set of protein ligands in both species that mediate additional mechanisms of gamete recognition and attachment . Indeed by analysing P . berghei Δp48/45 mutants [4] and mutants lacking expression of P47 and P230 ( this study ) we found that low levels of fertilisation did occur . Surprisingly , in all mutants significant higher fertilisation rates were observed in mosquito midguts compared to in vitro rates of fertilisation . Even in the mutant lacking expression of both P48/45 and P47 , the same low fertilisation rates are observed . Assuming that P . berghei Δp48/45 gametes lack P230 surface expression as has been shown for P . falciparum Δp48/45 , then gametes of the double knock-out mutant can fertilise in the absence of essentially all three fertility factors of the 6-cys family , albeit at a reduced rate . These observations indicate the presence of additional proteins that secure fertilisation in the absence of the three members of the 6-cys family . For unidentified reasons this alternative fertilisation pathway appears to be much more efficient in vivo than in vitro , suggesting that mosquito factors influence this alternative route of fertilisation . The observed oocyst formation in Δp48/45 and Δp47 P . falciparum parasites [4] , [12] might therefore also be explained by this route of fertilisation and the presence of relatively high numbers of oocysts might indicate that this alternative pathway is more efficient in P . falciparum in A . stephensi compared to P . berghei in A . stephensi . Such alternative pathways of fertilisation may have implications for development of transmission blocking vaccines that block fertilisation using antibodies directed against members of the 6-cys family of proteins and therefore it is important to identify the additional proteins involved in the process of recognition and attachment of gametes . It is possible that other members of the 6-cys family that are expressed in gametocytes ( P230p , P38 and P36 ) may be involved in the alternative pathways of fertilisation . Although we found that gametes lacking expression of these proteins did not show a significant reduction in fertilisation , the effect of their absence on gamete fertility may only become evident in the absence of P48/45 , P47 and P230 . Further research using mutants lacking multiple 6-cys members is required to reveal whether other 6-cys family members or other unrelated proteins play a role in alternative routes of fertilisation . For P48/45 , P47 and P230 in P . falciparum evidence has been published that these proteins are under differing rates of positive selection resulting in non-neutral sequence polymorphisms [28] , [29] , [30] , [31] . Polymorphisms in gamete proteins may be a consequence of sexual selection as is the case for gamete proteins of other organisms [3] , [66] . However , sequence polymorphism in these Plasmodium genes may also result from natural selection exerted by the adaptive immune system of the host . These three proteins are expressed in mature gametocytes , and as only a very small percentage of gametocytes ever get passed on to a mosquito , the vast majority of gametocyte proteins ( including these 6-cys members ) are eventually released into the hosts circulation where they are exposed to the host immune system . Indeed it has been shown that P48/45 and P230 both elicit humoral responses in infected individuals that can mediate transmission blocking immunity [22] , [24] , [67] , [68] , [69] , [70] . Our analyses on dN/dS values of the three rodent parasites provide additional evidence that directional selection pressures affect sequence polymorphisms of gamete surface proteins , especially evident for the female specific p47 which belongs to the top 4–6% fastest evolving genes in the rodent parasite genomes . Analysis of dN/dS variation across the genes by the sliding window approach on P230 identifies one region that is evolving rapidly in all the rodent parasites and , interestingly , this correlates with the same region in P . falciparum ( B-type domain IV ) that has the highest density of SNPs [7] . The correlation of the location of P . falciparum SNP's with increased dN/dS ratios in both P230 and P47 may indicate that similar selection pressures exists in different Plasmodium species . Whether this positive selection on these gamete proteins is driven by immune responses and/or mating interactions is presently unknown . However , insight into sequence polymorphisms in gamete surface proteins that are targets for TB vaccines and the influence of these polymorphisms on mating behaviour of parasites in natural populations of P . falciparum should help to improve TB vaccines development . | Sexual reproduction for malaria parasites is an essential process and is necessary for parasite transmission between hosts . Fertilisation between female and male gametes occurs in the midgut of the mosquito and proteins on the surface of gametes are principle targets in transmission blocking strategies . Despite their importance , relatively little is known about the malaria proteins involved in fertilisation . In this study we show that two gamete proteins , one expressed on the surface of males , the other on the surface of females , have important roles in the mutual recognition and attachment of gametes . Mutant parasites that lack the presence of these surface proteins show a strong reduction in fertility . Comparison of these gamete surface proteins in different malaria parasites showed that these proteins are evolving rapidly either across their length or at discreet regions/domains . We found , that despite the drastic reduction in zygote formation , low levels of fertilisation can still occur in the absence of these surface proteins , indicating that gametes can use alternative proteins to recognize each other . Both genetic variation of gamete surface proteins and the presence of different fertilisation pathways have important implications for transmission blocking vaccines targeting gamete surface proteins . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"genetics",
"and",
"genomics/microbial",
"evolution",
"and",
"genomics",
"cell",
"biology/microbial",
"growth",
"and",
"development",
"microbiology/parasitology",
"infectious",
"diseases/protozoal",
"infections",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases"
]
| 2010 | Three Members of the 6-cys Protein Family of Plasmodium Play a Role in Gamete Fertility |
The inoculum effect ( IE ) is an increase in the minimum inhibitory concentration ( MIC ) of an antibiotic as a function of the initial size of a microbial population . The IE has been observed in a wide range of bacteria , implying that antibiotic efficacy may depend on population density . Such density dependence could have dramatic effects on bacterial population dynamics and potential treatment strategies , but explicit measures of per capita growth as a function of density are generally not available . Instead , the IE measures MIC as a function of initial population size , and population density changes by many orders of magnitude on the timescale of the experiment . Therefore , the functional relationship between population density and antibiotic inhibition is generally not known , leaving many questions about the impact of the IE on different treatment strategies unanswered . To address these questions , here we directly measured real-time per capita growth of Enterococcus faecalis populations exposed to antibiotic at fixed population densities using multiplexed computer-automated culture devices . We show that density-dependent growth inhibition is pervasive for commonly used antibiotics , with some drugs showing increased inhibition and others decreased inhibition at high densities . For several drugs , the density dependence is mediated by changes in extracellular pH , a community-level phenomenon not previously linked with the IE . Using a simple mathematical model , we demonstrate how this density dependence can modulate population dynamics in constant drug environments . Then , we illustrate how time-dependent dosing strategies can mitigate the negative effects of density-dependence . Finally , we show that these density effects lead to bistable treatment outcomes for a wide range of antibiotic concentrations in a pharmacological model of antibiotic treatment . As a result , infections exceeding a critical density often survive otherwise effective treatments .
The inhibitory effects of antibiotics often decrease with increasing density of the starting microbial population , a phenomenon known as the inoculum effect [1] . While the IE is commonly attributed to enzymatic degradation of the drug [2 , 3]—with the classical example being the degradation of β-lactams by a β-lactamase enzyme—recent studies have pointed to a range of other potential mechanisms , including heat-shock mediated growth bistability [4] , intercellular signaling between resistant and sensitive cells [5] , and a decrease in per-cell antibiotic concentration and therefore the number of available drug molecules per cell [6] . However , the consequences of the inoculum effect for microbial population dynamics , the design of optimal treatment strategies , and the evolution of resistance are debated , in part because antibiotic efficacy is typically measured as a function of initial population ( inoculum ) size , not as an explicit function of population density . Specifically , the IE is commonly described as an increase in the minimum inhibitory concentration ( MIC ) of a drug as a function of inoculum size [1] . The MIC measurement requires cell populations to grow above a particular threshold level—measured by optical density or colony growth on solid media—meaning that cell density is changing by several orders of magnitude on the timescale of the experiment . In addition , the outcome measure in such experiments is binary , with cell populations either surviving or dying . Because clinical therapies often involve time-dependent drug doses that span both sub- and super-MIC levels [7] , it is difficult to systematically evaluate and optimize these dosing regimens without knowing the direct functional relationship between per capita growth rate , cell density , and drug concentration . Despite these inherent challenges , several recent studies have shown that a deeper quantitative understanding of the IE holds the promise of improved strategies for minimizing microbial growth or mitigating the evolution of resistance . For example , certain classes of protein synthesis inhibitors were recently shown to induce growth bistability , leading to a strong IE and population growth that depends markedly on the frequency of periodic drug dosing [4] . In cases where the IE can be traced to enzyme production , mathematical models suggest that optimized treatment protocols may resurrect otherwise ineffective first line drugs and prolong the efficacy of newer antibiotics [8] . In addition , intuitive treatment strategies—such as introducing an inhibitor of the enzyme—may actually promote , rather than inhibit , the spread of resistance [3] , while sophisticated measures of drug efficacy , such as the MIC of a single cell , may be required for predicting selection pressures driving resistance [9] . Even in the absence of detailed mechanistic information , pharmacological models based on classical MIC measurements suggest that otherwise successful therapies may fail to clear some infections [6] . Taken together , these studies suggest that the IE may play an important role in optimizing treatment strategies . Unfortunately , despite the promise of these emerging approaches , our quantitative understanding of the IE remains limited because the functional relationship between population density and drug inhibition is generally not known . To address these issues , we constructed multiplexed , computer-automated culture devices that allow us to monitor per capita growth in real time while maintaining bacteria at fixed population densities . The use of continuous culture devices in microbiology dates back more than 60 years [10] , and constant density bioreactors ( “turbidostats” ) have proven useful in a wide range of microbiology settings [11] . Unfortunately , commercially available bioreactors are expensive and somewhat inflexible , making the use of bioreactors far less common than traditional batch culture methods . Recently , however , a number of simple but elegant reactor designs has reinvigorated interest in continuous culture devices for applications ranging from synthetic circuit characterization [12] to cancer biology [13] , stochastic gene expression [14] , and antibiotic resistance under sustained selection pressure [15 , 16] . Inspired by these studies , we developed a simple , multiplexed turbidostat that would allow us to measure the growth of bacterial populations at fixed densities in response to a constant concentration of antibiotics . Our goal is to strip away technical ambiguities associated with classical measurements of the IE and provide the functional relationship between antibiotic inhibition and population density . In turn , we incorporate this density dependence into mathematical models to demonstrate its marked effects on population dynamics and potential treatments . While the IE has been observed across a wide range of bacterial species , here we focus on E . faecalis , a gram-positive species increasingly recognized as a clinically important pathogen [17] . E . faecalis underlies a host of nosocomial infections , and they rapidly acquire high-level resistance through genome mutations and horizontal gene transfer , making them important contributors to the spread of drug resistance in other bacteria [17–21] . In addition , E . faecalis are known to participate in community-like behavior , including cell-cell communication via two-component signaling [22–24] , sharing of extracellular DNA via fratricide [25 , 26] , phage-mediated gene transfer [27] , pheromone-induced quorum sensing [28 , 29] , and chromosomal transfer [30] . These studies offer molecular evidence of widespread intercellular interactions in E . faecalis populations , leading us to hypothesize that the behavior of these communities—in particular , their responses to antibiotic treatments—may depend strongly on population density . Indeed , the classical IE has been observed in E . faecalis for a wide range of antibiotics [31–33] , but little is known about how such density dependence might impact optimal therapies or ecological and evolutionary dynamics .
To determine explicitly the density dependence of single drug activity , we built multiplexed computer-automated microbial culture chambers ( Fig 1 , Figure A in S1 Text ) that maintain populations of E . faecalis at fixed densities while exposing them to different concentrations of antibiotics ( Methods ) . Cell density is monitored by light scattering and maintained through feedback control using a series of peristaltic pumps , which deliver fresh media and drugs and remove waste ( Fig 1A ) . Population growth ( specifically , per capita growth ) over time can then be calculated by monitoring the flow rates of the pumps ( Fig 1B–1D ) . By repeating this procedure at multiple cell densities—each in the exponential growth phase ( Fig 1D , inset ) —we can directly measure the density dependence of antibiotic inhibition for any given drug concentration ( Fig 1D ) . We measured the density dependence of nine antibiotics ( Table A of S1 Text ) , many of which represent current treatment options for E . faecalis infections [18] . Interestingly , we find that many antibiotics ( e . g . tigecycline ) exhibit density-dependent inhibition , while others ( e . g . ceftriaxone ) exhibit inhibition that is largely independent of density ( Fig 2 ) . While most drugs show decreased inhibitory activity as density increases—consistent with interpretations of the classical IE—ampicillin exhibits increased inhibition at higher densities . Of the nine drugs we tested , seven ( tigecycline , spectinomycin , daptomycin , nitrofurantoin , ciprofloxacin , linezolid , and doxycycline ) show a clear density dependent decrease in inhibition for at least one dosage , while ampicillin shows an increase in efficacy . In addition , one can estimate the drug’s half-maximal inhibitory concentration ( K ) at each density by fitting the growth , g , at each density to a sigmoidal dose-response function , g= ( 1+ ( D/K ) h ) −1 , where D is the drug concentration and h is a Hill-like steepness coefficient [34] . We find that six of the nine drugs show statistically significant change in K , and K can increase by a factor of three or more over this rather narrow density range ( Figure C of S1 Text ) . The observed density dependence of antibiotic inhibition may result from multiple different mechanisms , perhaps in combination [2–6] . Because E . faecalis are fermentative bacteria and pH is known to modulate antibiotic efficacy in a wide range of pathogens [31 , 35] , we asked whether density dependence could be partially explained by community-driven changes in local pH during exponential phase growth . While pH is known to modulate drug efficacy and to depend on bacterial growth phase , it has somewhat surprisingly not been linked with the IE , perhaps because IE measurements typically standardize only the initial pH ( which , like density , may change significantly during the experiment ) . Indeed , we found that exchanging regular media for highly buffered media ( see Methods ) modified the density dependence for some but not all drugs . For example , the density dependence was almost completely eliminated in highly buffered media for tigecycline and ampicillin ( Fig 3B ) . Note that growth in buffered and regular media did not differ significantly in the absence of drug ( Figure D in S1 Text , left panel ) . Furthermore , external modulation of pH recapitulated the effect in low density cultures ( Fig 3B ) . On the other hand , ciprofloxacin and spectinomyin continued to show significant , but decreased , density dependence in highly buffered media ( Fig 3C ) . Spectinomycin belongs to a class of protein synthesis inhibitors recently shown to exhibit an IE due to heat-shock mediated growth bistability [4] , and indeed spectinomcyin inhibition remained strongly density-dependent even in buffer , though the dependence was slightly weaker at high densities . At the other extreme , highly buffered media actually led to increased density dependence for ceftriaxone , indicating that changes in pH may counteract a second , unknown mechanism that decreases drug efficacy at high densities ( Figure D in S1 Text , right panel ) . Our results are also consistent with recent measurements in uropathogenic strains that show decreasing external pH will increase MIC for ciprofloxacin while decreasing MIC for ceftriaxone and ampicillin in disc diffusion assays [35] . These measurements indicate that density-mediated changes in pH represent one mechanism for the IE during exponential phase growth . For some drugs ( ampicillin and tigecycline ) , these pH changes account for the majority of the observed density-dependent changes , while for other drugs ( ciprofloxacin , spectinomycin , ceftriaxone ) , additional mechanisms appear to dominate . Because our experiments provide explicit measurements of per capita growth as a function of population density , we can quantify the impact of cell density on specific growth dynamics or treatment strategies . For example , what effect does the observed density dependence have on the time it takes a growing population to reach a particular threshold size in a constant drug environment ? To answer this question , we first developed a simple mathematical model for density-dependent turbidostat growth . Our model incorporates density-dependent inhibition as an increase or decrease in the effective drug concentration , a process governed by a phenomenological rate constant ε ( Methods; S1 Text ) . This change in drug concentration could represent , for example , the biochemical degradation of drug by enzymes , though recent work has shown that concentration rescaling can also describe changes in drug efficacy from a wide range of other sources , including genetic mutations and drug interactions [36 , 37] . We can directly estimate ε from the steady-state turbidostat data for each drug ( Table B of S1 Text ) ; ε quantifies the rate of change in the effective drug concentration—and therefore drug efficacy—due to cell density . Specifically , we assume that the effective drug concentration decays according to dD/dt = –εDnj , where ε is the phenomenological rate constant , n is the cell density , and j is a positive integer that describes the kinetic order of the decay ( j = 1 is linear with cell density , j = 2 quadratic; see S1 Text ) . We find both models ( j = 1 or j = 2 ) provide qualitatively accurate descriptions of density dependent turbidostat data for the drugs in this study ( Figure E of S1 Text ) . For 5 of the 9 drugs , the quadratic model ( j = 2 ) is quantitatively superior to the linear model ( j = 1 ) according to standard model selection methods ( S1 Text , including Table B ) , so we adopt that model in what follows ( though we note that both models provide qualitatively similar behavior; see Figure G in S1 Text ) . We stress that all information about density dependence is contained in the rate constant ε ( S1 Text , including Table B and Figures E , F ) . For example , the density dependence of tigecycline is well-described by a quadratic ( j = 2 ) decay with ε = 0 . 9±0 . 1 ( the units of ε are ( [time][cell density]2 ) -1 , where time is measured in units of inverse per capita growth rate without drug , and cell density is measured in units of OD ) . The model provides a good qualitative and quantitative fit to our growth rate measurements ( Figures E and F of S1 Text ) . To explore growth in a constant drug environment , we next incorporated these density effects into a classic logistic model of population growth [38] that includes a Hill-like dose-inhibition curve [34] ( Methods ) . We estimated parameters for logistic growth from drug-free growth curves and parameters of the ( low density ) dose-response function from a series of standard growth curves in early exponential phase ( OD<0 . 2 ) at different drug concentrations . For example , for tigecycline in regular media , the dose-response curve is described by half-maximal inhibitory concentration K0 = 19 . 3±1 ng/mL and Hill coefficient h = 2 . 1±0 . 1 . ( Figure F of S1 Text; estimates include ± standard error of fitting parameter; see also Methods ) . As in the turbidostat model , we incorporated changes in drug efficacy due to population density , n , by assuming that effective drug concentration changes according to the measured rate parameter ε . Using this simple mathematical model , we calculated the time required for a population with initial density of OD = 0 . 2 to reach a threshold of OD = 0 . 8 for different concentrations of tigecycline , which shows marked density-dependence ( ε = 0 . 9 ±0 . 1; see Fig 2 and Table B of S1 Text ) . Our model predicts that the time to threshold increases by a factor of approximately three as tigecycline concentration is increased from 0 to 2 . 5 K0 , and we were able to verify these predictions experimentally ( Fig 4A ) . To estimate the impact of density-dependence on time to threshold , we then repeated this calculation with ε = 0 ( no density dependence ) but all other parameters unchanged . Even over this somewhat limited density range ( 0 . 2 ≤ OD ≤ 0 . 8 ) , the density-dependence of drug efficacy can lead to a significant ( approximately two-fold ) decrease in time to threshold over a range of drug concentrations . Furthermore , because the density dependence of tigecycline is largely eliminated in highly buffered media ( Fig 3 ) , we experimentally confirmed these predictions by repeating the experiment with highly buffered media ( Fig 4 ) . Our results indicate that density-dependence can have a marked effect on growth dynamics , even in a simple constant drug setting . Because density dependence can impact constant drug treatments , we next asked whether the efficacy of such drugs could be improved by judiciously dosing the drug over time . To answer this question , we again consider a simplified scenario where a logistically growing population is treated with sub-MIC concentrations of a drug that exhibits density dependent efficacy ( measured again by ε ) . To be conservative , we restrict ourselves to populations growing in the approximate density regime we measured experimentally; specifically , we consider a population with a starting density of OD = 0 . 1—just below the densities measured—and limit growth by introducing a carrying capacity of C = 1 . 3 , just below that measured in our culture devices ( Fig 1D , inset ) . For simplicity , we do not consider clinically relevant periodic dosing schedules or super-MIC dosing regimens in this section , but we develop a more realistic model in a later section . Using this sub-MIC growth model , we compared a naïve dosing strategy where drug is added at a concentration D0 at time t = 0 to an optimal , step-like dosing protocol where drug is added at concentration ( D0/τ ) at time t = 0 and then “switched off” ( set to 0 ) at time τT , with τ ( 0≤τ≤1 ) chosen to minimize the total population density n ( T ) at the end of the treatment period T ( see Methods ) . In the absence of density dependence , both dosing strategies yield a time-averaged drug concentration 〈D〉=1T∫0TD ( t ) dt=D0 . Note that D ( t ) is the external concentration of drug at time t in the absence of density-dependent concentration changes; for simplicity , we do not consider the flow dynamics needed to establish and maintain such a concentration , though these details could be readily incorporated for any particular flow system . When a drug exhibits density-dependent inhibition ( ε>0 ) , the two dosing protocols often yield significantly different results . Specifically , the optimal step-like dosing protocol can decrease the final population size by more than 25% relative to the naïve protocol , depending on the strength of density dependence ( ε ) and the drug concentration <D> ( Fig 4B ) . For example , in the case of tigecycline ( ε = 0 . 9 ) , the optimal step-like dosing protocol decreases population size by more than 10% when D0/K0 ≈ 1 ( Fig 4B , main panel: dashed line and inset ) . These differences arise because the effective drug concentration decreases dramatically in growing populations due to density-dependent inhibition , rendering late-stages of the therapy less potent than expected ( Fig 4B , upper right inset ) . Our results suggest that even when density-dependent inhibition limits the efficacy of a given drug , an optimal step-like dosing strategy can mitigate the density dependence and reduce its impact ( Fig 4B , bottom right inset ) . The step-like protocol—which employs higher drug concentrations when drug is most effective—is reminiscent of the response-time strategy proposed to deal with cells expressing β-lactamase [8] , though here we derive it from experimental measurements and a phenomenological model , not a mechanistic model . The benefit of optimal treatment is typically maximized at <D> ≈ K0 , suggesting that this strategy may be useful when drug is limited because of cost , toxicity , acquired resistance , or delivery limitations . On the other hand , when D0 >> K0 , optimal dosing does not have a significant effect on the total population size because nearly complete inhibition can be achieved with either protocol . In the specific case of tigecycline , concentrations much larger than K0 should be achievable in vivo [39] . However , optimal dosing could be important after mutants acquire resistant or for other drugs with similar density dependence profiles . For example , E . faecalis V583 is more resistant to spectinomycin ( K0 = 102±1 ug/mL ) , and K0 is the same order of magnitude as in vivo serum levels in some cases [40] . As a whole , these calculations show that treatment strategies can , in principle , be improved using time-dependent dosing when density-dependent efficacy is significant . To gauge the impact of density dependence on more realistic treatment dynamics , we incorporated our measurements into a simple pharmacokinetic / pharmacodynamics ( PK/PD ) model of infection similar to those from [6 , 41] ( see Methods and S1 Text ) . Briefly , the model considers cyclic dosing of antibiotic with period T , leading to time-dependent effective drug concentration D ( t ) that starts at a maximum concentration D0 at the beginning of each dosing period . In the absence of density dependence , the concentration D ( t ) decays exponentially at a rate kd , which accounts for decay of drug in the clinical system of interest ( e . g . a patient ) . To incorporate density dependence , we again incorporate an additional decay term ( –εDnj ) in the equation for D ( t ) . Bacteria grow logistically and respond to drug according to a Hill-like pharmacodynamics function that ranges from maximum growth ( gmax , which we set equal to 1 without loss of generality ) to a minimum growth characterized by a maximum rate of kill ( gmin < 0 ) . The concentration , K0 , at which growth is zero is defined as the MIC ( by analogy with the half-maximal inhibitory concentration of the in vitro model , we choose the same notation K0 to now represent the MIC ) . Because we have not measured the maximum kill rates ( gmin ) for the drugs in this study—and because those rates may depend on specific environmental factors—we consider a wide range of kill rates both larger and smaller in magnitude than the native growth rate of the population . To study the effect of density-dependence on long-term treatment strategies , we analytically derived a full phase diagram ( Fig 5A ) describing the steady state behavior of the system as a function of initial drug dose D0 and population density n . The analytic calculations involve an adiabatic elimination of the fast timescale dynamics of D ( t ) ( S1 Text ) and are increasingly accurate in the limit of small maximum kill rate ( |gmin|<<1 . We find that for sufficiently small D0 , the infection cannot be cleared regardless of initial density . Similarly , for sufficiently large D0 , the infection will always be cleared . Interestingly , however , there is a range of drug concentrations where the outcome is bistable and high-density infections cannot be cleared ( S1 Text ) . Specifically , bistability occurs when ε>0 for initial drug concentrations in the range K0γ ( 0 ) < D0 < K0γ ( C ) , with γ ( n ) a density-dependent function that depends on the strength of the density dependence ( ε ) , the maximum kill rate ( gmin ) , the Hill coefficient ( h ) , and the dosing protocol ( specifically , the period T and the natural decay rate kd of the drug ) . For drug concentrations in the bistable region , infections with a density greater than some critical value cannot be cleared . These analytical approximations qualitatively describe numerical solutions of the model ( Fig 5 ) . For numerical analysis , we take T = 8 hours and kd = 0 . 5 hr-1 ( as in [6 , 41] ) , though similar qualitative results are obtained for a range of parameters ( see Figure G in S1 Text ) . In addition , for drugs whose efficacy increases with density ( ε<0 , such as for ampicillin ) , bistability cannot exist . Instead , the non-trivial fixed point becomes stable , leading to a drug-dependent stable population size ( Figure G in S1 Text ) . In that case , there is an intermediate range of drug concentrations for which population size will either increase or decrease from its initial value depending on the specific location in the phase diagram . Overall , the analytical approximation accurately predicts the entire phase diagram when the drug is characterized by a small kill rate ( Fig 5A ) but systematically underestimates the critical density for large kill rates ( Fig 5B ) . This underestimation arises , in part , because solutions that might decay in the long-time limit go extinct following a small number of doses when the maximum kill rate is large . In addition , large kill rates mean that n ( t ) is often changing on a timescale similar to that of D ( t ) , precluding the separation of timescales required for an accurate adiabatic approximation . To further explore these density-drive treatment effects , we numerically calculated the initial dose D0 required to clear infections at various densities ( Fig 5C ) . When there is no density dependence ( left ) , bistability is absent and infections are cleared for initial doses ranging from 5–20 times the MIC , depending on the kill rate of the drug . On the other hand , when modest ( middle ) or strong ( right ) density dependence exists—similar in size to the effects measured in this study—the treatment is bistable for a large range of concentrations , indicating that initial dosages hundreds or even thousands of times larger than the MIC can fail to clear particularly dense infections ( OD≈0 . 8 ) , especially when maximum kill rate is relatively small . Overall , our numerical and analytical results indicate that density-dependent drug efficacy can lead to bistability in treatment outcomes over a broad range of parameters , indicating that successful treatments based on traditional ( low density ) MIC measurements may often fail when applied to sufficiently dense populations .
Using computer-automated turbidostats , we have directly measured the effects of population density on antibiotic inhibition in E . faecalis . In contrast with traditional measurements of the IE , which provide only a binary measure ( survival , death ) of density dependence , here we provide unambiguous , fixed-density measurements of per capita growth rate as a function of density and drug concentration . Using this functional relationship—which is quantitatively described by a phenomenological rate parameter ε—we are able to optimize time-dependent drug dosing protocols , including a clinically-inspired pharmacology model that involves periods of both super- and sub-MIC drug exposure . More specifically , we show that density-dependence can significantly impact growth dynamics in constant drug environments , while time-dependent dosing strategies can partially mitigate these effects . Perhaps most strikingly , the observed density dependence can induce bistability of treatment outcomes over a wide range of antibiotic doses—in some cases , more than 1000 times the MIC of the drug—rendering otherwise successful protocols potentially ineffective for the treatment of dense populations . We analytically derive an expression for the size of the bistable region , and our analysis allows us to calculate the critical population density for treatment failure using common pharmacological parameters . Our results have implications for both basic and applied biology . From a basic science perspective , they underscore the notion that bacterial physiology cannot be cleanly divided into three growth phases: lag phase , exponential phase , and stationary phase . Instead , our measurements show that substantial changes in growth dynamics occur when bacteria are under stress , even in relatively narrow density ranges in exponential phase . In addition , we’ve shown that density-driven changes in pH modulate inhibition for some drugs . It is surprising that ( to our knowledge ) pH not been previously linked with the IE , because pH is known to modulate drug efficacy and to depend on bacterial growth phase . However , classical IE measurements involve standardizing only the initial pH of the media , but like density , that pH may change significantly during the experiment , perhaps obscuring any pH effects . Finally , in contrast to the classical inoculum effect , we found that the inhibitory effects of some drugs , such as ampicillin , increase at high densities . Overall , these findings represent a new entry in a growing catalog of density-mediated physiological changes in bacteria , including modulation of metabolism [42] , quorum sensing [43] , and synchronization of gene expression [44] . While it is well known that bacteria in different growth phases may respond differently to antibiotics—with biofilms [45 , 46] and dormant persister cells [47] representing two of the most salient examples—it is notable that such significant density-dependent effects can also arise in exponential phase , where growth in the absence of drug is approximately constant . On a practical level , our findings may offer a first step toward systematic optimization of antimicrobial-therapy based on population density . In addition , they may be important for industrial applications—including microbial fermentation [48]—where optimizing growth may also require consideration of density-mediated changes in population dynamics . However , it is important to keep in mind several limitations of our study . First , aside from the pH-mediated effects that explain a fraction of our results , we have not attempted to elucidate the molecular mechanisms responsible for the observed density dependence . Indeed , as suggested by other studies , it seems likely that such dependence—and the IE in general—results from a combination of factors and could differ for each drug [2–6] . For that reason , we focused instead on the functional implications of our measurements for population growth and drug response . Secondly , we measured population density using light scattering , which is a widely used method for inferring cell density but could conceivably conflate changes in cell number with changes in cell shape . Because growth rates are calculated from pump flow rates required to maintain constant light scattering intensity over a narrow range—not directly from time series of optical density—we do not anticipate significant artifacts from this limitation . We’ve also restricted our measurements to sub-MIC levels of antibiotics , where cell filamentation is minimal . Nevertheless , one should interpret the specific functional forms of the density dependence with some caution , particularly for drugs such as ciprofloxacin that can induce filamentation , as optical density may not be strictly proportional to viable cell number at high drug concentrations . In addition , the sensitivity of light scattering limits our measurements to relatively dense populations ( OD>0 . 1 ) , though measurements of the inoculum effect suggest that density-dependence may span several orders of magnitude [1 , 4 , 9] . It is not clear whether the density dependencies we observe may extend to lower densities . If so , that would dramatically increase the effect of density-dependence on time-to-threshold measurements , optimal dosing strategies , and PK/PD-based treatment protocols . We also stress that applying results from in vitro measurements to realistic in vivo scenarios is a significant challenge . In an attempt to emphasize density-dependent effects in a transparent setting , we incorporated our results into both a simple model for sub-MIC growth dynamics as well as a PK/PD model for antibiotic treatment . However , we acknowledge that these models neglect important factors such as the response of the host immune system [49] as well as other PD phenomena , including the post-antibiotic effect [50] , that may impact in-vivo treatments . Despite these caveats , our findings provide experimental and theoretical evidence that antibiotic efficacy depends dramatically on population density , and this dependence may , in turn , dictate the success or failure of standard treatment protocols . Bistability may be a particularly important consideration for drugs with low kill rates , and many drugs exhibit in-vitro kill rates on the lower end of the range considered here when applied to E . faecalis [51–53] . More generally , because our results are based on phenotypic measurements , they allow for optimization and prediction of drug dosing protocols even when molecular mechanisms are not fully known . Such mechanism-free phenomenological approaches are promising because of their potential applicability in more realistic clinical scenarios , where phenotypic measurements—for example , an estimate of the parameter ε—may be tractable even without mechanistic information [8 , 36 , 54–59] . Our hope is that these findings motivate continued quantitative studies on the IE and lay the groundwork for more detailed clinical models . Finally , our results also raise a series of fundamental questions at the interface of clinical , basic , and evolutionary microbiology . Perhaps most interestingly , the density dependence of antibiotic efficacy suggests that the relative fitness of drug-resistant cells may be intricately linked with the total density of the surrounding microbial population . As a result , the dynamics of evolving populations may exhibit many of the complex spatial and temporal dynamics described by evolutionary game theory [38] , even beyond the known cases that depend on enzymatic drug degradation [3 , 9] . At a practical level , these results also raise fundamental questions about the impact of antibiotic therapy on the evolution of drug resistance . In the long run , effective antimicrobial therapies must balance the inhibitory effects of treatment with its propensity to promote drug resistance , a topic of considerable recent interest [60 , 61] . Our results suggest that both of these factors could depend significantly on cell density , and we hope our work motivates continued efforts to understand the potentially complex interplay between population density , optimal treatments , and the evolutionary dynamics of resistance .
All experiments were performed with Enterococcus faecalis V583 , a fully sequenced vancomycin-resistant clinical isolate [20] . Cultures for experiments were taken from single colonies grown on plates of BHI agar and then cells were grown at 30°C in sterile Brain-Heart Infusion media ( Remel ) overnight before being diluted in fresh media 500X prior to experiment start . Experiments with pH-stabilized media had cells suspended in BHI supplemented with 50 mM Sodium Phosphate ( Fisher ) . Antibiotics used in the study include Ampicillin Sodium Salt and Doxycycline Hydrochloride ( Fisher ) ; Ciprofloxacin and Ceftriaxone Sodium Salt ( Acros Organics ) ; Nitrofurantoin and Spectinomycin Sulfate ( MP Biomedicals ) ; Tigecycline ( TSZ Chem ) ; Linezolid ( Chem-Impex Int’l LLC ) ; and Daptomycin ( Fisher ) . See Table A in S1 Text . We constructed customized continuous culture devices based on similar designs from several recent works [14–16] . In particular , the in-depth protocol from [16] was instrumental in our design process and is highly recommended to others looking to build customized instruments . Briefly , in our instrument , an infrared LED emitter ( Radioshack ) shines through the side of a flat bottom glass vial positioned on a multi-position magnetic stir plate . IR light passes through the culture solution ( approximately 15 mL ) and is scattered onto a photodiode detector ( Radioshack ) at an angle of approximately 130 degrees from the emitter . Photodiode voltage is directly proportional to the optical density of cells inside the vial . Populations are held at designated densities via custom MATLAB scripts , built on the Matlab Instrument Control Toolbox , which control a series of peristaltic pumps ( Boxer 15000 , Clark Solutions ) via multi-channel analog output device ( Measurement Computing USB-3103 and USB-3105 ) , which control pump speed , and a relay board ( Measurement Computing USB ERB-24 ) to allow us to turn pumps on or off . Voltages from up to 18 vials are simultaneously recorded from 18 photodiodes via a series of analog input boards ( Measurement Computing USB 1616FS ) . When the voltage is above a threshold value , a signal is sent to the relay board to turn on two pumps for that vial: one to introduce new media and the other to remove media and cells to keep vial volume constant . When voltage returns below a set value through dilution , the pumps are turned off . The CCD and all tubing is thoroughly cleaned with bleach and de-ionized water after each experiment to prevent contamination . At the onset of the experiment , overnight stationary phase bacterial cultures are diluted 500X and allowed to grow in the culture vials until a specific density is reached . At that point , drug is manually added at the desired concentration to both the culture vial and a connected chamber with fresh media . Drug concentration is the same in both the culture vial and the fresh media vial . Flow between the media chamber , the culture vial , and a waste vial is managed by a series of computer-controlled peristaltic pumps that maintain constant cell density ( Fig 1 ) . In our constant density turbidostat , the per capita growth rate , g , is equal to the relative dilution rate g = F/V , where F is the ( potentially time-dependent ) flow rate of the pumps and V is the total culture volume . We determined population growth rate by first setting the variable flow rate of each pump to a constant that we call Fmax ( Fmax ≈ 1 mL/min ) . We then switched the pumps “on” or “off” to maintain constant cell density according to the scheme in Fig 1A . To estimate population growth , we measured the state of each pump ( “on” or “off” ) at each time step ( Δt ≈ 1 . 5 seconds ) during this constant density phase . The fraction of time the pump is on , fon , is given by the number of time steps Non for which the pump was “on” divided by the total number of timesteps ( NTotal ) during a period of ~1–3 hours once population growth had reached a steady state . Flow rate F is related to fon according to F = fon Fmax , and hence growth is given by g=fonFmaxV . To verify that the growth was in steady state , we estimated real time growth rate ( Fig 1C ) using a moving average of the relative dilution rate ( fraction of time in “on” state ) over a window size of approximately 15 minutes . All values are relative to growth rates measured in the same vial and at the same density in the absence of drug ( doubling time of approximately 30–35 minutes ) . We consider two sources of experimental uncertainty in our measurements . First , there is uncertainty in the measurement of steady state growth rate in a single vial . The growth rate is given by g=fonFmaxV , where fon is the fraction of time the pump is “on” , Fmax is the maximum flow rate of the pump , and V is the volume of the culture . Fmax and V are assumed to be known exactly for a given vial , so the uncertainty comes from the measurement of fon . This uncertainty depends on the length of time over which growth rate is measured: very long measurements give an increasingly accurate picture of the growth rate . In steady state growth , we expect the uncertainty in the number of time steps for which the pump is on to be approximately ( Non ) 1/2 , so the fraction of time the pump is on is given by fon=NonNTotal±NonNTotal , where NTotal is the total number of time steps over which growth is measured ( and is known exactly ) . To estimate the relative per capita growth , we calculated fon with and without drug in the same vial and at the same density . The uncertainty in the ratio between fon with drug and fon without drug is then given by standard error propagation for a ratio of two variables . Specifically , the uncertainty in the ratio z of two uncorrelated variables x and y is given by δz=|z| ( δx/x ) 2+ ( δy/y ) 2 . The second source of variability comes from vial-to-vial variations . To account for these variations , we measured relative growth in duplicate or triplicate for each condition . The final growth rate measurement is the mean over these trials . Because each trial has an associated uncertainty , the total uncertainty is given by propagating the uncertainty from each individual measurement into a single uncertainty for the mean . This calculation draws on standard error propagation for sums of random variables . Specifically , for a sum of two uncorrelated random variables z = x+y , the uncertainty is given by δz=δx2+δy2 . From a growth measurement g with uncertainty δg , 95% confidence intervals are taken to be g ± 1 . 96 δg , which essentially assumes ( by the Central Limit Theorem ) that the total measurement noise is Gaussian . Measurement differences are deemed statistically significant if and only if 95% confidence intervals do not overlap ( this is a somewhat conservative estimate , as differences can be statistically significant even when confidence intervals do overlap , though the converse is never true; here we use the more stringent criteria of non-overlapping confidence intervals ) . We modeled density dependent drug inhibition in the turbidostat by assuming that the effective drug concentration D in each vial changes according to D˙= ( F/V ) ( Din−D ) −εDnj , where F is the flow rate of the pumps , V is the volume of media in the culture chamber , Din is the drug concentration in the feed chamber ( measured in units of K0 ) , and the last term ( -εDnj ) accounts for density-dependent effects on drug inhibition ( we restrict ourselves to models with j = 1 or j = 2 ) . The per capita growth g ( measured in units of growth rate in the absence of drug ) is related to the pump flow rate ( g = F/V ) and also to the effective drug concentration in the vial ( g ( D ) = ( 1+ ( D/K0 ) h ) −1 ) . Therefore , the steady state behavior is given by ( 1−gg ) 1/h=Din1+εnj ( g−1 ) , which can be solved numerically to give g for any combination of Din , h , j , and ε . We use this model along with standard model selection techniques [62] to estimate ε and determine the best decay model ( j = 1 or j = 2 ) for each drug directly from turbidostat data ( S1 Text , including Table B and Figures E and F ) . We modeled density dependent drug inhibition by extending a classic logistic model of bacteria growth to include density- and drug-dependent population growth . Specifically , we have n˙=g ( D ) ( 1−nC ) n ( 1 ) D˙=−εDnj ( 2 ) where n is the cell density , C is the carrying capacity of the environment , the factor of ( 1 − n/C ) corresponds to linear decrease in growth rate as n approaches the environmental carrying capacity , and g ( D ) is the per capita growth rate of the population exposed to drug at effective concentration D . Specifically , we have g ( D ) = ( 1+ ( D/K0 ) h ) −1 , with K0 the IC50 of the drug and h a Hill-like steepness coefficient , similar to dose response models in pharmacology [34] . To model the observed density dependence , we allow the effective drug concentration D to decrease ( ε>0 ) or increase ( ε<0 ) proportionally to Dnj , where j = 1 or j = 2 to indicate linear or quadratic decay with density , respectively ( Eq 2 ) . However , we stress that more general forms of drug decay may be better suited to some drugs . We choose the form above as a simple analytical parameterization that captures the primary features of our measurements ( Figure D of S1 Text ) . Alternatively , in cases where mechanistic understanding is available , the form of the decay may potentially be derived from a microscopic model ( for example , a linear function of n may be appropriate to describe density effects that arise from enzyme production ) . We also stress that K0 and h can be estimated from traditional growth curves , and only the parameter ε is estimated from our constant-density measurements . To investigate the effects of dosing schedule on sub-MIC growth dynamics , we compared total population size n ( T ) at the end of a treatment of length T for cells exposed to two different dosing protocols . In both cases , the dynamics are given by Eqs 1 and 2 , and T is chosen to correspond to the approximate time the population needs to reach carrying capacity in the absence of drug . In the absence of density dependence , both protocols yield an average effective drug concentration of 〈D〉=1T∫0TD ( t ) dt=D0 . Please see S1 Text for a detailed description of the PK/PD model as well as the derivation of the phase diagram for treatment bistability . Briefly , the PK/PD model describes antibiotic treatment that is dosed periodically with period T; at the beginning of each period , the effective drug concentration is set to D0 . During the dosing period , the drug decays according to Eq 2 ( due to density dependent inhibition ) , but we also include an additional linear decay ( with a rate kd ) to account for natural ( density independent ) drug decay in clinical setting . The effective concentration is then reset to D0 at the next cycle . The cell density n changes with instantaneous D ( t ) according to Eq 1 , but g ( D ) is now a monotonically decreasing function of D that crosses g ( D ) = 0 at D = K0 , where K0 is defined as the MIC value , and asymptotically approaches some maximum kill rate ( gmin<0 ) . The model can be easily solved numerically using any ODE integration software ( we used Matlab’s ode45 function ) . To study the long-time behavior of the system , we assumed n ( t ) is approximately constant on the timescale of D ( t ) . We then averaged the dynamics of D ( t ) over one period , yielding an approximation equation for n valid in the limit of small gmin . We then performed linear stability analysis on the fixed points of the averaged equation to analytically derive phase diagrams ( S1 Text ) . | The pace of antibiotic discovery has rapidly slowed in the last few decades , creating an urgent need to reevaluate and optimize therapies based on current drugs . In this work , we combine quantitative laboratory experiments on bacterial populations with mathematical models of antimicrobial therapies to demonstrate that bacterial populations of different sizes ( densities ) may respond very differently to the same antibiotic treatment . As a result , otherwise successful antibiotic treatments may fail to eradicate bacterial infections that have reached a critical population density . Our findings indicate that it is important to consider the size of a bacterial infection when designing effective antimicrobial therapies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"classical",
"mechanics",
"pathology",
"and",
"laboratory",
"medicine",
"fluid",
"mechanics",
"pathogens",
"drugs",
"microbiology",
"antibiotic",
"resistance",
"pharmaceutics",
"antibiotics",
"enterococcus",
"pharmacology",
"population",
"biology",
"bacteria",
"bacterial",
"pathogens",
"population",
"density",
"antimicrobial",
"resistance",
"medical",
"microbiology",
"microbial",
"pathogens",
"fluid",
"dynamics",
"enterococcus",
"faecalis",
"continuum",
"mechanics",
"physics",
"flow",
"rate",
"population",
"metrics",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"drug",
"therapy",
"drug",
"interactions",
"organisms",
"population",
"growth"
]
| 2016 | Population Density Modulates Drug Inhibition and Gives Rise to Potential Bistability of Treatment Outcomes for Bacterial Infections |
Soil- and waterborne bacteria such as Pseudomonas aeruginosa are constantly challenging body surfaces . Since infections of healthy skin are unexpectedly rare , we hypothesized that the outermost epidermis , the stratum corneum , and sweat glands directly control the growth of P . aeruginosa by surface-provided antimicrobials . Due to its high abundance in the upper epidermis and eccrine sweat glands , filaggrin-2 ( FLG2 ) , a water-insoluble 248 kDa S100 fused-type protein , might possess these innate effector functions . Indeed , recombinant FLG2 C-terminal protein fragments display potent antimicrobial activity against P . aeruginosa and other Pseudomonads . Moreover , upon cultivation on stratum corneum , P . aeruginosa release FLG2 C-terminus-containing FLG2 fragments from insoluble material , indicating liberation of antimicrobially active FLG2 fragments by the bacteria themselves . Analyses of the underlying antimicrobial mechanism reveal that FLG2 C-terminal fragments do not induce pore formation , as known for many other antimicrobial peptides , but membrane blebbing , suggesting an alternative mode of action . The association of the FLG2 fragment with the inner membrane of treated bacteria and its DNA-binding implicated an interference with the bacterial replication that was confirmed by in vitro and in vivo replication assays . Probably through in situ-activation by soil- and waterborne bacteria such as Pseudomonads , FLG2 interferes with the bacterial replication , terminates their growth on skin surface and thus may contributes to the skin’s antimicrobial defense shield . The apparent absence of FLG2 at certain body surfaces , as in the lung or of burned skin , would explain their higher susceptibility towards Pseudomonas infections and make FLG2 C-terminal fragments and their derivatives candidates for new Pseudomonas-targeting antimicrobials .
Human skin and mucosal surfaces harbor and encounter a high number of diverse microorganisms [1 , 2] , but are rarely infected . Apart from the physical skin barrier of the stratum corneum ( reviewed in ref . [3] ) , a “chemical barrier” , consisting of various antimicrobial peptides and proteins ( AMPs ) , was described to act as innate defense barrier and contributes to the control of microbial growth at body surfaces [4 , 5] . AMPs can be expressed constitutively or induced by microbe-associated molecular patterns—even at a distance across the physical barrier [6] . In recent investigations a significant role was also ascribed to AMPs as coordinators or alarmins of the adaptive immune response [7] . Despite the increasing number of identified AMPs , there is a growing number of reports describing bacterial defense and immune escape mechanisms to these proteins [8–11] implying that epithelial layers possess additional defense strategies . Humans are always in contact with microbes , in particular with those from aquatic environment and soil . Especially ubiquitous Pseudomonads are highly abundant on human skin [12] but cause skin infections only when the cutaneous barrier is disturbed , such as in toe web infections [13] or hot tub folliculitis [14] . More severely , under conditions where the cutaneous barrier is completely missing as in burn wounds , Pseudomonas aeruginosa is a major cause of morbidity and mortality [15] . As waterborne bacteria , Pseudomonads would thrive at healthy human skin surfaces primarily on areas with an adequate content of moisture and humidity . Besides mucosal epithelia , suitable areas seem to be the lumen and ducts of eccrine sweat glands of skin . Eccrine sweat glands are producing the sweat-specific antimicrobial peptide dermcidin [16] . Interestingly , whereas dermcidin shows antimicrobial activity against various bacteria , there is hardly any against P . aeruginosa [17] . This surprising finding suggests that sweat glands and the epidermis are producing additional , yet unknown factors , which control the growth of especially Pseudomonas spp . . The sole of the foot is rich in sweat glands and extracts of plantar stratum corneum have been shown to be a rich source of antimicrobial peptides [18–21] . Heparin-bound compounds of these extracts were subjected to high performance liquid chromatography and peptide fragments of filaggrin-2 ( FLG2 ) , formerly also known as ifapsoriasin [22 , 23] , were identified by MS/MS analyses in antimicrobially active fractions . To test our hypothesis if protein fragments of FLG2 contribute to the resistance of human skin against P . aeruginosa infections , recombinant FLG2 protein fragments of the 248 kDa full length protein , representing typical regions of the protein , were generated and their antimicrobial properties investigated . Our results show that C-terminal FLG2 protein fragments are able to kill Gram-negative bacteria , and here most efficiently P . aeruginosa , by interfering with the bacterial replication .
Immunodetection revealed the presence of FLG2 in stratum granulosum and stratum corneum [23–26] ( S1 Fig ) , and FLG2 fragments were also detected in stratum corneum extracts [23] . To test the hypothesis that , apart from its physical barrier function , FLG2 contributes to the innate skin defense , the following FLG2 protein fragments were recombinantly generated: a protein fragment consisting of both the last repeat domain B14 together with the C-terminus ( FLG2-4 , amino acids 2244–2391 ) , the FLG2 B-repeats B13 ( FLG2-B13 , aa 2172–2246 ) and B14 ( FLG2-B14 , aa 2247–2321 ) , as well as the FLG2 C-terminus ( FLG2-C-Term , aa 2322–2391 ) ( Fig 1 ) . In addition , corresponding polyclonal antibodies were generated by immunizing goats with recombinant FLG2 fragments and purified by protein G- and affinity chromatography . Immunohistochemistry revealed strong FLG2 immunostaining in the thickened stratum corneum of palmar ( S1 Fig and [23] ) and plantar skin [23] , areas that are frequently in contact with the environment and particularly rich in eccrine sweat glands . To investigate whether FLG2 protein is also present in eccrine sweat glands , plantar biopsy material was analyzed with antibodies directed against C-terminal FLG2-4 and intense immunoreactivity was found in sweat gland cells ( Fig 2A and 2B ) . Interestingly , glands showed also luminal FLG2 immunoreactivity with strong staining associated with the luminal surface of the ducts as well as luminal particle-like structures . To investigate whether FLG2 is also present in sweat , sediments as well as supernatants of collected healthy donors' sweat were separately analyzed by Western blotting ( Fig 2D ) . FLG2-4 staining was almost exclusively detected in DTT-treated sediments with the most intense bands corresponding to 70 and 140 kDa , suggesting that sweat gland FLG2 is crosslinked to particular matter and needs to be processed to release C-terminus-containing FLG2 protein fragments . Especially plantar and palmar skin is frequently in contact with soil- and/or waterborne microbes . Hence , it is tempting to speculate that these skin areas are protected from infections by Pseudomonads by means of water-insoluble , stratum corneum-associated antimicrobial compounds . To test this hypothesis , the antimicrobial activity of the different recombinantly expressed FLG2 protein fragments was analyzed in the radial diffusion assay ( RDA ) system and identified the FLG2 C-terminal protein fragment ( FLG2-4 ) as potent Pseudomonas aeruginosa ATCC 33354 killing factor with a minimal effective concentration ( MEC ) of 0 . 4 μM in this assay system ( Fig 1 ) . Whereas FLG2-4 was also bactericidal for Escherichia coli ATCC 11775 ( MEC: 2 . 4 μM ) , no killing activity was seen for the yeast Candida albicans ATCC 24433 and the Gram-positive Staphylococcus aureus ATCC 6538 ( Fig 1 ) . FLG2-4 is composed of the FLG2 repeat domain B14 and the FLG2-C-terminus . In order to identify the structural elements responsible for bactericidal activity within the FLG2-4 fragment , FLG2-B14 , the FLG2-C-terminal fragment ( C-Term ) , and the second to last repeat , FLG2-B13 , were analyzed in the RDA system . The antimicrobial activity against the tested bacteria was solely located within the FLG2 C-Term protein fragment , although the MECs for E . coli ATCC 11775 ( 13 . 0 μM ) and P . aeruginosa ATCC 33358 ( 3 . 3 μM ) were about 5 . 5–8 . 3 times higher than seen for the FLG2-4 ( Fig 1 ) . With the hypothesis that the C-terminal fragment FLG2-4 is active against various soil- and waterborne bacteria , further P . aeruginosa strains , including P . aeruginosa ATCC 10145 , 33348 , 333358 , 39324 , PAO1 , the Cystic Fibrosis ( CF ) clinical isolates CF 636 , 640 , 645 , 646 , and other Pseudomonas sp . as P . stutzeri RV A2/1990 , P . syringae ATCC 10205 , P . paucimobilis RV A2/1994 , P . fluorescens ATCC 49323 , and P . putida RV A1/2000 were tested . Antimicrobial activity was observed against all tested strains with MECs between 0 . 1 and 7 . 0 μM ( S1 Table ) . Most of the cystic fibrosis ( CF ) clinical isolates as well as the P . aeruginosa strain ATCC 39324 , P . fluorescens ATCC 49323 , and P . putida RV A1/2000 showed turbidity within the clearing zones . Since all of the subsequent experiments were performed in liquid systems , we also analyzed antimicrobial activity of FLG2-4 against E . coli ATCC 11775 and P . aeruginosa ATCC 33354 in the microdilution assay system; the LD90 of 0 . 8 μM and 0 . 2 μM , respectively , confirmed bactericidal activity found in the RDA ( S2 Table ) . In summary , these results indicate that the C-terminal peptide fragment of FLG2 , FLG2-4 , is a potent , preferentially Pseudomonas ssp . killing antimicrobial peptide . Whereas the antimicrobial activity is located solely within the C-terminus , the presence of the last B-repeat ( B14 ) increases the killing potency of the protein . Immunoreactive FLG2 fragments are present in sweat and extraction was only possible in the presence of the reducing agent DTT ( Fig 2D ) . Also , FLG2 extraction from stratum corneum was seen to be possible in the presence of SDS and DTT [23] , non-ionic detergents , 8M urea or EDTA [24] . The observation of smaller immunostained bands , apart from a band of the expected size of the full length protein , suggests that FLG2 is cleaved within the transition zone from stratum granulosum to stratum corneum and sweat glands , presumably necessary for generation of soluble antimicrobially active fragments . P . aeruginosa is metabolically versatile [27] and known for the degradation of even insoluble substances [28] . Therefore , P . aeruginosa was cultivated with plantar stratum corneum ( PSC ) and the release of immunoreactive C-terminal FLG2-fragments was monitored . In this experimental system , P . aeruginosa was able to break down FLG2 fragments from stratum corneum that were recognized by the specific C-terminal directed FLG2-4 antibody ( Fig 3A ) . Interestingly , the major bands between 50 and 70 kDa seem to resemble that of DTT-treated sweat ( compare Figs 3A and 2D ) . The efficacy of antimicrobials highly depends on the type of interaction with their target microbes . To analyze whether FLG2 breakdown products of PSC associate with or are taken up by the pathogen , P . aeruginosa was spatially separated from PSC by a transwell ( pore size 0 . 45 μm ) . After incubation , bacterial cells were taken from the upper transwell compartment , washed , and subjected to immunofluorescence analyses . Despite the separation of P . aeruginosa and stratum corneum , bacterial cells showed strong FLG2-4 immunoreactivity ( Fig 3B , left panel ) when compared to untreated cells ( Fig 3B , right panel ) . To get further insight into the mechanism by which FLG2-4 kills P . aeruginosa , morphology of FLG2-4-exposed P . aeruginosa were examined by transmission electron microscopy ( TEM ) . As shown in Fig 4A , 4B and 4C , P . aeruginosa exhibits massive bleb formation already within 30 min of treatment , when compared to control bacteria ( Fig 4D ) . Prolonging exposure to FLG2-4 resulted eventually in a significant destruction of the bacterial cells ( S2 Fig ) . To localize FLG2-4 during treatment , confocal laser scanning microscopy was performed . FLG2-4 immunofluorescence could be detected mainly at or within the membrane of the bacteria ( Fig 4E ) and was consistent with the subcellular fractionation of FLG2-4-treated bacteria ( Fig 4F ) . Nonetheless , confocal laser scanning microscopy revealed also FLG2-4 immunoreactivity within the blebs released by P . aeruginosa ( S3 Fig ) . Bleb formation of P . aeruginosa provides an active external decoy to trap cell surface-acting antimicrobial agents [29] , and is described as a sign of perturbations in the bacterial envelope [30] or DNA damage [31] . Analyzing the effect of FLG2-4 treatment on the bacterial outer membrane , a lysozyme lysis assay was performed in the presence of chloramphenicol to prevent self-induced lysis of P . aeruginosa . As shown in Fig 5A , FLG2-4 treatment did not impair the integrity of the outer membrane by means of lysozyme penetration . FLG2-4-treated bacteria do not exhibit a difference in the OD595nm compared to control cells . Using the pore forming antibiotic Polymyxin B [32 , 33] , treated bacteria showed a marked decrease in OD already within less than 10 min . Furthermore , the uptake of the membrane-impermeable DNA-stain SYTOX Green was only slightly increased during the first 10 min after FLG2-4 treatment ( Fig 5B ) , while Polymyxin B treatment showed an increased uptake of SYTOX Green almost immediately after addition to the bacterial cells . Since the observed blebbing might also be a reaction to DNA damage ( 31 ) , the ability of FLG2-4 to interact with DNA was analyzed . Electromobility shift assays showed that the basic protein FLG2-4 ( pI: 10 . 18 ) is able to shift supercoiled as well as linearized plasmid DNA , when using concentrations in the range of the LD90 for the tested Gram-negative bacteria ( Fig 6A ) . Interestingly , the antimicrobially active C-terminal peptide ( pI: 9 . 8 ) also interacted with and shifted DNA ( S4 Fig ) , but even though having almost identical pI values , neither the antimicrobially inactive FLG2-B13 ( pI: 9 . 95 ) nor FLG2-B14 peptide ( pI: 8 . 59 ) caused a DNA shift ( S4 Fig ) . To prove an in vivo interaction of FLG2-4 with the bacterial DNA , treated P . aeruginosa were fixed with formaldehyde and subsequently DNA was isolated . FLG2-4 crosslinking to DNA was monitored by dot-blot analyses using the anti FLG2-4 antibody and reveal that FLG2-4 was detected within the DNA fraction ( Fig 6B ) . Many cationic proteins display binding affinities to DNA , e . g . LL37 [34] , thereby affecting also replication events [35] . Therefore , the impact of FLG2-4 was investigated in a simplified model of bacterial DNA replication , the polymerase chain reaction ( PCR ) . Using conventional Taq Polymerase , PCR analyses were performed in the absence and presence of FLG2-4 . Whereas concentrations of 0 . 32 μM and 0 . 16 μM did not affect the reactions , in samples containing FLG2-4 at concentrations reflecting the LD90 for E . coli ( 0 . 78 μM or 250 ng/20μl ) , no amplicon was detected using GAPDH primers on cDNA ( Fig 7A ) . Even coincubation of the FLG2-B13 and FLG2-B14 protein fragments did not hamper the FLG2-4 mediated PCR inhibition ( S5A Fig ) . Merely increasing the polymerase concentration in the PCR reactions could partly abolish the FLG2-4 inhibition ( S5C Fig ) . Neither the structurally unrelated AMP hBD-2 ( corresponds to ~1 . 25 fold of LD90 for E . coli [36] ) nor BSA did affect PCR reactions at the applied concentrations ( S5B Fig ) . To test the effect of FLG2-4 on bacterial replication in vivo , a plasmid-based replication assay was developed . The assay based on the finding that following chloramphenicol treatment , the FLG2-4-induced lysis of bacteria is similarly prevented ( Fig 5A ) as known for some quinolone antibiotics [37 , 38] . Since in the presence of the translation-inhibitor chloramphenicol bacterial chromosomal replication is also blocked [39] ( S5D Fig ) , plasmid replication of pBR322 was monitored in E . coli ATCC 11775 over time . This plasmid , which contains a closely related origin of replication ( ori ) of the ColE1 , the pMB1 ori , replicates in an autonomous , self-controlled way [40] . While plasmid concentration increased over time in bacteria treated solely with chloramphenicol , addition of FLG2-4 in concentrations reflecting its LD90 ( 0 . 8 μM ) impede plasmid replication ( Fig 7B and 7C ) .
Human skin acts as barrier between the outer environment and the susceptible body interior . To manage this task , the cells of the epidermis , the keratinocytes , undergo a complex and tightly regulated process called epidermal terminal differentiation . Eventually this process causes the formation of a protective physical skin barrier , the stratum corneum . This cell layer , consisting of dead cornified keratinocytes , called corneocytes , and lipids , blankets the vulnerable layers of living keratinocytes . The keratinocytes themselves are able to induce more complex innate immune responses , e . g . by providing several constitutive and inducible antimicrobial peptides to protect skin from invading microbes [5 , 41 , 42] . Eccrine sweat glands support the maintenance of the barrier by secreting antimicrobial fragments of dermcidin [43] or supporting the process of re-epithelialization during wound healing [44] . Upon endo- or exogenous breakdown of stratum corneum compounds , C-terminus-containing FLG2 protein fragments are generated beyond the living epidermis . Here we could show that C-terminal FLG2 fragments exhibit antimicrobial activity , preferentially against P . aeruginosa and other water or soil bacteria , which are constantly challenging our skin . In contrast to the mechanisms assigned to many other AMPs [45–48] , the bactericidal mechanism of FLG2-4 is not based on pore formation by inserting into membranes . Instead , our results show that association of FLG2-4 with the cytoplasmic membrane and impediment of the bacterial replication machinery causes the bacterial death . Similar to the quinolone antibiotics that interfere with the bacterial replication by acting on its DNA gyrase finally causing the bacterial death [49 , 50] , our results suggest that FLG2-4 impair the DNA polymerases processivity , thereby stalling bacterial replication and thus preventing the survival of certain pathogens on or in human skin . As the majority of the proteins involved in terminal differentiation and formation of the cornified cell envelope , the S100 fused-type protein ( SFTP ) FLG2 is encoded within the epidermal differentiation complex on chromosome 1q21 . 3 [51] . In human epidermis , FLG2 colocalizes partially with filaggrin , one of the best characterized member of the SFTP family [23 , 24 , 26] . Similar to filaggrin , or its precursor profilaggrin , FLG2 is proteolytically processed during epidermal terminal differentiation [23] and possesses similar biochemical properties [24] . The biological functions of FLG2 are mostly unknown . The N-terminal part of FLG2 , including the A-type repeats , shares identity to the SFTP member hornerin , whereas the C-terminal part , including the B-repeats , rather shows similarities to filaggrin [23] . For the repetitive region of the murine B-repeats a keratin-bundling activity similar to filaggrin subunits has been shown [52] , assuming a similar function for the human FLG2 B-repeats . The antimicrobially active C-terminal region , however , does not belong to the repetitive sequences . About the function of the FLG2 C-terminus just as little is known as for the filaggrin C-terminus; it was suggested that it is necessary for proper processing of the precursor profilaggrin into smaller subunits [53 , 54] . Maybe an equally important role , in addition to the antimicrobial properties , can be ascribed to the C-terminus of FLG2 . As potential crosslinking partner of the cornified cell envelope , FLG2 seems to be retained within the stratum corneum ( S1 Fig ) and is somehow particle-bound and hardly water-soluble in eccrine sweat ( Fig 2 ) . FLG2 is detectable in the stratum corneum , especially of palmar and plantar skin [23] . These skin locations , enriched in eccrine sweat glands , in principle provide proper habitats for soil- and waterborne pathogens as Pseudomonas spp . . Even though our hands and feet are frequently in contact with the outer environment and therefore with a multiplicity of these microbes , infections are virtually excluded . The adaptable , non-stringent metabolic requirements enable Pseudomonas spp . to occupy a diversity of biological niches . Regarding its keratinolytic [55] and ceramidase activity [56] as well as its frequent detection on human skin [12] , it is not astonishing that P . aeruginosa is able to break down C-terminus-containing FLG2 fragments from stratum corneum ( Fig 3A ) . More strikingly , these fragments seem to accumulate in bacterial cells ( Figs 3B , 4E and 4F ) . Many antimicrobial peptides interact with the bacterial membrane , leading to pore formation and finally leakage of the intracellular components [57] . Ultrastructural analyses as well as confocal microscopy revealed bleb formation on the bacterial surface shortly after treatment with the C-terminal FLG2-4 fragment . Further , FLG2-4 immunoreactivity could be detected at these blebs and surrounding the cells ( S3 Fig , Fig 4E ) . Due to the high energy consumption , bleb formation of bacteria is a short term bacterial stress response to haul out detrimental factors and thus diminishing the efficacy of AMPs [58] . Especially antimicrobial factors that cause perturbations in the bacterial cell envelope or induce DNA damage , are known to trigger this stress response [29 , 59] . Subfractionation of FLG2-4-treated P . aeruginosa particularly showed FLG2-4 immunoreactivity in the inner membrane fraction ( Fig 4F ) . In addition , FLG2-4 interacts with bacterial DNA ( Fig 6 ) . These results suggest that FLG2-4 might colocalize and interfere with the membrane-associated bacterial replication complex . The detailed molecular mechanisms of bacterial replication are not fully understood , but replication always starts at one particular site , the origin of replication or oriC , and it is a multiprotein process [60] . During the initiation of replication in bacteria , the membrane-associated protein DnaA binds the oriC and subsequently enforces unwinding and opening of the double helix , finally enabling the assembly of the replisome [61–64] . In further steps , helicases and topoisomerases unwind and relax the coiled DNA helix . Once the DNA is melted and primed , the general DNA replication process starts . The presence of FLG2-4 in the inner membrane fractions and its ability to bind DNA indicated an interference with the DNA polymerase reaction , the central stage of replication . In concordance , in a first , simplified model of replication , FLG2-4 was able to stall the in vitro polymerase chain reaction in a dose-dependent manner ( Fig 7A ) . This effect could be rescued partly by increasing the polymerase concentration ( S5C Fig ) . To monitor the effect of FLG2-4 on in vivo replication , an assay was developed that avoided the stress-induced disintegration of the bacterial cell shortly after FLG2-4 treatment by addition of the translation inhibitor chloramphenicol . Since chloramphenicol also blocks the replication of the bacterial chromosome [65] , the autonomous replication of the transformed pBR322 plasmid in the presence of chloramphenicol [39 , 66] was investigated and , in support of our hypothesis , it was found to be significantly reduced upon FLG2-4 treatment in E . coli ( Fig 7B and 7C ) . Taken together , these results lead to a model where Pseudomonas spp . , when encountering human skin , are capable of degrading available substrates ( e . g . FLG2 among other proteins of the stratum corneum ) , and in turn generate C-terminus-containing FLG2 fragments . These fragments accumulate at the bacterial surface , eventually integrate into the bacterial inner membrane and interfere with the replication machinery . In consequence , the inhibition of the replication induces a bacterial SOS response , similar as observed for certain quinolone antibiotics [37 , 38] that ultimately causes bacterial cell death [49 , 50] . The present study indicates that C-terminal FLG2 fragments might be promising alternatives to traditional antibiotics used to cure especially Pseudomonas infections . Considering the fact that FLG2 transcripts could not been detected in the human lung [23] ( nor its mouse orthologue in the murine lung [67] ) might emphasize a potential application of C-terminal FLG2 fragments in the eradication of early P . aeruginosa infection in cystic fibrosis in order to prevent a chronic infection state and its clinical consequences .
The expression and protein synthesis was performed as described before [23] . The protein fragment for the C-terminal fragment of FLG2 , FLG2-4 , was produced using the pETSUMO expression system ( Invitrogen ) , whereas the other C-terminal fragments , FLG2-B13 , -B14 , and-C-Term were produced using the pSUMO3 expression system ( LifeSensors ) . Immobilized metal ion affinity chromatography-purified SUMO-FLG2 fusion proteins were cleaved using SUMO protease 1 and 2 ( LifeSensors ) , respectively , and the FLG2 proteins were subsequently purified by reversed phase HPLC . Polyclonal goat-anti-FLG2-4 antibodies were generated and purified by affinity chromatography with the respective protein fragment , as described earlier [23] . All experiments were performed according to the Declaration of Helsinki protocols and under protocols approved by the Ethics Committee at the Medical Faculty of the Christian-Albrechts-University , Kiel ( AZ 104/06 ) . Sweat and stratum corneum samples derived from healthy volunteers , skin explants for histology derived from patients undergoing surgery , and were obtained after written informed consent . Commercial production of customized polyclonal antibodies in the goat was in accordance with the German Animal Welfare Act ( §10a ) . The protocol was approved by the ethics committee of the “Landesamt für Landwirtschaft , Lebensmittelsicherheit und Fischerei Mecklenburg-Vorpommern“ , Germany ( AZ: LVL M-V 7221 . 3–2 . 5-010/2004 ) . FLG2 protein localization was visualized in skin paraffin sections as described previously [23] . Microbes were cultivated in either brain heart infusion medium ( BHI ) , lysogeny broth ( LB ) , tryptic soy broth ( TSB ) or M9 medium [68] supplemented with 0 . 5% casamino acids and 1μg/ml thymine . If not otherwise stated , microbes were incubated under shaking conditions ( 37°C at 170 rpm ) . The purified proteins were tested against different bacteria in a modified radial diffusion ( RDA ) or microbroth dilution assay [69] . The underlay agarose used for the RDA was inoculated with approximately 105 colony forming unit ( CFU ) per ml , the protein fragments were applied using a twofold dilution series . Evaluation was performed by calculating the linear regression over each measured clearing zone replicate ( GraphPad Prism version 5 . 04 for Windows ) . The x-intercept , that represents the "minimal effective concentration" ( MEC ) according to [69] , was calculated as a regression over all measured clearing zone units . The x-axis intercept and its 95% confidence interval are given in μM for each microorganism . For the microbroth dilution assay , bacteria were grown to early log-phase in BHI , washed with 10 mM sodium phosphate/1% TSB , and adjusted to a concentration of 104−105 CFU/ml , followed by an incubation with FLG2-fragments for 2h at 37°C . The samples were plated out on BHI agar in appropriate dilutions and CFU were counted after an overnight incubation at 37°C . The LD90 defines the protein concentration killing 90% or more of the bacteria . Pooled powdered stratum corneum ( SC ) ( ~15 mg ) was exposed to UV light for 60 min to minimize microbial load . Thereafter the SC was washed once with sterile H2O . Remaining liquid was removed by a centrifugation step ( 2000 x g , 5 min ) and 200 μl of a logarithmic-grown P . aeruginosa culture ( OD600 0 . 2 ) were added . Breakdown products were analyzed by Western Blot using goat-anti-FLG2-4 as described [23] . Autoclaved and washed plantar stratum corneum ( ~15 mg ) was placed in a 6-well-microplate . 100μl of a logarithmic-grown P . aeruginosa culture were transferred onto the stratum corneum , separated by a transwell ( 0 . 45μm ) . After incubation for 30 h at 37°C , bacteria were washed and heat-fixed to glass slides . After blocking with Roti-Block ( Roth ) , immunoreactive FLG2 was detected with goat-anti-FLG2-4 antibody ( 0 . 5 mg/ml , 1:100 ) and monitored using an AlexaFluor488-coupled chicken-anti-goat IgG ( Invitrogen ) secondary antibody . Slides were analyzed using a Zeiss Axioskop . Logarithmic grown P . aeruginosa ATCC 33354 were washed and concentrated to an OD600nm of 4 in 10 mM sodium phosphate/1% TSB . The amount of FLG2-4 used was about 2 x 107 molecules per colony forming unit ( CFU ) . Samples were incubated at 37°C for 30 min , 90 min , and 180 min . After incubation , the samples were fixed in 2 . 5% glutaraldehyde . Afterwards , the pelleted samples were resuspended in 2% Noble Agar , washed several times with PBS pH 7 . 4 , and incubated with 1% osmium tetroxide in PBS for 45 min . Again , this was followed by several washings in PBS . Samples were then dehydrated in an ascending graded ethanol series . For embedding , the EtOH was replaced stepwise by a polyhydroxy-aromatic acrylic resin ( LR White ) , starting at a resin:EtOH ratio of 1:2 , followed by 1:1 , 2:1 , and three times in resin only , each for 30 min . Finally , samples were embedded in resin at 60°C . The hardened resin was then cut into 5 nm sections and transferred onto a grid . The samples were analyzed with a transmission electron microscope ( Philips TEM 208 or FEI Tecnai G2 Spirit BioTwin ) . P . aeruginosa ATCC 33354 was cultured as described for electron microscopy and FLG2-4 concentration was adjusted to an identical molecule:CFU ratio . Bacteria were incubated for 30 min , 60 min , and 90 min with the protein , washed twice with PBS and resuspended in 2% formaldehyde/0 . 2% glutaraldehyde/PBS to a theoretical OD600nm of 1 . The cells were then fixed for 1 h at room temperature and subsequently stored overnight on glass slides at 4°C in a wet chamber . The next day , the samples were dried at room temperature , washed three times with PBS , and incubated for 5 min in 50 mM NH4Cl in PBS . Two washing steps in PBT ( PBS/0 . 1% BSA+0 . 05% Tween 20 ) were followed by a 5 min incubation in 50 mM Tris/1 mM EDTA . For blocking , the samples were treated for 20 min with 0 . 1% BSA/0 . 2% glycine/Tris buffered saline . Incubation with goat-anti-FLG2-4 ( 0 . 5 mg/ml , 1:50 in PBT ) was done overnight at 4°C . After three washing steps in PBT the samples were incubated 1 h with AlexaFluor 488-labeled chicken-anti-goat IgG ( Molecular Probes ) 1:200 and the DNA stain DRAQ 5 ( 1:1000 ) in PBT . The samples were mounted with Mowiol and analyzed with a confocal laser scanning microscope ( Philips/FEI CM 10 ) . Mid-log-phase cultures of P . aeruginosa PAO1 grown in brain heart infusion ( BHI ) medium were resuspended in 10 mM sodium phospate buffer supplemented with 1% TSB and adjusted to OD600nm 0 . 6 . Bacteria were exposed to 0 . 8 μM FLG2-4 for 30 min at 37°C . Periplasmic fractions were collected as described [70] . Briefly , bacterial pellets were washed in 10 mM sodium phosphate buffer with 20% sucrose and 1 mM EDTA ( pH 8 . 0 ) , incubated with agitation for 10 min at room temperature , and centrifuged at 10000 × g for 10 min . Bacterial pellets were resuspended in ice-cold H2O and shaken on ice for 10 min . The suspension was centrifuged at 10000 × g for 10 min at 4°C . Supernatants containing the periplasmic fractions were saved . To isolate the cytosolic content , the bacteria were resuspended in H2O prior to heat inactivation and lysis at 95°C for 5 min . After centrifugation the supernatants were saved as cytosolic fractions . To separate the inner and outer membrane fractions , the remaining pellets were treated with 1% sarkosyl in 30 mM Tris ( pH 8 . 0 ) [71] . After centrifugation the supernatants were saved as inner membrane fractions and the sarkosyl-insoluble pellet as outer membrane fractions and finally resuspended in SDS PAGE loading buffer . The fractions were separated by SDS PAGE using 12% acrylamid gels and analyzed by Western Blot as described [23] . Inner membrane fractions were verified with antibodies directed against XcpY [72] . The Lysozyme Lysis Assay was conducted as described earlier [73] , except that chloramphenicol ( 15μg/ml ) was used instead of streptomycin and the decrease in optical density was monitored at 595 nm . For the SYTOX Green uptake , mid-log-phase BHI cultures of P . aeruginosa PAO1 were suspended in 10 mM sodium phosphate buffer , supplemented with 1% TSB and adjusted to OD600nm 0 . 2 . SYTOX Green was added at 1 μM final concentration and bacteria were exposed to 0 . 8 μM FLG2-4 , 2 μM Polymyxin B , or kept untreated . DNA-associated fluorescence ( Excitation 485 nm/Emission 535nm ) was time-dependently quantified using a Berthold Twinkle microplate fluorometer . pUC18 plasmid DNA was isolated with the GeneJET Plasmid Miniprep Kit ( Thermo Scientific ) . To linearize the plasmid , pUC18 was digested with HindIII and purified with the GeneJET Extraction Kit ( Thermo Scientific ) . Approximately 120 ng of supercoiled or linearized plasmid and increasing concentrations of FLG2 peptide fragments ( 0 . 3 μM , 0 . 9 μM , 1 . 8μM ) were incubated for 30 min at room temperature in 670 mM Tris-HCl , pH 8 . 8 , 160 mM ( NH4 ) 2SO4 , 25 mM MgCl2 , 0 . 1% Tween20 . Samples were loaded immediately onto a 0 . 8% agarose gel in 0 . 5 x TAE pH 8 . 0 buffer , supplemented with 0 . 05% ethidium bromide , and run at a field strength of 5 . 7 V/cm . After electrophoresis , bands were visualized under UV light . Logarithmic grown P . aeruginosa PAO1 were washed and concentrated to an OD600nm of 2 in 10 mM sodium phosphate/1% TSB . The amount of FLG2-4 used was about 2 x 107 molecules per CFU . Samples were incubated at room temperature for 30 min and subsequently treated with formaldehyde ( 1% final ) for 30 min . Crosslinking was stopped by addition of glycine ( 125mM final ) and bacteria were washed . Cosslinked DNA was immediately isolated with peqGold TriFast according to the manufacture’s protocol and dot-blot analyses were performed using the anti-FLG2-4 antibody . The influence of different FLG2 C-terminal peptides on a PCR was analyzed using a conventional Taq S polymerase ( Genaxxon ) . Human keratinocyte cDNA , corresponding to 10 ng total RNA , was used as template in a 20 μl PCR reaction , including 200 μM dNTPs , 200 nM each primer and 0 . 5 U polymerase for 30 cycles . Primer pairs for the housekeeping gene GAPDH ( 5'-CCAGCCGAGCCACATCGCTC-3' ( forward ) and 5'-ATGAGCCCCAGCCTTCTCCAT-3' ( reverse ) ) with an annealing temperature of 60°C were used . FLG2-4 was used in three different concentrations , 0 . 78μM ( 250 ng/20 μl PCR reaction ) , 0 . 32 μM ( 100 ng/20 μl ) , and 0 . 16 μM ( 50 ng/20 μl ) , whereas both hBD-2 and BSA were only used at 250 ng/20μl ( 2 . 88 μM and 0 . 19 μM , respectively ) . PCR samples were subsequently loaded on a 1 . 2% agarose gel containing 0 . 5 μg/ml ethidium bromide and were run in a 1 x TAE buffer system . Bands were visualized under UV light . E . coli bearing plasmid pBR322 was grown in supplemented M9 at 30°C to mid-log-phase . Pelleted cultures were resuspended in 10 mM sodium phospate buffer supplemented with 1% TSB , adjusted to OD600nm 0 . 2 , and supplemented with 170 μg/ml chloramphenicol . Aliquots were treated with 0 . 8 μM FLG2-4 or kept untreated . Incubation was continued and samples were taken after 0 , 60 , 120 , 180 , 240 min , heat inactivated at 95°C , and subsequently analyzed by PCR . 1μl was used as template in a 20 μl PCR reaction including 200 μM dNTPs , 200 nM each primer and 0 . 5 U Taq S polymerase ( Genaxxon ) for 18 cycles . The sequences of the used pBR322 primer pair for the bla gene is 5'-TTGCCGGGAAGCTAGAGTAA-3' ( forward ) and 5'-GCTATGTGGCGCGGTATTAT-3' ( reverse ) with an annealing temperature of 60°C . As control for stalled chromosomal DNA replication the 16S rRNA gene was amplified using 5'- AGAGTTTGATCMTGGCTCAG-3' ( forward ) and 5'-TACGGYTACCTTGTTACGACTT-3' ( reverse ) with an annealing temperature of 56°C . The PCR samples were subsequently loaded on a 1 . 2% agarose gel containing 0 . 5 μg/ml ethidium bromide and run in a 1 x TAE buffer system . Bands were visualized under UV light . Filaggrin-2: NP_001014364 XP_371313; Filaggrin: NP_002007 XP_048104 XP_378901; Human beta-defensin-2: AAC69554; Dermcidin: ABQ53649 | Pseudomonas aeruginosa is able to cause severe infections that increasingly threaten patients with cystic fibrosis and burns . The emerging antibiotic resistance of those bacteria exigently necessitates the development of new effective drugs . Since healthy skin is unexpectedly resistant towards P . aeruginosa infections , it constitutes a promising source of new antimicrobials . We identified fragments of the insoluble skin protein filaggrin-2 as P . aeruginosa-bactericidal proteins that can be released by the action of P . aeruginosa from the outermost skin layer . Unlike many other antimicrobial proteins , filaggrin-2 fragments target bacterial replication , thus presenting a new mode of antibacterial action . Our findings could initiate the urgent development of newly designed antimicrobials and effectively tackle the challenges of Pseudomonas infections | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
]
| []
| 2015 | Skin-Derived C-Terminal Filaggrin-2 Fragments Are Pseudomonas aeruginosa-Directed Antimicrobials Targeting Bacterial Replication |
Cardiac fibrosis occurs in many forms of heart disease and is considered to be one of the main arrhythmogenic factors . Regions with a high density of fibroblasts are likely to cause blocks of wave propagation that give rise to dangerous cardiac arrhythmias . Therefore , studies of the wave propagation through these regions are very important , yet the precise mechanisms leading to arrhythmia formation in fibrotic cardiac tissue remain poorly understood . Particularly , it is not clear how wave propagation is organized at the cellular level , as experiments show that the regions with a high percentage of fibroblasts ( 65-75% ) are still conducting electrical signals , whereas geometric analysis of randomly distributed conducting and non-conducting cells predicts connectivity loss at 40% at the most ( percolation threshold ) . To address this question , we used a joint in vitro-in silico approach , which combined experiments in neonatal rat cardiac monolayers with morphological and electrophysiological computer simulations . We have shown that the main reason for sustainable wave propagation in highly fibrotic samples is the formation of a branching network of cardiomyocytes . We have successfully reproduced the morphology of conductive pathways in computer modelling , assuming that cardiomyocytes align their cytoskeletons to fuse into cardiac syncytium . The electrophysiological properties of the monolayers , such as conduction velocity , conduction blocks and wave fractionation , were reproduced as well . In a virtual cardiac tissue , we have also examined the wave propagation at the subcellular level , detected wavebreaks formation and its relation to the structure of fibrosis and , thus , analysed the processes leading to the onset of arrhythmias .
The contraction of the heart is controlled by propagating waves of excitation . Abnormal regimes of the wave propagation may cause cardiac arrhythmia , asynchronous contractions of the heart and even lead to cardiac arrest and sudden cardiac death . Cardiac arrhythmias often originate from blocks of propagation [1] . In that case , the wave goes around the block , reenters the same inceptive region and , thus , forms a persistent rotational activity called cardiac reentry , which is one of the main mechanisms of lethal cardiac arrhythmias . A normal heart has a complex structure , which is composed of the bundles of elongated cardiac cells . Apart from premier excitable cardiac cells , there are also inexcitable and non-conducting cells of connective tissue: cardiac fibroblasts . Their role is to maintain the structural integrity of the heart [2] and repair injuries [3] . Fibroblasts outnumber cardiomyocytes in a healthy human heart although occupying a much smaller total volume [2 , 4] . However , many pathological conditions are associated with an excessive growth of the fibrous tissue , called cardiac fibrosis , which is , therefore , considered to be one of the major arrhythmogenic factors [5 , 6] . The mechanisms of arrhythmia onset in fibrotic tissue remain poorly understood but generally believed to be associated with the increased probability of waveblock formation . It is a well-established fact that the presence of the non-conducting cells slows down wave propagation [7] and can completely block it if the non-conducting cells’ density is high . The critical density of non-conducting cells above which the conduction terminates is called percolation threshold . This concept originates from the percolation theory , and , by definition , it specifies the point of long-range connectivity loss/formation in random systems . Connectivity here refers to electrical synchronisation in the tissue , or , in other words , the ability to transmit electrical waves of excitation . Percolation threshold , i . e . the critical density of non-conducting cells which breaks long-range connectivity , plays an important role in arrhythmogenicity . It was shown that cardiac tissue is most susceptible to arrhythmias if the density of non-conducting cells is only slightly ( ∼2-3% [8] ) below the percolation threshold . There are two main factors that may facilitate reentry formation . First , a large amount of non-conducting cells ( acting as heterogeneities ) increases the probability for waveblock formation [9] . Second , a high fraction of non-conducting cells creates a ‘maze’ that effectively lengthens the travel distance for the waves as they follow a longer zig-zag path [10] and , thus , provides sufficient room for the emerging reentrant loops . As a result , high density of non-conducting cells both facilitates the initiation and creates conditions for the existence of reentrant cycles , resulting in a highly arrhythmogenic substrate . Up to now wave propagation failure was studied only in generic mathematical representations of heterogeneous cardiac tissue with each cell randomly chosen to be either conductive or non-conductive ( see e . g . [9 , 11] ) . In this kind of 2D computer models , the propagation of excitation failed at 40% of non-conducting cells [12] , which is also within the range of values predicted by classic mathematical models ( e . g . 37-44% of the area uncovered by conducting elongated ellipses with the shape similar to cardiomyocytes [13] ) . However , experimental measurements [14] indicate that wave propagation and synchronous contraction in 2D cardiac monolayers is observed for up to 75% percentage of non-conducting cells . In this paper , we study the phenomenon of wave propagation in cardiac tissue with a high density of non-conducting cells using a joint in vitro-in silico approach . We performed experiments in 25 monolayers with various percentages of non-conducting cells and detected wave propagation to determine the percolation threshold . We have found that , indeed , the experimentally measured threshold ( 75% of the area covered by non-conducting cells ) is substantially higher than what was predicted in conventional computer modelling ( 40% [12] ) or classic mathematical models . Further morphological examination revealed that the key mechanism of conduction in highly heterogeneous tissue is likely to be tissue patterning . The cardiomyocytes were not located randomly but organised in a branching network that wired the whole sample . Next , in order to explain cardiac network formation , we applied a virtual cardiac monolayer framework developed in [15] , based on the Cellular Potts Models [16–18] . We proposed a hypothesis that such self-organisation occurs due to cytoskeletons’ alignment . Based on this hypothesis , we were able to obtain branching patterns , as well as reproduce the decrease in conduction velocity and wave percolation observed in experiments . We have further studied in silico the process of formation of the wavebreaks leading to reentry formation and analysed the tissue structures that caused them . This paper is organized as follows . First , we describe the experiments conducted with the neonatal cell cultures . Second , we analyse the patterns and formulate the hypothesis on the mechanism of their formation . Third , we implement the hypothesis and reproduce pattern formation in silico in a Cellular Potts Model ( CPM ) of cardiac tissue [15] . Next , we compare electrical signal propagation in computer modelling and in experiments . Finally , we show a spontaneous formation of uni-directional blocks in the model resulting in reentry formation and locate and analyse the structures causing it .
We cultured neonatal rat ventricular cardiomyocytes mixed with cardiac fibroblasts in variable proportion ( 20-88% fibroblasts ) and studied electrical activity in these monolayers using optical mapping . In Fig 1a and in S1 Video the wave propagation in a sample with high fraction of non-conducting cells is shown ( 66% ) . In spite of high percentage of non-conducting cells , this sample is still conducting electrical waves , however the wave propagation pattern is complex . The wave originates from the stimulation point marked by a yellow spike at the bottom of the tissue and initially spreads in all directions . However , due to a large number of non-conducting cells , the wave is blocked at multiple sites . After a short delay ( in areas outlined with dashed ellipses ) , the wave propagates further into the left and the right parts of the sample , and again spreads in various directions . This process repeats multiple times , resulting in a complex , fractionated wave pattern containing narrow pathways and some bulk excited regions . The conduction blocks in Fig 1a are shown in red , which are the places where the wave propagation was blocked and the wave had to go around . The main propagation paths are shown with white and black arrows . In this sample , the electrical wave propagation was still possible and long-range connectivity was still present , however the amount of non-conducting cells was close to the percolation threshold . We have found that the percolation threshold for the neonatal rat cardiac monolayers was 75 ± 2% of non-conducting cells . We have also measured conduction velocities in the samples below the percolation threshold ( Fig 1b ) . The measurements show that the velocity decreased when approaching the percolation threshold . In the samples with low level of non-conducting cells , the velocity was approximately 10-14 cm/s , and it decreased twofold in the samples with 70% non-conducting cells . The number of conduction blocks was higher in samples with high percentage of non-conducting cells . As a result , the mean conduction velocity decreased with the increase in a percentage of non-conducting cells . After optical mapping of the wave propagation , we have fixated the samples and studied their morphology using immunohistochemical labeling . We have found that the cardiomyocytes in the samples have formed connected networks capable of electrical wave propagation . In Fig 2 cardiomyocytes are shown in pink , and the cluster that they have formed is outlined with a white contour . The cardiomyocytes were organised in a branching structure that wired the whole sample . We have followed the pathway using a confocal microscope and it was possible to find long-range connectivity in the tissue . Therefore , we observed that the cardiomyocytes were organised in conduction pathways and assumed that there must be a mechanism responsible for their self-organisation . Pattern formation in cell populations during development was extensively studied using Cellular Potts Models [16 , 17 , 19] , including many studies [20–22] of the branching structures similar to the one in Fig 2 . Therefore , we have considered several existing hypotheses to explain the formation of the cardiac pathways . First , we suggested that differential adhesion together with cell elongation may be enough to explain the observed patterning . A similar system with autonomously elongating cells was successfully used to describe vasculogenesis [20] . However , we did not impose obligatory elongation on the cardiac cells , because they are not necessarily elongated according to our experimental observations ( see Fig 3 below or S2 Video ) . Cardiomyocytes obtain their typical brick-like shape with the guidance of the extracellular matrix over the course of development . However , there is no evidence for any internal autonomous mechanism for elongation . In experiments on the glass , cells were not only bipolar but also tripolar and multipolar ( see S2 Video ) . In our tissue growth model [15] , which was accurate at replicating realistic cell shapes , this feature was reproduced with explicit introduction of the actin bundles . Cooperation between aligned actin bundles pulling in one direction shaped the cell more efficiently , which led to clustering of these virtual actin bundles and resulted in multipolar cell shapes . Similar approach was also previously used to describe the shapes of dendritic cells [23] . Therefore , since the elongation was not imposed , there was no mechanism forcing cardiomyocytes out of clusters to search for new connections . Next , we considered various mechanisms that were previously used to explain similar branching patterns in angiogenesis . The main sources of instability in those models were chemotaxis and contact-inhibition [21] . The percolation in the networks formed due to sharp gradients of chemoattractants was also studied previously in a continuous model [22] . However , there was no evidence of any directed migration of cardiac cells and for type-specific contact-inhibition , similar to those that select a tip cell in a growing blood vessel . Therefore , we discarded this hypothesis either . After trying several approaches used before , we have not found an existing model that could have been 1 ) applicable to the cardiac tissue and at the same time 2 ) could reproduce the experimentally observed branching structures . Finally , after careful analysis of our experimental preparation , we found that an essential feature of our structure was the alignment of the cytoskeletons in the neighbouring cells . This alignment of the actin fibres is clearly seen in our experimental preparations . In Fig 3 the bundle of neonatal rat cardiomyocytes is shown . The red arrows indicate the intercalated disks between the cells . The white arrows point at the actin bundles on the opposite sides of the intercalated disk , that smoothly continue one another . The precise biological mechanism of such alignment is unknown . However , in our view , it can be derived from a well-known property of actin cytoskeleton reinforcement in response to the external force [24] . Actin filaments , as well as the adaptors that link them , remodel under applied tension . In case of contact of two cells , the actin filaments are connected through adherens junctions , which transmit the tensile forces between the filaments of the neighbouring cells . It was shown that higher tension stabilises the whole complex [24] . The tension is maximal , if the actin filaments are aligned with each other , which gives a preference for intercellular alignment of the cytoskeletons . Therefore , we have incorporated this mechanism into our Cellular Potts Model ( CPM ) of cardiac cells [15] . This model was already adjusted to reproduce characteristic shapes of the cardiac cells in virtual tissue model , and here we extended it with a new energy term , that corresponded to the alignment of actin bundles in the neighbouring cells . In our CPM model , cell-substrate adhesion sites , to which actin bundles are anchored , were represented as separate entities: specially labeled subcells of the lattice . If two adhesion sites of two cells came into contact , we established a new connection between them . The connection means , that an additional bond energy was applied to them . This energy depended on the angle between the linked actin bundles ( see Fig 4 ) . The minimum of the energy corresponded to smooth coupling between the bundles , or zero angle . In this case , the bond was the most stable , but couplings with the non-zero angle between the bundles had a tendency to break apart . With this new energy term that favours cytoskeletons alignment , the cardiomyocytes in simulations created branching patterns . In Fig 5a , a resulting simulated structure of the sample with 70% non-conducting cells is shown . We see that in this sample only 30% of cardiomyocytes were able to build a network , and even with such a high density of non-conducting cells , this network was fully interconnected . Our further studies showed that such interconnection was established for every sample with 30% of cardiomyocytes ( n = 10 ) of 1 cm × 1 cm size , which is close to the size of the Petri dish used in our experiments . For samples with 28% of cardiomyocytes , the network was interconnected with 20% probability . We also see that the patterns in computer simulation ( Fig 5a ) and in experiment ( Fig 5b ) have similar features , such as long single-cell-wide connections that bridged the gaps between cell clusters , or isolated non-conducting clusters trapped within the main cardiac pathway . Therefore , using the hypothesis on cytoskeletons’ alignment allowed us to reproduce not only the percolation threshold , but also the main features of the pattern . S2 Video shows the growth of the branching pattern , highlighting the large connected clusters of the cardiomyocytes by different colours . One can see from the video , that subtle movements of the cells ( left ) result in dramatic changes in connectivity ( right ) . This emphasises the important role of the protrusions of the cells , which are taken into consideration in our model , in electrical signal propagation . Using this model with cytoskeleton alignment , we have studied further wave excitation patterns and the percolation threshold for electrical waves in the system . We have reproduced the experiments from Fig 1a in silico using Majumder et al . [25] model for neonatal rat ventricular cardiomyocytes . Fig 6a and S3 Video show wave propagation in a virtual sample with 70% non-conducting cells . The wave propagation pattern is complex and similar to one observed in experiment ( see Fig 1a ) . One can see , that the number of propagation blocks per unit area is similar to that of the experimental activation pattern , and the trajectories of the waves share the same features . Note , that the spatial scale of the simulated activation map is slightly smaller than those of the experimental one . The percolation threshold in virtual cardiac monolayers was equal to 71 . 5 ± 1 . 5% of non-conducting fibroblasts , meaning that the samples with such level of fibroblasts had 50% probability of percolation . In our simulations , 100% of the samples ( n = 10 ) with 70% fibroblasts were interconnected , whereas samples with more than 73% fibroblasts ( n = 10 ) were never conducting . For 72% fibroblasts 20% of the samples were functional and for 71% fibroblasts , those were 80% . The conduction velocity dependence on the density of fibroblasts is shown in Fig 6b . For each simulated sample we have indicated the mean value of the velocity , and the standard deviation of the velocity distribution was shown with error bars . One can see , that the dependency of the velocity on the fibroblasts’ density is similar to one measured in experiments ( see Fig 1b ) . Closer to the percolation threshold the fluctuations in conduction velocities amplified due to the stochastic nature of the percolation block . Moreover , variations in conduction velocities within one sample were very high in that range , because of a large number of conduction blocks . Thus , the percolation threshold in our simulations with Ebond = 5 . 0 was only slightly lower ( ≈ 72% ) than in experiment ( 75% ) . However , both values are much higher than the predictions of the models with random cells distribution ( 40% ) . Fig 6b also shows the percolation threshold in a computer model without cytoskeleton alignment ( Ebond = 0 , blue ) , which was equal to 64% . It is still higher than the threshold for randomly distributed cells ( 40% ) since some clustering of the cells takes place . The reason for this clustering is that the adhesion between FB is slightly stronger than between cardiomyocytes or between cells of different types ( JFB−FB = 500 vs . JCM−CM = JCM−FB = 700 ) . Such a difference is required to reproduce the difference in cell shapes correctly: all surface energies of the fibroblasts have to be lower than those of cardiomyocytes to reproduce dynamic fluctuations of their boundaries . The model in this study inherits some parameters from our previous paper [15] , however it was reparameterized to ensure long-term stability of the model ( on the scale of 20’000-50’000 MCS ) . We have first tried to enhance differential adhesion , but discovered that the differential adhesion alone was not capable of reproducing the branching pattern observed in experiments ( as was discussed above ) . Thus , differential adhesion was tuned down as much as possible ( JCM−CM set to be equal to JCM−FB ) , preserving the desired cell shapes and volumes . Nonetheless , the resulting parameters cause some clustering even for Ebond = 0 , which , in turn , raises the percolation threshold from 40% up to 64% . These data show to which extent the accurate representation of cell shapes and clustering of the cells of different types affect the percolation threshold ( +24% compared to a simple square lattice model ) , and how much it is affected by cytoskeletons’ alignment ( extra +8% ) . We have studied various values of Ebond and have found that there is an optimal range of values that provides the highest percolation threshold . One can see in Fig 7 that both low ( Ebond = 0 . 0 ) and high ( Ebond > 7 . 5 ) cytoskeletons’ coupling energies result in the low percolation thresholds . When the coupling is weak , the cardiomyocytes do not form strong connections between each other and form clusters rather than networks ( see top-right pattern ) . When the coupling is strong , the cardiomyocytes form very strong connections instead , which effectively “freeze” these cells in place and do not allow any motility , which is essential for network formation ( see bottom-right pattern ) . Therefore , intermediate values of the Ebond are optimal for network formation , as they favour cytoskeleton alignment sufficiently , yet without imposing too much restriction on cells’ motility . We have chosen Ebond = 5 . 0 for this study , since it belongs to the optimal range of values and , at the same time , reproduces the most qualitative features of the experimental samples . One can see in the middle-right image in Fig 7 , that there are some fibroblasts ( white cells ) trapped in the clusters of cardiomyocytes ( red cells ) . It is an important characteristic feature of the pattern , which is also present in our experimental samples ( see right-hand image in Fig 5 ) . In Fig 7 , we have shown two definitions of the percolation threshold corresponding to 10% and 50% probability of successful network formation . The 50% is typically used in theoretical studies [8] , however here we compare the probability corresponding to 10% with the experiment . The reason for this is that it is not possible to carefully assess the probability of network formation in experiments . There are many other factors that may cause the propagation failure in cell culture and large statistics is needed to find the exact value corresponding to 50% probability of conducting network formation . Therefore , we have recorded all successfully conducting samples including those that belong to a vicinity of the percolation threshold and were still , by chance , conducting . Therefore , we suggest that 10% probability threshold provides a practical estimation of the highest percentage of non-conducting cells which we are likely to witness electrically connected , and thus this value is better for comparison with the experiment . We have shown that in virtual tissues a spontaneous onset of reentry can occur . It was observed in samples with a high level of non-conducting cells . Fig 8 shows such process for a sample with 70% non-conducting cells . We see that after the first stimulus applied to the left border of the sample ( shown in yellow ) , the wave propagation was blocked at the lower part of the sample , but it pursued through the upper part . After reaching the right border , the wave turned around and entered the bottom region . However , this first wave was blocked in the middle of the sample , as the tissue in the upper part of the sample have not been recovered yet . The wave from the second stimulus , applied at the same site , has followed the same path and formed a sustained circulation along the path shown with the red dashed line in Fig 8 . Detailed analysis of the structure revealed that formation of the reentry here is solely due to specific structure which is shown in the middle ( inside the yellow square ) , which acts as an area of uni-directional block ( or “diode” ) : the waves can propagate from right to left , but not in the opposite direction . The diode is formed by two cell clusters that barely touch each other . These cell clusters have slightly different areas adjacent to this connection . Therefore , if the wave propagates from left to right ( from a smaller cluster to a larger one ) , then the small cell cluster does not produce enough current and can not depolarise a bigger one . The transmembrane potential in the largest cluster raises ( which is shown in pink colour ) , but not enough for the sodium channels to open . This effect is called source-sink mismatch [1] . When the source ( a smaller cluster ) is insufficient compared to the sink ( a larger inactive cell cluster ) , the wave propagation is blocked . This effect was observed in chemical systems [26] and later in cardiac monolayers [27] . In a sample in Fig 8 , the “diode” could initiate a reentry if the sample is stimulated from the left or from the top with a high frequency ( 4 Hz or more ) . We have performed studies in 6 large samples with 70% of non-conducting cells . All of these samples had 2-5 areas of the uni-directional block and many bi-directional blocks , but only one of these samples was arrhythmogenic . Thus , in addition to “diodes” , some extra geometrical conditions are required . The precise conditions are to be studied in the future , however , they are related to the presence of the long conducting circuits in the tissue . In fact , in Fig 8 we see that apart from the diode , there is also a loop , which is shown with a red dashed curve in the bottom-right image . The diode is a part of this loop , however , the loop is large enough to account for recovery of the bottom part of the tissue after one rotation . The samples with a higher density of non-conducting cells were even less likely to have long circuits , thus we did not observe sustained reentry there . We concluded , that reentry formation requires not only uni-directional blocks but also long circuits , and the densities of non-conducting cells slightly below the percolation threshold are the most arrhythmogenic ones . This result was previously shown for random cell distributions , that reentry is most likely to occur 5-10% below the percolation threshold [8] . The principle holds true in our model , but quantitatively the densities of non-conducting cells are different . Our model shows , that the areas of the uni-directional block can be naturally formed during tissue growth . Every time one spreading cell comes in contact with the other cell cluster , there is a chance that this connection will be asymmetric . It may explain the fact that reentries are frequently observed in the experimental setups with neonatal cardiac monolayers .
We observed paradoxical electrical wave propagation in samples with up to 73–75% of non-conducting cells instead of mathematically predicted 40% for randomly distributed cells . We have shown both in vitro and in silico , that electrical wave propagation was possible due to the formation of the conduction pathways that rewired the whole monolayer . We have proved the existence of this branching network with immunohistochemical images . We have measured the conduction velocity , which decreased with the increase of the portion of non-conducting cells in the monolayer in a similar way in experimental and computational studies . To explain the formation of the pathways , we have considered several existing hypotheses based on differential adhesion , cells elongation and directed migration . However , all of them were drawn inapplicable to the cardiac tissue . As a result , we developed and proposed a new hypothesis , based on cytoskeleton alignment . Assuming that cardiomyocytes align their cytoskeletons to fuse into cardiac syncytium , the morphology of conductive pathways was successfully reproduced in computer modelling . The virtually generated monolayers were then used for the studies of the electrical wave propagation , and experimentally observed velocity decay and high percolation threshold were reproduced as well . The proposed hypothesis of cytoskeletons alignment has never been considered before . This mechanism , however , has some similarities to diffusion limited aggregation ( DLA ) : a process in which random-walking particles form fractal aggregates . In our model , a cell ‘sticks’ to the growing pattern , but , unlike DLA , our cells also align cytoskeleton’s orientation with the pattern . Such alignment actually takes place as a part of syncytium formation , when cardiomyocytes fuse and align their cytoskeletons . Once a cell aligns cytoskeleton with its neighbours , this structure maturates and fixes the cell in place . In some sense , this process causes contact inhibition only between cardiomyocytes and not between cardiomyocytes and non-cardiomyocyte cells . Such type-specific contact inhibition results in the branching structure like one shown in angiogenesis [21] . Formation of conduction pathways and complex texture of the tissue affects arrhythmogenicity , and may also be important for arrhythmia treatment . For example , it was shown that texture of cardiac tissue at subcellular level can substantially affect the propagation of external current during defibrillation [28 , 29] . It would also be interesting to quantify the excitation patterns in terms of the number of re-entrant sources and wavefront complexity [30] . Therefore , it will be interesting to perform a similar study for the textures generated with our model . There are several limitations of the methods used in this study . First of all , we have conducted experiments with cell cultures , which are different from the real cardiac tissue . It would be interesting , yet more complicated , to study patterning of the real 3D cardiac tissue . However , the mechanisms of patterning that we discovered here are likely to be universal and might be applicable to the real cardiac development as well . Second , fibrosis is a more complex condition than just excessive growth of fibroblasts . It also involves collagen deposition that insulates the cardiac fibers from one another and is also considered to increase arrhythmogenicity . In this study , we did not take the extracellular matrix into account , but it would be also interesting to measure its effect on the percolation threshold in the future studies . Third , recent studies suggest that there are other conducting cell types in the heart , such as myofibroblasts [31] and macrophages [32 , 33] . The fraction of these other cell types in the heart does not exceed 5-7% [34 , 35] , and , moreover , their amount in ventricular cell population , used in this study , is negligible [32] . However , presence of such cells in other cell populations may affect the percolation threshold . Finally , percolation depends on the size of the sample and scales with this size . For random systems , the scaling laws are known . In our case , we did not consider scaling and used similar size of the samples in silico as in the experiment . However , it would be interesting to study scaling of the percolation threshold and see how the size can affect the probability of wavebreak formation . We conclude , that the cardiomyocytes in heterogeneous tissues with high portion of non-conducting cells can form a connected network and allow electrical signal propagation in monolayers containing up to 75% of non-conducting cells .
All studies conformed to the Guide for the Care and Use of Laboratory Animals , published by the United States National Institutes of Health ( Publication No . 85-23 , revised 1996 ) and approved by the Moscow Institute of Physics and Technology Life Science Center Provisional Animal Care and Research Procedures Committee , Protocol #A2-2012-09-02 . In this study , we have used a mathematical model that was previously developed for cardiac monolayers formation [15] . It is based on the Cellular Potts Model ( CPM ) formalism , which describes cells as the domains of the regular lattice and assigns energy to the system of cells to describe their growth and motility . In our previous work [15] , we have selected the main features of the cardiac cells ( such as area , number of protrusions , etc . ) and parametrised our model to reproduce the cell shapes . In this study , we extended the model with a new energy term responsible for syncytium formation . The evolution of the cardiac monolayer is described by the Hamiltonian: H = H adhesive + H elastic + H protr + H nuclei + H junctions , ( 1 ) where Hadhesive + Helastic is the basic CPM model and Hprotr is the term describing the protrusion dynamics of the cardiac cells , which produces a characteristic polygonal shapes of these cells . Hnuclei corresponds to higher rigidity of the nuclei compared to the cell body . Finally , Hjunctions is a new term , that describes the stability of adherens junctions and alignment of the cytoskeletons of the neighbouring cells . The core feature of the model of cardiac cells is the explicit representation of cell attachments as the labelled subcells of the lattice . They are first assigned when the cell expands , and can be destroyed with a certain penalty , if , for example , the cell is stretched due to its movement . The number of attachments per cell is limited by the amount of actin present in the cytoplasm , which is also reflected in the model . If the maximal number of attachments is reached , new attachments do not form . The spreading of virtual cells is shown in S5 Video . One of the important features of our approach , is that elongation of the cells is not imposed , but rather emerges due to interactions with the environment . Cardiac cells in a real heart are indeed elongated , as they are guided by the extracellular matrix , whereas in a Petri dish cells can spread in any direction with no preference . However , this does not result in a circular shape due to the limited number of attachment sites where spreading occurs . Instead of explicit declaration of elongation , we suggested that virtual cells have only a small ( ≈10 ) number of well-developed mature attachment sites . Moreover , these sites in a model tend to cluster spontaneously ( see S5 Video ) , as the actin strands pull the cell more efficiently acting together , rather than separately . Therefore , these clustered attachment sites turn cell shapes into bipolar , tripolar or multipolar states . The bipolar state is the most stable one for isolated cells , however tripolar cells are also relatively common . This correctly represents the population of cells observed in experiments . Hprotr is an energy term , that decreases with the distance from the centre of mass of the cell , and which is applied only to these labeled attachment sites . As a result , the attachment sites spread out and reproduce characteristic polygonal shape of the cardiac cells . Here , in this model , we omit the details of the spreading process , which involves polymerization and depolymerisation of the actin , attachment/detachment , etc . , but we mimic the overall dynamics of the protrusions . Also , we assume , that for every attachment site a corresponding actin bundle exists , which stretches from the attachment site towards the proximity of the nuclei . If two attachment sites in the neighbouring cells come into contact , the bond between these attachment sites can be formed . In the algorithm , this bond appears if one attachment site attempts to move over the other . In such case , instead of copying of the subcell , the connection establishes . A new energy term Hjunctions applies to the subcells , which are involved in a newly established junction . This term determines the stability of the cell-to-cell junction and depends on the angle between the actin bundles , associated with the attachment sites involved ( see Fig 4 ) . It is determined as follows: Hjunctions=∑i→ , j→labeledjunctionEbond ( 1−cos ( αi , j ) ) , where αi , j ( shown in Fig 4 ) is the angle between two cytoskeleton bundles of two neighbouring cells which ends at points i and j labelled as the parts of the junction . The wider is the angle , the less stable is the junction . Therefore , the junctions with continuous actin bundles on both sides persist , but the kinked bundles tend to lose the connection . As a result , the actin bundles of the neighbouring cells tend to align and stay in the aligned state . The parameters of the model used in simulations ( see Table 1 ) in this paper were adjusted to compensate the additional energy term Hjunctions . The most of them are close to the parameters used in our original paper [15] . The value of the new energy term was set to Ebond = 5 . 0 , which provides enough stability for the junctions to maintain the branching structure but at the same time not too much stability to allow cells to search for possible new connections . Addition of Hjunctions effectively increased the adhesion between cardiomyocytes . Therefore , the differential adhesion was toned down in a model to allow cardiomyocytes to migrate randomly before they maturate and stick to the pattern . In this study , we have changed type-specific adhesion coefficients Jcell−cell . Our choice of the parameters was guided by the desired balance between branching and clustering: 1 ) lower adhesion energy ( J ) to the cells of the same type increased the size of unstructured , irregularly shaped clusters; 2 ) on the other hand , high adhesion energy forced cells to migrate out of clusters . We have chosen neutral values of JCM−CM = JCM−FB , which meant that cardiomyocytes had no preference in neighbours . Their energy was equal for being surrounded with either other cardiomyocytes , or with the non-conducting cells . The non-conducting cells had a slight tendency to cluster ( JFB − FB < JCM − FB ) . This choice of parameter allowed us to qualitatively reproduce the patterns observed in experiments ( see Figs 2 and 5 ) . The code used in this study is available on Github: https://github . com/NinelK/VCT . Examples 3 and 4 in the repository correspond to formation of cardiac pathways with or without cytoskeletons alignment . The structure of the virtual samples was first generated with the Cellular Potts Model . After the cells were grown and statistical characteristics of the cells and clusters stopped changing ( 50’000 MCS ) , the resulting lattice was converted into a matrix of coupling coefficients for electrophysiological studies . The electrical coupling between non-conducting cells and any other cells was set to zero , coupling of the lattice points within the cells was set to Din , and , finally , coupling between cardiomyocytes was set to either DL or DT depending on the orientation of the cells’ virtual cytoskeletons . The methodology was already described in detail in our previous paper [15] and the ratio between coefficients was adjusted to match the anisotropy of the wave propagation in aligned cells ( Din = 100 × DL , DT ≪ DL and negligible ) . The magnitude of the coupling coefficients was slightly adjusted ( Din = 0 . 4 cm2/s , DL = 0 . 4 × 10−2 cm2/s , DT = 0 cm2/s ) to reproduce experimentally measured maximal conduction velocity in the control samples with low portion of non-conducting cells . The same coefficients were then used for all simulations . | Cardiac arrhythmias are one of the major causes of death in the industrialized world . The most dangerous ones are often caused by the blocks of propagation of electrical signals . One of the common factors that contribute to the likelihood of these blocks , is a condition called cardiac fibrosis . In fibrosis , excitable cardiac tissue is partially replaced with the inexcitable and non-conducting connective tissue . The precise mechanisms leading to arrhythmia formation in fibrotic cardiac tissue remain poorly understood . Therefore , it is important to study wave propagation in fibrosis from cellular to tissue level . In this paper , we study tissues with high densities of non-conducting cells in experiments and computer simulations . We have observed a paradoxical ability of the tissue with extremely high portion of non-conducting cells ( up to 75% ) to conduct electrical signals and contract synchronously , whereas geometric analysis of randomly distributed cells predicted connectivity loss at 40% at the most . To explain this phenomenon , we have studied the patterns that cardiac cells form in the tissue and reproduced their self-organisation in a computer model . Our virtual model also took into account the polygonal shapes of the spreading cells and explained high arrhythmogenicity of fibrotic tissue . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"statistical",
"mechanics",
"muscle",
"tissue",
"fibroblasts",
"connective",
"tissue",
"cells",
"waves",
"cellular",
"structures",
"and",
"organelles",
"cytoskeleton",
"cardiology",
"contractile",
"proteins",
"actins",
"computer",
"and",
"information",
"sciences",
"arrhythmia",
"animal",
"cells",
"proteins",
"connective",
"tissue",
"biological",
"tissue",
"muscle",
"cells",
"physics",
"biochemistry",
"cytoskeletal",
"proteins",
"wave",
"propagation",
"computer",
"modeling",
"cell",
"biology",
"anatomy",
"cardiomyocytes",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences"
]
| 2019 | Self-organization of conducting pathways explains electrical wave propagation in cardiac tissues with high fraction of non-conducting cells |
Trachoma is endemic in several Pacific Island states . Recent surveys across the Solomon Islands indicated that whilst trachomatous inflammation—follicular ( TF ) was present at levels warranting intervention , the prevalence of trachomatous trichiasis ( TT ) was low . We set out to determine the relationship between chlamydial infection and trachoma in this population . We conducted a population-based trachoma prevalence survey of 3674 individuals from two Solomon Islands provinces . Participants were examined for clinical signs of trachoma . Conjunctival swabs were collected from all children aged 1–9 years . We tested swabs for Chlamydia trachomatis ( Ct ) DNA using droplet digital PCR . Chlamydial DNA from positive swabs was enriched and sequenced for use in phylogenetic analysis . We observed a moderate prevalence of TF in children aged 1–9 years ( n = 296/1135 , 26 . 1% ) but low prevalence of trachomatous inflammation—intense ( TI ) ( n = 2/1135 , 0 . 2% ) and current Ct infection ( n = 13/1002 , 1 . 3% ) in children aged 1–9 years , and TT in those aged 15+ years ( n = 2/2061 , 0 . 1% ) . Ten of 13 ( 76 . 9% ) cases of infection were in persons with TF or TI ( p = 0 . 0005 ) . Sequence analysis of the Ct-positive samples yielded 5/13 ( 38% ) complete ( >95% coverage of reference ) genome sequences , and 8/13 complete plasmid sequences . Complete sequences all aligned most closely to ocular serovar reference strains . The low prevalence of TT , TI and Ct infection that we observed are incongruent with the high proportion of children exhibiting signs of TF . TF is present at levels that apparently warrant intervention , but the scarcity of other signs of trachoma indicates the phenotype is mild and may not pose a significant public health threat . Our data suggest that , whilst conjunctival Ct infection appears to be present in the region , it is present at levels that are unlikely to be the dominant driving force for TF in the population . This could be one reason for the low prevalence of TT observed during the study .
Trachoma , caused by ocular strains of Chlamydia trachomatis ( Ct ) , is the leading infectious cause of blindness worldwide [1] . Ocular infection with Ct is associated with a characteristic follicular conjunctivitis , known as “trachomatous inflammation—follicular” ( TF ) [2] , which can persist for some time after the initiating infection has been cleared . In some individuals , infection can also cause the sign “trachomatous inflammation—intense” ( TI ) . Repeated and prolonged bouts of severe inflammatory disease can lead to trachomatous scarring [3] ( TS ) which , in some individuals , can eventually cause the eyelashes to turn inwards , producing trachomatous trichiasis ( TT ) , a condition in which the lashes painfully abrade the cornea . In combination with other trachoma-induced changes to the ocular surface , this may lead to corneal opacity ( CO ) and blindness [2 , 4] . In 2014 , the World Health Organization ( WHO ) estimated trachoma to be a public health problem in 51 countries , and responsible for approximately 2 . 2 million cases of visual impairment . Efforts to globally eliminate the disease as a public health problem are promising , with several countries having reported reaching elimination goals [1] . The island states of the Western Pacific Region are made up of several thousand widely dispersed volcanic islands and coral atolls . It has long been suspected that trachoma is endemic in these islands , with reports from the early twentieth century indicating that trachoma was present [5–9] . More recently , trachoma rapid assessments ( TRA ) conducted in the Pacific indicated the presence of trachoma in Fiji , Vanuatu , Solomon Islands , Nauru and Kiribati . Although TRAs do not give accurate estimates of disease prevalence , the TRA data suggested that , whilst TF levels appeared high , both TI and TT were surprisingly scarce [10] . A population-based prevalence survey ( PBPS ) is the gold standard for estimating district-wide trachoma burden [11] and trachoma PBPSs took place between 2011 and 2014 in some districts of the Solomon Islands , Kiribati and Fiji [12] . They reported that TF prevalence was above the WHO-recommended threshold ( ≥10% TF in children aged 1–9 years ) for public health interventions in the Melanesian-dominated districts studied ( Solomon Islands and Fiji ) . Those surveys also reported surprisingly low levels of TT [12 , 13] ( at or below the WHO elimination threshold of 0 . 1% in the all-ages population[14] ) compared to those observed in other populations with highly prevalent TF [15 , 16] . The seemingly discrepant ( with respect to out-of-region comparators ) finding of high prevalence of TF in populations with negligible levels of TI or TT led us to question whether ocular Ct infections are present in Melanesia . We augmented GTMP mapping of Temotu and Rennell and Bellona provinces in the Solomon Islands , with tests for infection and next-generation sequencing , to determine the prevalence of Ct , and whether Ct and active trachoma were associated . If Ct was detected in conjunctival specimens , we considered whether those strains were of an ocular or genital genotype . We considered how our results compare to other published datasets .
The study adhered to the tenets of the Declaration of Helsinki . The London School of Hygiene & Tropical Medicine Ethics Committee ( 6319 and 6360 ) and the Solomon Islands National Health Research and Ethics Committee ( HRC13/18 ) granted ethical approval for this study . Village and household heads were consulted prior to enrolment . Individuals were informed of the nature and requirements of the study prior to enrolment , by a staff member fluent in local dialects , and were asked to provide written evidence of consent . For those aged 18 years and under , written consent of a parent or guardian was required . Our study was conducted alongside GTMP survey teams as they undertook mapping of one evaluation unit ( EU ) in the Solomon Islands in October and November , 2013 . In the GTMP study , an EU was defined as a single administrative province , however , Temotu and Rennell and Bellona were grouped together into a single EU due to their small populations . Local healthcare workers identified Temotu and Rennell and Bellona as the provinces with the highest suspected trachoma burden in the country . The study was a cross-sectional cluster-randomised PBPS of trachoma . We determined a priori that 1019 children aged 1–9 years should be sampled to estimate infection prevalence of 10% with a precision of ±3% at the 95% confidence level , assuming a design effect of 2 . 65 [17] . Our sample size was framed around the population of children in this age range , as they are the group most likely to harbour infection [15 , 18] . The Land Registry listed 533 villages across the two provinces in 2013 . Local healthcare workers identified villages that were not currently inhabited , and remaining villages were eligible for simple random selection . In each cluster a targeted number of households were randomly selected from a full list of village households to recruit the required number of children . All household residents over the age of 1 year in each selected household were eligible for inclusion [17] . Clinical examination was conducted by graders who had been certified according to GTMP protocols [17] . Clinical grading was carried out using the WHO simplified system , in which trichiasis is defined as at least one eyelash in contact with the eyeball ( or evidence of recent removal of in-turned eyelashes ) , TF is defined as 5 or more follicles of >0 . 5mm diameter on the central part of the upper tarsal conjunctiva , and TI is defined as pronounced inflammatory thickening of the upper tarsal conjunctiva obscuring >50% of the deep tarsal vessels [2] . Conjunctival photographs were taken using a Nikon D3000 SLR camera and graded by an independent photograder who had previously had a kappa agreement score in excess of 0 . 9 when grading photographs also graded by a master grader . The photograder was masked to the corresponding field grade for each photograph . Conjunctival swabs were collected from all children aged 1–9 years . Polyester-coated cotton swabs ( Puritan Medical Products , Guilford , ME , USA ) were passed three times over the right tarsal conjunctiva with a 120° turn in between each pass [19–21] . The examiner changed their gloves between participants to avoid cross contamination in the field . 1 in 30 swabs did not touch the conjunctiva but were passed within 15 cm of a seated participant then stored and processed in identical fashion to other study specimens , to act as field contamination controls . Swabs were stored immediately in RNAlater ( Life Technologies , Paisley , UK ) , kept on ice packs in the field before short-term storage at 4°C , and frozen within 48 hours of collection [22] . Specimens were transported to the UK on dry ice where they were stored at -80°C until they were extracted with Qiagen AllPrep DNA/RNA mini kits ( Qiagen , Manchester , UK ) according to manufacturer’s recommendations . DNA specimens were tested for the Ct plasmid using a droplet digital PCR assay targeting a single Ct plasmid target in duplex with a human ribonuclease gene , which acted as endogenous control . The published assay methodology was used [23] with minor protocol adjustments to the tested sample volume ( 4 . 95μL increased to 8μL ) and oligonucleotide concentrations ( primer concentration increased from 300nM to 900nM; probe concentration decreased from 300nM to 200nM ) . Samples were considered valid if there was >95% confidence in non-zero endogenous control concentration , and positive according to published criteria ( >95% confidence in non-zero chlamydial plasmid load [23] ) . Samples from children with TF were retested with an alternative , quantitative assay targeting both chlamydial chromosomal and plasmid targets [24] . The oligonucleotides used are shown in Table 1 . Chlamydial DNA was preferentially enriched in clinical samples using custom Ct-specific RNA baits in a SureSelect system , as developed by the PATHSEEK consortium . Due to the low biomass of conjunctival swabs , human carrier DNA was added to achieve the required DNA input concentration for sequencing . Samples were sequenced on the Illumina MiSeq platform [25] . Articles were identified through two specific literature searches to illustrate how our data compared to existing published studies . PubMed hits , references contained within those articles and relevant articles from the authors’ own archives were considered eligible for inclusion . Using the terms “population-based” and “trachoma” , we identified population-level studies where TF had been reported in children aged 1–9 years and TT had been reported in adults over the age of 15 years in the same population . District-level data were extracted where more than a single district was reported on in a single publication . Using the search terms “trachoma” and “infection” , we identified studies reporting population-level nucleic acid-based infection data alongside clinical data on TF ( with or without TI ) . Data from communities who had received one or more rounds of MDA were excluded . During both literature searches , studies where age-specific trachoma prevalence in 1- or 2-year age bands was presented were also reviewed . Field data were recorded using a purpose-built Open Data Kit ( https://opendatakit . org/ ) app [17] . Age adjustments were carried out using 5-year census age bands [26] . All analyses were carried out using R version 3 . 2 . 2 [27] . The overall agreement between field exam and photographic grade was determined using kappa agreement scores . The relationship between TF and Ct infection was tested using logistic regression . Ct loads in children with and without TF were compared using the Mann Whitney U test . Following preliminary assessment of mapping quality to various Ct strains , 251-bp paired end reads were trimmed and mapped to Ct reference strain AHAR-13 genome sequence ( GenBank accession CP000051 . 1 ) and B/Jali20 plasmid genome sequence ( FM865436 . 1 ) using Bowtie 2 [28] . SAMtools and BCFtools were used to index and assemble reads , and bases were called by collapsing reads vertically [29] . Trimmed reads were also mapped to E/Bour ( genome HE601870 . 1 , plasmid HE603212 . 1 ) to determine whether the choice of reference influenced branching points in the phylogram . Consensus sequences were submitted for megablast search on National Centre for Biotechnology Information ( NCBI ) GenBank to determine nearest relatives based on genetic sequence . Whole genomes were aligned with progressiveMauve [30] . A core alignment was generated by extraction and amalgamation of locally collinear blocks using stripSubsetLCBs [31] . Distance matrices and bootstrapped phylogeny was inferred using phangorn , ape and SeqinR packages in R [32–34] . Regions orthologous to ompA , trpA and the plasticity zone ( PZ; a ~20 kb region of the Ct chromosome between dsbB and ycfV [35] ) were analysed in isolation due to their disproportionately high variability and influence on pathogenicity compared to the rest of the Ct chromosome . These were extracted from consensus sequences using BLASTn in the NCBI BLAST+ suite , and aligned using MUSCLE alignment software [36] .
The combined population of Temotu ( 21 , 362 ) , and Rennell and Bellona ( 3041 ) comprises 4 . 7% ( 24 , 403/515 , 870 ) of the total population of the Solomon Islands , according to the 2009 national census [26] . We surveyed 959 households in 32 clusters throughout this EU . 4049 people were enumerated , and 3674 ( 91% ) consented to participate ( a total of 17% of the provincial population , average 4 . 2 people per household ) . Data was not collected on the reasons people did not take part . The examined population included 1135 children aged 1–9 years ( 53% male ) , and 2061 adults aged 15 years and over ( 42% male ) . The median age was 18 years ( min 1 , max 100 , inter-quartile range [IQR]: 8–38 years ) . In this population there were 397 ( 10 . 9% ) cases of TF , 5 ( 0 . 1% ) cases of TI , and 2 ( 0 . 1% ) cases of trichiasis in either eye of subjects of all ages identified by field grading . 84% of cases of TF were bilateral . In children aged 1–9 years , the prevalence of active trachoma ( defined as presence of TF and/or TI in either eye ) was 26 . 3% ( TF: 26 . 1% [296/1135]; TI: 0 . 2% [2/1135] ) . The proportion of males in this age group with active trachoma was significantly higher than that of females in the same age group ( 28 . 9% [176/608] versus 23 . 1% [122/527] , p = 0 . 027 ) . When adjusted for age and sex , the prevalence of active trachoma was 22% ( 95% confidence interval [CI]: 18 . 5–26 . 0% ) . In adults aged 15 years and over , the prevalence of active trachoma was 1 . 2% ( TF: 1% [21/2061]; TI: 0 . 1% [3/2061] ) . The prevalence of trichiasis in adults was 0 . 1% ( 2/2061 ) . When adjusted for age and sex , the prevalence of TT was 0 . 04% ( 95% CI: 0–0 . 3% ) . The field team did not recall one case of trichiasis and the other was documented as mild , with a single lash contacting the globe away from the cornea . Photographs were taken from 3110/3674 ( 85% ) of participants . Preliminary quality control yielded 2418 ( 78% ) photographs that were unsuitable for grading due to quality issues . A total of 692 photographs from study participants of all ages ( 238 in children aged 1–9 years ) were evaluated by an independent grader . Photo-grading according to the simplified grading system agreed with TF in 94% of cases leading to Fleiss’ kappa agreement scores of 0 . 88 , which indicated excellent agreement between photo grader and field grading in TF calls . Cases of TS were noticed in the photo set , although the evidence was insufficient to determine the population prevalence of this sign . Fig 1 shows exemplars of both mild and more severe TF in Solomon Island children , as well as a normal conjunctiva for contrast . The age-specific prevalence of TF cases is shown in Fig 2 . Swabs were collected from 1076/1135 ( 94 . 8% ) child participants aged 1–9 years , along with 41 field controls . Of those , 1002 ( 93 . 1% ) passed quality control by testing positive for the endogenous control gene H . sapiens RPP30 . All blank field controls and all known-negative extraction and PCR controls tested negative for endogenous control and microbial targets . Evidence of Ct infection was found in 13 ( 1 . 3% ) of 1002 specimens . Of those who tested positive for plasmid on diagnostic screen , 9 also tested positive for Ct chromosomal target omcB . Of the individuals whose swab was positive for the endogenous control RPP30 , 257/1002 ( 25 . 7% ) had TF and/or TI in the right eye ( Table 2 ) . The prevalence of Ct infection was 10/257 ( 3 . 9% ) in those with TF and/or TI , and 3/745 ( 0 . 4% ) in those without . Active disease status was highly significantly associated with current infection ( odds ratio: 10 . 0 , p = 0 . 0005 ) . While TF was observed in all 32 villages that were surveyed , we observed Ct infection in just eight villages . While the study was not designed to detect sub-EU-level differences in prevalence , post hoc analysis indicated there were significantly more cases of infection per capita in Rennell and Bellona than in Temotu province ( 6/131 [4 . 6%] versus 7/871 [0 . 8%] , respectively; Mann Whitney U: p = 0 . 0004 ) . TF levels in the 1–9 year old indicator group were in excess of 10% in both provinces included in the EU , but significantly lower in Rennell and Bellona ( 27 . 3% in Temotu versus 17 . 2% in Rennell and Bellona; Mann Whitney U: p = 0 . 0001 ) . The mean load of endogenous control target was 12 , 560 copies/swab ( IQR: 872–13 , 980 copies/swab ) . In positive swabs , the median load of Ct plasmid was 13 , 840 copies/swab ( IQR: 3599–84 , 990 copies/swab ) . There was a large difference in median load between Ct positive samples from children with active disease when compared to those without active disease , although the difference was not statistically significant ( median 13 , 840 versus 782 copies/swab; Mann Whitney U test p = 0 . 81 ) . The median omcB load was 7725 copies/swab ( IQR: 1696–22 , 110 copies/swab ) and the mean plasmid:genome ratio ( i . e . , plasmid copy number per bacterium ) was 4 . 4 ( IQR: 3 . 7–5 . 6 ) which is similar to that described elsewhere [24] . Sequencing was successful in 11/13 strains . The mean number of paired reads per specimen was 2 . 3 million ( IQR: 2 . 1–2 . 7 million; Table 3 ) . The median percentage of reads mapping to A/HAR-13 reference genome was 10 . 1% ( IQR: 1 . 5–24 . 4% ) per specimen . The median percentage of reads mapping to B/Jali20/OT reference plasmid was 2 . 1% ( IQR: 0 . 3–4 . 8% ) per specimen . Complete genome sequences ( >95% coverage ) were obtained from five of 13 specimens , whilst partial genome sequences ( <95% coverage ) were obtained from six specimens . Complete ( >95% ) or partial ( <95% ) plasmid sequences were obtained from eight and three specimens , respectively . One specimen failed to sequence . The median coverage of at least 1× read depth of the reference genome was 51 . 5% ( IQR: 12 . 4–98 . 2% ) ; the median coverage of at least 1× of the reference plasmid was 99% ( IQR: 61 . 4–99 . 7% ) . BLASTn analysis of five complete Solomon Islands Ct consensus genome sequences against archived Ct genome sequences revealed that all five had closest sequence homology and alignment to serovar A type ocular strains . BLASTn analysis of eight complete Solomon Islands plasmid consensus sequences revealed highest sequence homology with published serovar B ocular strain plasmids . Phylogenetic analysis of the complete Solomon Islands Ct genomes showed that these samples formed a single clade of closely related genotypes that formed a sub-clade of the ocular strains . The five complete genomes were more closely related to each other than they were to the nearest reference neighbour ( A/HAR-13 ) . Strains from the Solomon Islands were most closely related to the ocular serovars in the T2 chlamydial clade , as shown in Fig 3 . Bootstrapping supported our phylogram by indicating resampling did not alter the branch position in >80% of bootstrap runs . Outer membrane protein A ( ompA ) sequences from the ocular reference strains do not cluster in a single clade , as has been described previously [37] . OmpA orthologs from the five Solomon Island sequences were more closely related to each other than to their nearest neighbour which was the C-TW3 strain . Their relationship to other reference sequences is shown in S2A Fig . Additionally , tryptophan synthase alpha subunit ( trpA ) orthologs in Solomon Island sequences were most similar to ocular strains ( S2B Fig ) and featured a single nucleotide deletion leading to a premature stop codon and truncation of the open reading frame when compared to urogenital sequences . Finally , the relationship between the plasticity zone ( PZ ) of Solomon Island sequences compared to references was evaluated and is shown in supplementary Fig 2C . A low prevalence of TT has been observed in other populations in which TF was highly prevalent . Fig 4 shows data taken from published PBPSs in Nigeria [38–44] , Niger [45] , Sudan [46] , Kenya [16] , Ethiopia [47] and Cameroon [48 , 49] comparing the prevalence of TF in those aged 1–9 years with the prevalence of TT in 15+ year-olds in the same EUs , all of which were treatment-naïve . Of 58 identified EUs with comparable prevalence of TF in the 1–9 year olds ( 10% < TF < 40% ) to that observed in the current study , 50% had a TT prevalence greater than 1% in those over the age of 15 years; the median TT prevalence was 1% , compared to 0 . 1% in our survey . Our population had a high TF prevalence when compared to other districts in this analysis with <1% TT . Published data from studies in Ethiopia [50] , Niger [50] , Tanzania [18] , Gambia [18] , Cameroon [48] , Mali [51] and Brazil [52] indicate that in other trachoma-endemic populations , the highest age-specific TF prevalence is generally in those aged 3–4 years . When published age-specific TF profiles from other parts of the world are compared to the age distribution in this study ( Fig 5 ) , the peak age-specific TF prevalence in our data is shown to be in an older age group . We identified a number of previous studies have reported concomitant Ct infection and active trachoma prevalence estimates [19–21 , 53–70] in children aged 0–9 years or a subset of that group in 35 districts . At the population level there is a good correlation between the two ( R = 0 . 84 ) ( Fig 6 ) . Our Ct infection estimate does not conform to the patterns observed in those other populations , with infection being substantially less prevalent than might be expected given the TF prevalence .
We report an apparently mild trachoma phenotype in which TF is moderately prevalent yet Ct infection , TI and TT are rare . The findings reflect those of other studies; over 2300 adults were surveyed in Makira , Isabel and Central provinces yet only 3 cases of TT identified , indicating an unadjusted prevalence among those adults of 0 . 1% despite 22 . 2% of the 1–9 years population in the same survey having signs of TF [12] . The TF prevalence in Temotu , Rennell and Bellona is the highest of the populations surveyed during GTMP mapping . Choiseul province had low levels of both TF ( 6% ) and TT ( 0% ) . Interestingly , Western Province had more TT cases than Temotu , Rennell and Bellona [71] . Further studies are warranted to determine whether infection prevalence is also higher in that region . The prevalence of TF in these communities qualifies this EU for priority implementation of the A , F and E components of the SAFE strategy ( surgery to treat TT , mass antibiotic distribution to treat infection , promotion of facial cleanliness and environmental improvement to reduce transmission [72] ) , but the trichiasis data suggest the elimination target for TT has already been met . The evidence of this survey suggests that prevalence of TF in children may not be an appropriate marker of disease burden in this setting . Prolonged infection with Ct , intense transmission of Ct , presence of TI and other markers of inflammation have been associated with progression to TS [3 , 73–75] , the precursor of TT . It is therefore feasible that the low prevalence of TI and Ct are related to the paucity of TT in this population . In Fig 4 we demonstrate that the correlation between reported TF and TT prevalence is weak , suggesting that high TF prevalence is not always indicative of a significant TT burden . This is not an unexpected finding; the signs of TF are transient , being instigated and cleared over the course of weeks or months , whereas TT has a much longer-term onset and is influenced by an accumulation of a lifetime of microbiological and immunological stimuli . Field grading has not previously been standardised across studies and must therefore be compared between studies with caution . TF is used in part for ease and uniformity of field grading rather than specificity for chlamydial infection , and Fig 6 shows that there is a moderate correlation between population Ct infection prevalence and population TF prevalence in most districts studied to date . The search criteria for this literature search were relatively lenient; controlling for field grading discrepancy , sample collection methodology , and diagnostic test would likely result in a stronger correlation . At the individual level , TF and Ct prevalence do not always correlate well; for example in a rural Tanzanian community where the TF prevalence was >10% , only 6 . 1% of those with TF had Ct infection and no association was found with clinical signs disease [74] . Follicular inflammation of the conjunctiva can have many different causes , such as viruses , nonchlamydial bacteria , chemical exposure and allergic reactions [76] . While detailed eyelid examination may be able to distinguish these infections phenotypically , the WHO simplified grading system is not sufficiently detailed to do so . Streptococcus pneumoniae and Haemophilus influenzae have been shown to be significantly associated with TF in the Gambia and Tanzania [74 , 77] . Numerous other species of the Staphylococcus , Streptococus , Moraxella , Hameophilus and Corynebacteria genera among others have been cultured from the conjunctivae of children and adults living in trachoma-endemic areas . Although many have not been shown to associate with clinical signs of TF , there are indications that in adults these bacteria can drive inflammation which leads to increased scar tissue deposition [78] or recurrent TT after surgery [79 , 80] . It is therefore likely that a proportion of all TF cases globally may not be chlamydial in origin; it seems that in this Pacific Island setting , this proportion is high and this is translated into a reduced prevalence of end-stage trachomatous disease . Fig 4 also indicates that this may also be true of other parts of the world , where districts with sufficient TF to qualify for intervention under WHO guidelines do not necessarily have a significant TT burden . Exposure to circulating Ct may modulate the immune response at the conjunctiva to increase inflammation in response to otherwise commensal organisms . In turn this could drive the immunopathology that leads to scarring . It is possible that the low prevalence of Ct observed in this population is insufficient to drive intense transmission , and children are exposed less frequently than in other trachoma-endemic populations and therefore are not as susceptible to such intense or regular periods of inflammation . The highest age-specific TF prevalence in this study was in those aged 6 years ( Fig 1 ) . The difference between that and other published data , as documented in Fig 5 , may imply a different mode or intensity of transmission , or may reflect reduced accumulation of partial immunity . It is not possible to determine the true mechanism without intensive longitudinal study , but this observation supports the case that the epidemiology of TF in our population is atypical . While the majority of our Ct-positive swabs were taken from eyes with active trachoma , we consider the low absolute number of infections insufficient to drive the moderate burden of TF . It is not clear from our cross-sectional study whether a non-chlamydial microbial agent is causing TF , or whether those with TF had suffered a relatively recent Ct infection and had persistent inflammatory disease causing the follicular inflammation we observed . The use of ddPCR has not yet become widespread in infectious disease studies . While it is not suitable for all applications , it offers the significant benefit of reference-free quantification of nucleic acids . In the present study , the load of Ct in positive samples was substantial , which are thought to be more transmissible than low-load infection . There was also substantially higher infection load in those with TF as compared to those without TF , although this difference was not statistically significant . Despite recent advances in culture-free sequence methodologies , low-load infections are known to yield poor quality or no sequence data . The technique we used in this study reportedly provides high quality sequence data ( 20× read depth over at least 95% of the genome ) when the input specimen has above ~12 , 000 and ~98 , 000 chlamydial genome copies in vaginal swab and urine samples , respectively [25] . A lower load limit for ensuring high quality sequencing in ocular samples is not yet known , but we were encouraged that partial or complete sequence data were yielded from 11/13 Ct positive swabs strains that were sequenced . Those where complete genome coverage ( >95% ) was achieved appeared to be most closely related to ocular serovars , and appeared to be very closely related to each other . The sequence information suggested that , at trpA and the wider PZ , the Solomon sequences were closely related to ocular strains , and ocular and urogenital strains were distinct from each other at these loci . The trpA open reading frame was truncated in these sequences . This region contains key determinants of Ct tissue tropism and further supports the close relationship of these strains to classical ocular references . The small number of sequences available makes it difficult to identify differences potentially related to pathogenicity . Urogenital strains are known to be able to infect conjunctival epithelium [81] , and given the high prevalence of sexually transmitted Ct infections in the Solomon Islands [82] , we may have expected some contamination of the conjunctivae with urogenital chlamydial strains . Our data did indicate urogenital strains were present in several conjunctivae , but the quality of those sequences aligning to urogenital references was uniformly low . It is not possible to determine whether this was because these were urogenital strains that had not established a sufficiently fulminant infection to yield enough material for sequencing , or whether the matches obtained were an artefact of the low sequence coverage . We can say , though , that our next generation sequencing confirmed that strains with high sequence homology to well-defined ocular Ct strains are present in the conjunctivae of children in the Solomon Islands . One limitation of our study is the absence of an alternative explanation for the discrepancy between Ct and TF levels . Only samples for which testing passed various quality control steps were included in this paper and our test had been previously validated against an external standard . We therefore do not believe that simple diagnostic failure has significantly influenced our data . Of the four signs of trachoma described in this paper ( TF , TI , TT and current infection ) , three ( infection , TI and TT ) are present at low levels in this population . Further studies are underway to test for potential alternative pathogens such as S . pneumoniae and H . influenzae , and we are investigating longer-term markers of Ct infection by screening the population of Temotu , Rennell and Bellona for both trachomatous scarring and antibodies against chlamydial pgp3 antigens . We have not addressed the genetics of the population in this study and while host genetic factors have been shown to associate with an increased risk of scarring [83 , 84] very little is known about diversity in immune response genes in the Solomon Islands . The limited amounts of immunogenetic typing data that are available indicate that some HLA epitopes associated with increased risk of scarring ( e . g . HLA-C2 ) are moderately prevalent in the Pacific region , while other putative protective alleles ( e . g . the HLA-B*08:01~C*03:04 haplotype ) are almost absent ( data taken from allelefrequencies . net , search March 2016 ) . Investigating this is beyond the scope of this study , but accumulated evidence on the genetics of trachoma indicate that both pathogen and host are sufficiently well adapted to coexistence that a host polymorphism that makes the host entirely refractory , or pathogen variation that completely ameliorates the infectivity and/or pathogenicity of Ct seems unlikely . Polymorphisms in key immune genes such as IL-10 and gamma-interferon have been shown to be more frequent in cases with severe trachoma than in normal controls [85] , although these were not replicated in genome-wide association screening . It is possible that variation in immune responsiveness may influence the susceptibility of this population to Ct infection but a specific immune pathway that is expressed significantly differently between those in whom scarring progresses and those in whom it doesn’t has yet to be identified [75] . In addition to ocular Ct infection , we observed signs of both active trachoma and trachomatous conjunctival scarring in this sample indicating trachoma is or has recently been endemic in these islands . However , the prevalence of Ct infection appeared to be too low to be the sole explanation for the high burden of TF , while TI and TT were curiously scarce given the substantial amount of TF that was present . Whilst Ct may account for some of the TF in this population , we expect that the majority of TF-like disease is either caused by a single as-yet undetermined factor , or by multiple contributory aetiologies . This form of disease , if not unique to the Solomon Islands , might inflate estimates of trachoma burden and could lead to unwarranted mass drug administration in other world populations . In several European settings , there has been a steady increase in incidence and reinfection rates of urogenital Ct despite enhanced detection and treatment , which some have hypothesised could be attributed to interruption of the natural acquisition of immunity [86] . We do not have good markers of what constitutes ‘acquired immunity to Ct’ to measure this , but it is relevant in this context to consider the possibility of negative effects of MDA in addition to the potential positive ones . The findings of this study may have profound impacts on approaches to trachoma programme monitoring in the peri-elimination period . | Trachoma is the most common infectious cause blindness worldwide , and the target of a global elimination initiative . A package of community-wide interventions is recommended to treat trachoma , which aim to reduce transmission of the causative agent Chlamydia trachomatis . These interventions require significant , prolonged investment . Clinical observation of follicles on the underside of the eyelid is used to assess requirement for and success of intervention . However , we now know that there are nonchlamydial causes of these follicles , so the observation of this clinical sign may not be specific for chlamydial infection . A recent study showed that infection testing can be cost-effective in low infection prevalence settings , as the money spent on infection testing is more than offset by the savings from avoiding additional rounds of community-wide interventions . We show here that , despite a high prevalence of clinical signs of disease , the estimated prevalence of chlamydial infection is low . The prevalence of the sight-threatening end stage of disease is also low , so we must consider whether the costly community-wide interventions are appropriate in this setting . The use of molecular tools at the population level to guide trachoma policy is still under investigation; this study will contribute to the pool of data required to assess the utility of these tools . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"sequencing",
"techniques",
"medicine",
"and",
"health",
"sciences",
"solomon",
"islands",
"pathology",
"and",
"laboratory",
"medicine",
"chlamydia",
"trachomatis",
"pathogens",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"microbiology",
"bacterial",
"diseases",
"signs",
"and",
"symptoms",
"eye",
"diseases",
"sexually",
"transmitted",
"diseases",
"chlamydia",
"neglected",
"tropical",
"diseases",
"molecular",
"biology",
"techniques",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"infectious",
"diseases",
"chlamydia",
"infection",
"inflammation",
"sequence",
"alignment",
"medical",
"microbiology",
"microbial",
"pathogens",
"molecular",
"biology",
"immune",
"response",
"people",
"and",
"places",
"diagnostic",
"medicine",
"oceania",
"ophthalmology",
"biology",
"and",
"life",
"sciences",
"trachoma",
"organisms"
]
| 2016 | Low Prevalence of Conjunctival Infection with Chlamydia trachomatis in a Treatment-Naïve Trachoma-Endemic Region of the Solomon Islands |
We model the regulatory role of proteins bound to looped DNA using a simulation in which dsDNA is represented as a self-avoiding chain , and proteins as spherical protrusions . We simulate long self-avoiding chains using a sequential importance sampling Monte-Carlo algorithm , and compute the probabilities for chain looping with and without a protrusion . We find that a protrusion near one of the chain’s termini reduces the probability of looping , even for chains much longer than the protrusion–chain-terminus distance . This effect increases with protrusion size , and decreases with protrusion-terminus distance . The reduced probability of looping can be explained via an eclipse-like model , which provides a novel inhibitory mechanism . We test the eclipse model on two possible transcription-factor occupancy states of the D . melanogaster eve 3/7 enhancer , and show that it provides a possible explanation for the experimentally-observed eve stripe 3 and 7 expression patterns .
Polymer looping is a phenomenon that is critical for the understanding of many chemical and biological processes . In particular , DNA looping has been implicated in transcriptional regulation across many organisms , and as a result plays a crucial role in how organisms develop and respond to their environments . While DNA looping has been studied extensively over the last several decades both experimentally [1–5] and theoretically [6–11] , many aspects of looping-based transcriptional regulation remain poorly understood . In vivo , the simplest looping-based regulatory architecture is comprised of a protein , or activator , that interacts with a distal site via DNA looping . Such simple architectures can be found in bacteria , where , for example , σ54 ( σN ) promoters are activated via such a mechanism [12–14] . In eukaryotes , DNA-looping-based regulation is associated with the interaction between the core promoter and distal regulatory regions called enhancers . These ∼500 bp enhancers typically contain clusters of transcription factor ( TF ) binding sites , and may be located from 1 kbp to several Mbps away from their regulated promoters . Detailed studies of enhancers from several organisms [15] have revealed that TFs can upregulate , inhibit , or both upregulate and inhibit gene expression via a variety of mechanisms . For example , repressors like D . melanogaster Giant , Knirps , Krüppel , and Snail inhibit expression either by partial overlap of their binding sites with that of an activator , or via a short-range repression mechanism termed “quenching” , whereby TFs positioned several tens of bps either upstream or downstream from the nearest activator inhibit gene expression [16–19] . In quenching , a bound protein inhibits gene expression , apparently without any direct interaction between the protein and either the promoter or the nearest-bound activator . In the prevailing model , quenching is assumed to be a result of histone deacetylation , which is facilitated by the formation of a ∼450 kDa DNA-bound chromatin remodeling complex made of a DNA-binding protein such as Knirps , C-terminus binding protein ( CtBP ) [20–22] , and the histone deacetylases ( HDACs ) Rpd3 [23] and Sin3 [24] . However , this model falls short of providing a full description of quenching . In particular , short-range repressors such as Knirps have been shown to retain a regulatory function even without the CtBP binding domain [22 , 25 , 26] , and tests for HDAC activity using either Rpd3 knock-outs [27] or the HDAC inhibitor trichostatin A failed to alleviate the observed repression effects [26] . Consequently , the mechanistic underpinnings of quenching are still poorly understood . Previously , we studied looping-based regulatory mechanisms for synthetic bacterial enhancers , in which the activator was bound close to the promoter ( within ∼1 Kuhn length , or ∼300 bp ) [28] . We successfully modeled the experimental results using a modified worm-like chain model that takes excluded-volume considerations into account [29–31] . We found that the regulatory effects are maximal for TFs bound at the center of the looping segment ( delimited by the activator and the promoter ) , and decrease as the looping segment length is increased . In this report , we focus on regulation for looping segments much longer than the Kuhn length , which are more typical for eukaryotic enhancers . In this case , the elastic regulatory effects that we observed previously are negligible . Using the modified worm-like chain model , we show that the excluded volumes of DNA and a TF bound within ∼1 Kuhn length ( ∼300 bp ) either upstream or downstream of one of the loop termini can block the “line-of-sight” of the other terminus , generating an eclipse-like effect , which reduces the probability of looping . Unlike the elastic effects that we reported previously , this eclipse-like effect is independent of looping-segment length for sufficiently long looping segments . Thus , our model offers a looping-based mechanistic model for quenching repression .
Linear polymers have been studied via simulations using a variety of methods [32 , 33] . To model DNA fully it is necessary to take into account the experimentally observed non-entangled chromosomal DNA structure [34] , and to include details regarding bound proteins . Recently , polymer rings have been studied as a model system for topologically constrained polymer melts , such as chromosomal DNA [35 , 36] . In this work , we address looping of a linear polymer , neglecting the additional topological constraints , but including bound protrusions . We chose to simulate a linear DNA chain with a bound protein using a sequential importance sampling Monte-Carlo approach that we used previously to simulate the configurational space of bare DNA [29] . To adapt our algorithm to the case of protein-bound DNA , we take into account not only the growing chain but also the location of the protrusion ( see Eq ( 4 ) ) . During chain generation , upon reaching link k , the simulation adds a hard-wall spherical protrusion with radius Ro at the location robject . If the protrusion overlaps any of the previously-generated chain links or protrusions , the chain is discarded . After generating the configurational ensemble , we identify the subset of “looped” chains . We provide the essential details of the simulation in this Section . Additional details can be found in S1 File .
To model the quenching effect of a bound repressor on DNA looping , we generated configurational ensembles for DNA with a spherical protrusion of size Ro = 9 . 2 nm or Ro = 18 . 4 nm located a distance of K = 95 bp or K = 135 bp from the chain origin along the chain , oriented either in the same direction as the looping volume δr or 180° from it ( see Fig 1 ) . We plot F ( L ) for the various configurations of Ro and K in Fig 2A . The data show that in the elastic regime ( L ≤ b ) , protrusions bound in-phase with δr ( solid lines , ↑ ) strongly reduce the looping probability relative to that of the bare DNA , while protrusions positioned out-of-phase to δr ( dashed lines , ↓ ) increase the looping probability , as we showed previously [28] . However , in the entropic regime ( L ≫ b ) , all chains converge to values of F → L ≫ b F ∞ ≤ 1 . For protrusions that are bound in-phase ( i . e . γk = 0° ) , F∞ is distinctly smaller than one , and strongly depends on both the distance to the nearest terminus and the size of the protrusion . Conversely , for protrusions bound out-of-phase ( γk = 180° ) , F∞ is only slightly smaller than 1 , with weak dependence on both protrusion size and position . To further explore the extent of the quenching effect in the entropic or long-chain-length regime , we plot in Fig 2B the value of F∞ as a function of a wide-range of Ro and K , for γk = 0° . The heatmap shows both a non-linear decrease in F∞ as a function of protrusion size and a non-linear increase as a function of protrusion distance from the chain origin . The red line in the figure demarcates the closest possible location at which a protrusion can be bound without physically penetrating part of the looping volume δr . Thus , for volumes and protrusion positions that fall below this line , a second excluded volume effect contributes to the reduction in the probability of looping , leading to a sharp increase in the overall effect . Together , the panels in Fig 2 show that a sufficiently-large protrusion can substantially reduce the probability of looping , independent of loop length , provided that its binding site is within a small distance from either of the loop termini . For further details regarding the computation of F∞ and its error estimation , see S1 File . To understand the long-range , length-independent effect shown in Fig 2 , we examine the chains’ terminating segments of length T ≪ L . If T ∼ max r ∈ δ r | r − r o b j e c t | ≪ b , these segments resemble stiff rods , and the object obstructs the line-of-sight of one chain terminus from the other . This eclipse-like phenomenon is manifested by a reduction in the number of polymer chains that are able to reach δr . This , in turn , results in a smaller Plooped as compared with the case in which no protrusion is present . In the entropic regime and in the absence of protrusions , the generated “rods” approach the looping volume δr from all directions that are unobscured by the volume of the polymer in a homogeneous fashion [7] . Due to this isotropy in the distribution of the chain termini orientations within δr , the reduction in Plooped can be approximated by the solid angle that the eclipsing object subtends at δr . Consequently , F ( Eq ( 6 ) ) can be approximated for this “rod model” by: F ∞ r ˜ ′ , R o = P looped object L , { r ˜ ′ , R o } P looped baseline L L ≫ b ≈ 4 π δ r - I chain - I object r ˜ ′ , R o + I chain ∩ object r ˜ ′ , R o 4 π δ r - I chain = 1 - I object r ˜ ′ , R o 4 π δ r - I chain + I chain ∩ object r ˜ ′ , R o 4 π δ r - I chain , ( 12 ) where Ro is the radius of the spherical object , r ˜ ′ is the location of the object , which could be located statically at point r′ ( in which case r ˜ ′ ≡ r ′ ) , or located on the chain a distance K from the chain origin ( in which case we use the terminology r ˜ ′ ≡ r ˜ ′ ( K ) figuratively to specify the progression of the protrusion along the chain ) . ℐ c h a i n ≡ ∫ δ r Ω c h a i n ( r ) d 3 r , where Ωchain ( r ) is the solid angle subtended at r by the polymer chain links . ℐ o b j e c t ( r ˜ ′ , R o ) ≡ ∫ δ r Ω o b j e c t ( r ˜ ′ , R o ) ( r ) d 3 r , where Ω o b j e c t ( r ˜ ′ , R o ) ( r ) is the solid angle subtended at r by the object , and ℐ c h a i n ∩ o b j e c t ( r ˜ ′ , R o ) corresponds to the solid angle contained in both ℐ c h a i n and ℐ o b j e c t . In order to test the eclipsing hypothesis , we first computed F∞ for the case of an object statically positioned at an off-chain location r ′ = d u ^ 0 , and without chain-chain interactions . In this simplified case , F∞ in Eq ( 12 ) can be approximated by the following eclipsing expression: F ∞ r ′ , R o ≈ 1 - I object r ′ , R o 4 π δ r = 1 - ∫ δ r 2 π 1 - 1 - R o + w / 2 r - r ′ 2 d 3 r 4 π δ r ≡ F ∞ static , ( 13 ) where we substituted Ω object 2 π = ∫ 1 - R o + w / 2 2 | r - r ′ | 2 1 d cos θ = 1 - 1 - R o + w / 2 2 | r - r ′ | 2 . ( 14 ) In Fig 3 , we compare the value computed from Eq ( 13 ) ( dashed cyan line ) to F ( L ) computed by our sequential importance sampling algorithm for the same conditions ( solid blue line ) . The data show that the eclipsing approximation F ∞ s t a t i c overestimates F∞ . We reasoned that the main cause for this estimation error is that Eq ( 13 ) disregards the flexible polymer nature of the chain . We ran an additional Monte-Carlo simulation to quantify the correction resulting from polymer flexibility . Here , we generated pairs consisting of an end-terminus point in δr and a direction vector of the terminal link , both distributed uniformly . Short polymer chains of length T originating at the chosen points were grown with their first links oriented in the chosen directions . These chains can be thought of as the terminating segments of long chains that have a uniform distribution of their end-termini in δr . We found that the probability of a flexible-polymer chain to overlap the object increased relative to the probability within the “rod model” , resulting in a decrease in the probability of the chain to form a loop ( magenta dashed line in Fig 3 ) . Using this “terminating-segments” correction , the discrepancy between F∞ from the simulation and F ∞ s t a t i c from Eq 13 is partially accounted for . We attribute the additional reduction in the simulated F∞ to interactions between the object and the remaining L − T length of the chain . In Fig 4 we plot F∞ as a function of an on-chain object of radius Ro , for several values of K . To compare the results of the numerical simulation to the full eclipsing model ( Eq 12 ) , we first note that when K is kept constant , ℐ c h a i n ∩ o b j e c t ( r ˜ ′ , R o ) | K = c o n s t ≈ c o n s t , as can be seen from the inset in Fig 4: the overlap between Ωchain ( orange cones ) and Ωobject ( red cones ) changes only slightly when the object grows by a factor of two . Furthermore , | r ˜ ′ ( K ) − r | K = c o n s t is approximately independent of Ro if ( R o + w / 2 ) ≪ | r ˜ ′ ( K ) − r | for all r ∈ δr . Thus , the dependence of F∞ on the radius Ro of an on-chain object can be derived from Eq ( 12 ) : F ∞ r ˜ ′ , R o | K = c o n s t ≈ ≈ 1 - I object r ˜ ′ , R o 4 π δ r - I chain + I chain ∩ object r ˜ ′ , R o 4 π δ r - I chain ≈ 1 - A I object r ˜ ′ , R o + B K ≈ 1 - A K R o + w / 2 2 + B K ≡ f K R o , ( 15 ) where we approximated ℐ o b j e c t ( r ˜ ′ , R o ) ≈ ( R o + w / 2 ) 2 π ∫ d 3 r | r ˜ ′ − r | 2 using Eq ( 14 ) and ( R o + w / 2 ) ≪ | r ˜ ′ − r | r ∈ δ r . In Fig 4 , we fit the numerical results for different values of K with functions of the form fK ( Ro ) ( Eq ( 15 ) ) . The fits are in excellent agreement ( R2∼ 0 . 99 ) with the numerical data . A salient feature of enhancers is that binding sites for TFs are positioned both upstream and downstream of the activators . The eclipse model predicts that F ∞ ( r ˜ ′ ( K ) , R o ) = F ∞ ( r ˜ ′ ( − K ) , R o ) , since for long chain lengths the correlation between the positions of both chain termini is completely abolished . Thus , from the perspective of one terminus , looping events can be initiated from any possible direction . To see if our simulation captures this symmetry , we explored a geometry in which the protein-like protrusion was positioned at negative K values , corresponding to a location outside of the looping segment delimited by the looping-volume ( activator ) and distal-terminus ( promoter ) locations . To do so , we generated an additional chain segment of length Q in the direction opposite to t ^ 1 , starting from link 0 , where Q ≫ K . We plot the results in Fig 5 . The data show that in the elastic regime , the “outside” geometry ( green line , K < 0 ) generates a significantly smaller effect on the probability of looping as compared with the “inside” architecture ( blue line , K > 0 ) . This lack of symmetry is probably due to the propensity of short loops to form a “tear-drop” shape , thereby reducing the quenching effect of any elements bound outside the looping segment [28] . However , for sufficiently large L/b , F ( L ) for both ±K enhancer geometries converge to the same value , as predicted by the eclipse model . So far , we modeled an enhancer-promoter region by a chain of discrete semi-flexible links , with the activator and the promoter located at both ends of the chain . However , in vivo , the DNA chain extends far beyond the enhancer-promoter region in both directions , and is subject to confinement . We address the first issue in this section , and the issue of DNA confinement in the next section . We computed the probability of looping for a generic loop of length L , located between two internal links of the chain . To do so , we extended the original length of the chain L by two flanking segments of 10 Kuhn lengths ( 10b ) on both ends of the chain . We assigned the cumulative Rosenbluth weight of the extended chain to the central segment of length L containing the enhancer-promoter region . See S1 File , Section 1 . 3 . We ran a simulation with K = 60 bp and R0 = 11 . 9 nm ( Fig 6 , magenta line ) and compared F ( L ) to the same case without flanking segments ( Fig 6 , blue line ) . The data show that there is a discrepancy in the long-range looping probability ratio between the two looping models . In particular , the down-regulatory effect of the protrusion is stronger when the enhancer-promoter region is modeled as part of a larger chain . In order to determine the individual contributions of the two flanking segments to the discrepancy , we ran additional simulations with different flanking-segment configurations . We found that only the trailing segment beyond the distal terminus of the looping segment contributes to this discrepancy . The addition of a single chain link of length w = 4 . 6 nm ( 0 . 092b ) after the terminus of the looping segment ( Fig 6 , green line ) diminishes F ( L ) , accounting for approximately half of the discrepancy . The addition of only 4 links ( 0 . 368b ) after the terminus of the looping segment ( Fig 6 , aqua line ) accounts for the entire discrepancy , fully agreeing with the results for the case with 10b flanking segments on both ends of the looping segment . While an addition of a leading segment before the looping volume diminishes the looping probability ( data not shown ) , it does not alter the looping probability ratio F ( L ) . This can be seen from the comparison between F ( L ) for the case with a single link after the terminus of the looping segment ( Fig 6 , green line ) and the case of a segment of length b before the looping volume and a single chain link after the terminus of the looping segment ( Fig 6 , red line ) . We believe that these results depend strongly on the looping conditions . For the conditions used here ( δω′ = 0 . 1 × 2π , see Fig 1 ) the leading flanking segment is inconsequential to the looping probability ratio . However , utilizing a uniform looping volume ( δω′ = 4π ) would lead to a diminished looping probability ratio as a result of leading segment addition . Similarly , the trailing flanking segment diminishes F ( L ) due to the fact that chain configurations that approached the looping volume from a direction colliding with the chain are no longer possible when the chain is extended by a trailing segment . If , however , the looping conditions were such that the chain terminus was required to approach the looping volume from a direction perpendicular to the looping volume cone axis , the trailing segment would become inconsequential to the looping probability ratio . These results suggest that while there is a discrepancy between the two simulation modes ( with flanking regions and without them ) , it arises from regions that are immediately adjacent to the enhancer-promoter region , and further downstream or upstream segments of the chain do not contribute to the down-regulatory effect . To check the effects of polymer confinement on the looping probability , we generated chains confined by spheres of various radii ( for full simulation details , see S1 File ) . In Fig 7 , we present the normalized looping probability P l o o p i n g s p h e r e / P l o o p i n g f r e e and the looping probability ratio for the confined chains , for the range of confining-sphere radii 125–625 nm . Note that the looping probability increases relative to that of the unconstrained polymer when the gyration radius of the chain becomes comparable to the size of the confining sphere ( Fig 7A ) . However , as can be seen in Fig 7B , the introduction of a bound protrusion does not influence the looping probability ratio F ( L ) up to a chain length of 10 kbp for confining spheres with radii ≥250 nm . For the smallest confining sphere ( radius of 125 nm ) we were only able to simulate chains of length ≤4 . 5 kbp ( see S1 File Section 3 . 4 ) , and here too the looping probability ratio remained unaffected . Finally , we applied our model to a real enhancer-promoter system . We chose the experimentally well-characterized eve 3/7 enhancer of D . melanogaster , which is separated from its promoter by about 4 kbp . See Discussion for a detailed examination of the validity of this application . We computed our model’s predictions for the active eve 3/7 enhancer , i . e . in the early developmental stages ( before gastrulation—stage 14 ) , and in the anterior region of the D . melanogaster embryo . To check whether the resultant probability of looping could provide a mechanistic explanation for the observed expression pattern , we simulated two states: first , a chain representing the enhancer that is fully occupied by spherical protrusions representing the activators dStat , Zld and the Knirps repressor–co-repressor complexes , and a second state where the chain is entirely devoid of bound protrusions . We tested three possible protrusion sizes , corresponding to reported repressor complex sizes: Knirps alone ( 46 kDa ) , Knirps bound to CtBP dimers ( 130 kDa ) , and a full putative 450 kDa complex , as reported by [23] . Since the exact looping conditions of the enhancer looping are unknown , we considered three choices of δω′ ( see Materials and Methods ) that represent different extreme cases . The results , plotted in Fig 8 , show that for all chosen looping conditions , there is a decrease in the probability of looping when the chain is fully bound by the protrusion representing Knirps . However , only the protrusion representing the full Knirps–co-repressor complex ( containing dCtBP and the HDACs Rpd3 [23] and Sin3 [24] ) was large enough to generate a significant reduction in looping .
We previously established [28] that a bound protein inside a loop can alter the probability of looping in a manner proportional to its size , when the chain length is of the order of the Kuhn length ( i . e . < 300 bp ) . The simulations and theory presented here identify a separate regulatory effect that is relevant to much longer chain lengths . In particular , our model predicts a decrease in the probability of looping that is independent of chain length for long chains in the entropic regime ( i . e . L ≫ 300 bp ) , provided that a sufficiently large protrusion oriented in-phase with the looping volume δr is positioned within one Kuhn length of one of the chain termini . We further showed that the reduction in looping probability resembles an eclipse-like phenomenon , where the protrusion blocks the line of sight of one chain terminus from the other . Our model provides a biophysical mechanism for so-called short-range repression or quenching by enhancers , for sytems with sufficiently long separation between the enhancer and the core promoter ( L ≫ b ) [16–19] . The model successfully captures many of this phenomenon’s salient features . These include lack of dependence on chain length , symmetry with respect to binding-site positioning inside or outside of the looping segment , the dependence of the regulatory effect on both the size of the bound complex and the distance of the TF from the nearest terminus , and the typical distances of the TF from the terminus ( ≲ 150 bp ) for which significant quenching can be generated . We computed the looping probability ratio for the D . melanogaster eve 3/7 enhancer-promoter system in the segment of the embryo where the repressor Knirps is expressed . We found that using realistic repressor-co-repressor-HDAC complex sizes for the simulated protrusions , a significant reduction in looping probability can be obtained , showing that the looping-based mechanism can account for at least some of the quenching generated by Knirps in the context of early fly development . Since there is also a substantial body of evidence supporting HDAC catalysis as a major mechanism for repression [40] , we conclude that it is likely that both HDAC activity and looping-related effects combine to generate the quenching observed in D . melanogaster and other organisms . There have been previous attempts to model enhancer regulatory logic for the gap genes of early fly development [41–43] . Using a semi-empirical approach , which coupled a thermodynamic model to an empirically-based regulatory scoring function , these works demonstrated that TF occupancy of enhancers can determine the gene-expression outcome . By contrast , our model does not compute the actual expression pattern , but rather the probability of looping for a given occupancy state . For this computation , our model requires only knowledge of the enhancer structure ( TF binding sites , protein-complex sizes , and over-all enhancer-promoter distance ) . This implies that no gene-expression experimental data is needed as an input to predict the down-regulatory effect of bound TFs . While this work focused on bare dsDNA with a protrusion , our results are applicable to any linear polymer as long as its length is significantly larger than its Kuhn length . However , to apply our results to chromosomal DNA , other constraints must be taken into account , some of which we attempted to address in this paper . First , natural DNA loops occur between sites within the chromosome , and not between ends . We showed that looping of a linear segment with additional upstream and downstream flanking segments qualitatively resembles looping without flanking segments . Second , chromosomal DNA is subject to confinement . Third , DNA is typically a mixture of chromatinized and non-chromatinized DNA . We adress the second and third points below . Due to confinement , chomosomal DNA is compacted into a globular state . The nature of the globular state varies between organisms ( equilibrated globule in yeast vs . a possibly non-crosslinked globule in higher eukaryotes [34 , 44–46] ) and chromatin spans a wide range of volume fractions [36] . For lower volume fractions , DNA can be viewed as a semi-dilute polymer solution , in which coils are strongly overlapping . This semi-dilute solution may be pictured as a system of domains ( blobs ) . Inside each blob , the chain behaves as an isolated macromolecule with excluded volume , and different blobs are statistically independent of one another [8 , 47] . If the volume fraction is sufficiently low , chromosomal DNA within a blob can explore nearly the entire volume of the blob without interacting with other parts of the chromosome . Our model is applicable to non-chromatinized dsDNA within a blob , provided that the characteristic blob size of the organism is large enough to contain a sufficiently-long length of dsDNA . For active eukaryotic promoters , there is evidence that the enhancer and the promoter regions are depleted of nucleosomes for up to 500–1000 bp both upstream and downstream from the center of each regulatory element [48] . This implies that the ends of the looping segment are non-chromatinized dsDNA , while the looping segment itself may be chromatinized . We argue that our model may be applied also for this case , provided that the looping segment is contained within a blob . This is because the eclipse effect is a result of the loss of correlation between the two chain termini . Thus , the effect should occur for any sufficiently-long chain ( Lchain ≫ bchain ) , independent of the exact chain parameters ( Lchain , bchain , wchain , etc . ) . The effect is sensitive to the TF size and location , and to the looping criteria . However , these are assumed to equal the conditions of non-chromatinized DNA at the chain termini ( i . e . in close proximity to the enhancer and promoter ) . We calculated typical blob sizes for a few model organisms ( see S1 File Section 2 ) . For bacteria , there are few examples of looping with lengths considered in this work , while the vast majority of the loops are short and are covered by the model described previously [28] . For D . melanogaster , a typical blob contains many kbp of chromatinized DNA . Thus our model may be applicable to enhancer-promoter systems in D . melanogaster and other organisms with low chromatin volume fractions , and to the eve 3/7 enhancer in particular . For enhancer-promoter systems for which a non-interacting blob cannot be assumed ( small blob size due to high chromatin volume fraction , or very long looping segment ) , we must consider the interaction of the looping segment with the rest of the chromosome . Recent results regarding chromatin structure [34] indicate that regions within eukaryotic chromosomes are unentangled but do interact ( for example , by confining each other sterically ) . We attempted to simulate such confinement using a hard-wall sphere around the simulated chains . We found that the probability of looping is increased by confinement . This is perhaps not surprising , since regions of the order of 1 Mbp within chromosomes have been found to interact with a contact probability with power-law constant ∼-1 [34 , 44] , as compared to the ∼-1 . 5 prediction of the equilibirum globule model [8] . This behavior can be explained by the formation of self-organized non-entangled regions , which can form via a variety of mechanisms [45 , 46 , 49 , 50] . Unlike the looping probability , the eclipse effect was found to be unaffected by confinement , up to the minimal confining sphere size that we were able to simulate . Note that the simulation dimensions are scalable: if we scale the bare dsDNA parameters by α = 10/4 . 8 , the 104-link simulation confined by a sphere of radius r also describes a polymer with width of αw = 10 nm , length of αL ≈ 6800 nm , and persistence length of αlp ≈ 110 nm , which is comparable to ∼100 kbp of 10 nm chromatin fiber confined in a sphere of radius αr , if we assume a density of 15 bp/nm for the fiber [51] . The range of confining radii and chain lengths corresponds to polymer volume fractions of up to 0 . 2% . Despite being limited to low volume fractions , the result that the eclipse effect is independent on volume fraction suggests that our model might be applicable to chromatinized enhancer-promoter looping segments that extend beyond the blob size . Our work indicates that we should consider three ranges for the physics of looping of chromatinized DNA . For short ranges which we studied previously [14 , 28] , looping is dominated by elastic energy . As shown in this work , there is an intermediate entropic looping range of up to ∼10 kbp of non-chromatinized DNA ( or ∼100 kbp of partially-chromatinized DNA ) . Recent studies indicate that there may be an additional long-range regime in which interactions between neighboring regions of the globular DNA must be taken into account [34 , 46] . | Biological regulation-at-a-distance , whereby a transcription factor ( TF ) is able to generate susbstantial regulatory effects on gene expression even though it may be bound a large distance away from its target ( 500 bp–1 Mbp ) , is only partially understood . Using a biophysical model and a computer simulation that take dsDNA and TF volumes into account , we identify a downregulatory mechanism which functions at large distances , whereby a TF bound within ∼ 150 bp from an activator decreases the probability of looping-based interaction between the activator and the distant core promoter . This “eclipse” mechanism provides insight into the question of how enhancer architecture dictates gene expression . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"invertebrates",
"gene",
"regulation",
"radii",
"geometry",
"animals",
"simulation",
"and",
"modeling",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"mathematics",
"materials",
"science",
"experimental",
"organism",
"systems",
"epigenetics",
"macromolecules",
"drosophila",
"chromatin",
"materials",
"by",
"structure",
"research",
"and",
"analysis",
"methods",
"polymers",
"polymer",
"chemistry",
"chromosome",
"biology",
"gene",
"expression",
"chemistry",
"insects",
"arthropoda",
"biochemistry",
"biochemical",
"simulations",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"organisms"
]
| 2017 | A Looping-Based Model for Quenching Repression |
Chronic viral infections lead to CD8+ T cell exhaustion , characterized by impaired cytokine secretion . Presence of the immune-regulatory cytokine IL-10 promotes chronicity of Lymphocytic Choriomeningitis Virus ( LCMV ) Clone 13 infection , while absence of IL-10/IL-10R signaling early during infection results in viral clearance and higher percentages and numbers of antiviral , cytokine producing T cells . IL-10 is produced by several cell types during LCMV infection but it is currently unclear which cellular sources are responsible for induction of viral chronicity . Here , we demonstrate that although dendritic cells produce IL-10 and overall IL-10 mRNA levels decrease significantly in absence of CD11c+ cells , absence of IL-10 produced by CD11c+ cells failed to improve the LCMV-specific T cell response and control of LCMV infection . Similarly , NK cell specific IL-10 deficiency had no positive impact on the LCMV-specific T cell response or viral control , even though high percentages of NK cells produced IL-10 at early time points after infection . Interestingly , we found markedly improved T cell responses and clearance of normally chronic LCMV Clone 13 infection when either myeloid cells or T cells lacked IL-10 production and mice depleted of monocytes/macrophages or CD4+ T cells exhibited reduced overall levels of IL-10 mRNA . These data suggest that the decision whether LCMV infection becomes chronic or can be cleared critically depends on early CD4+ T cell and monocyte/macrophage produced IL-10 .
The functional down regulation of antiviral T cells , also termed T cell exhaustion , is a major hurdle inhibiting the control or even clearance of chronic infections . T cell exhaustion is characterized by a gradual loss of cytokine producing antiviral CD8+ T cells [1] . The host-derived anti-inflammatory cytokine IL-10 is an important player in driving T cell exhaustion and viral chronicity in LCMV Clone 13 infected mice [2] , [3] , a commonly used murine model for chronic viral infections . Elevated IL-10 levels were also found to correlate with HIV replication in humans and transition to chronicity during HBV and HCV infection [4] , [5] , [6] . Since disruption of IL-10 receptor signaling enhances CD8+ T cell effector functions not only after LCMV infection [2] , [3] , but also in case of HIV- , HBV- or HCV-specific T cells [6] , [7] , [8] , [9] , interference with IL-10 signaling is currently proposed as a target for immune-based interventions during chronic viral infections . IL-10 is expressed during several persistent infections . It may on the one hand favor viral chronicity by suppressing the antiviral immune response , but on the other hand also protect the host from immunopathology [10] , [11] . IL-10 acts mainly by modulating the expression of proinflammatory cytokines and chemokines , by modulating the function of antigen presenting cells ( i . e . down-regulating for example the expression of MHCI , MHCII , B7-1 and B7-2 ) and by directly or indirectly suppressing proliferation , functional differentiation and effector activity of antiviral T cells [10] , [11] . Suppression of the antiviral immune response through IL-10 is a strategy actively exploited by herpes- and poxviruses which can encode viral IL-10 homologues to weaken the antiviral immune response [12] . Accordingly , rhesus macaques infected with rhesus CMV deficient for rhcmvIL-10 exhibit a T and B cell response of higher quality and quantity [13] . In addition , genetic polymorphisms in the IL-10 promoter are associated with decreased IL-10 production , leading to enhanced control of HCV , HBV , HIV and Epstein Barr Virus ( EBV ) [14] , [15] , [16] , [17] , [18] . Despite the advantages of more effective immune responses in absence of IL-10 , this may come at the expense of immunopathology [19] , [20] , [21] . It is currently unclear which cellular source ( s ) of IL-10 are decisive in promoting chronicity of high dose LCMV infection . Several cell populations have been reported to produce IL-10 after LCMV infection: LCMV specific CD4+ T cells secrete IL-10 on day 5 post infection [22] , IL-10 mRNA is upregulated in dendritic cells ( DCs ) , B cells , macrophages , CD4+ and CD8+ T cells on day 9 post infection [2] and purified CD4+ T cells and DCs produce IL-10 upon LCMV infection of splenocytes in vitro [3] . Furthermore , high percentages of DCs and macrophages , but also B cells , CD4+ and CD8+ T cells produce IL-10 nine and thirty days after Clone 13 infection when assessed in IL-10 reporter ( VertX ) mice [23] . Significant IL-10 production was found mainly in CD8− DCs and IL-10 serum protein levels were decreased in mice with selective IL-10 deficiency in CD11c+ cells [24] . In contrast , the IL-10 levels remained unaffected in B cell or T cell specific IL-10 knockouts on day 10 post infection [24] . Importantly and surprisingly , functional data evaluating the consequences of cell-specific IL-10 deficiency are lacking . It is conceivable that IL-10 is not primarily acting on a systemic level but in a microenvironment where IL-10 secreting cells and the immediate surrounding target cells are spatially and temporarily connected . Therefore , it is of great importance to identify the biologically relevant IL-10 producing cell populations responsible for impairment of LCMV-specific T cell immunity and consequently viral chronicity . We show here that even though DCs are a major source of IL-10 , specific IL-10 deficiency in DCs neither leads to functional improvement of the LCMV-specific T cell response nor does it prevent viral chronicity . Likewise , even though high percentages of NK cells produced IL-10 early after infection , NK cell specific IL-10 knockout mice were not able to clear LCMV and exhibited exhausted T cells . However , depletion of CD4+ T cells or monocytes/macrophages resulted in decreased overall amounts of IL-10 mRNA and selective IL-10 deficiency in T cells or myeloid cells prevented viral chronicity , identifying CD4+ T cells and monocytes/macrophages as the decisive IL-10 source responsible for suppressing T cell immunity and viral control during high dose LCMV infection .
To evaluate the timing when IL-10 determines about persistence or clearance of LCMV , the kinetics of viral titers were compared in C57BL/6 and Il-10−/− mice early after infection ( Figure 1A and Figure S1A–E ) . Interestingly , viral titers were comparable in all tested organs in both mouse strains until day 5/6 post infection . Afterwards , titers gradually decreased in Il-10−/− mice before the virus was cleared around day 10 post infection , whereas the titers remained high in C57BL/6 mice . Next , the question when IL-10R signaling is crucial to establish chronicity was addressed by starting IL-10R blockade at different time points after Clone 13 infection ( Figure 1B ) . Most mice that were treated with αIL-10R antibodies on day 0 and 1 succeeded in clearing LCMV Clone 13 , while the groups where αIL-10R treatment was started on day 2 or 3 developed severe immunopathology , which is often observed in situations of concomitant presence of high amount of antigen in presence of high numbers of functional T cells [25] , [26] . High viral titers were observed in the group where αIL-10R treatment was started on day 4 post infection , comparable to the untreated group . Similar observations were made in kidney , liver and lung ( data not shown ) . Thus , IL-10 acts during the first two days after infection to prevent immunopathology and to establish viral chronicity . To evaluate the dynamics of IL-10 production within the first days after LCMV infection , IL-10 mRNA levels were measured with daily intervals for the first 4 days . IL-10 mRNA was maximally induced during the first two days after infection and decreased afterwards in spleen ( Figure 1C ) , lung ( Figure S1G ) and liver ( Figure S1I ) . In the mesenteric lymph node IL-10 mRNA expression peaked on day 3 post infection ( Figure S1H ) . Quantification of IL-10 mRNA expression revealed that the highest amounts of IL-10 mRNA were expressed on day 2 post infection in the spleen ( Figure S1F ) . This prompted us to concentrate further analysis of IL-10 producing cell populations on spleen of day two post infection . FACS sorting of various cell populations and subsequent measurement of IL-10 mRNA levels by qPCR revealed that CD4+ T cells , CD8+ T cells , CD25+ CD4+ T cells , pDCs , CD8− DCs , neutrophils , macrophages and NKT cells upregulated IL-10 mRNA compared to naïve C57BL/6 splenocytes ( Figure 1D ) . However , only in purified NK cells the level of IL-10 mRNA increased to levels above those present in unsorted splenocytes . On day 5 post infection , IL-10 mRNA could no longer be detected in NK and NKT cells , whereas expression persisted in macrophages ( Figure 1E ) . Consistent with these data , analyses in TIGER IL-10-GFP reporter mice revealed around 10% of NK cells expressing GFP , while only between 0 . 5 and 2% of the NKT cells , DCs , CD4+ and CD8+ T cells were GFP+ on day 2 post infection ( Figure S2 ) . On day 5 post infection around 6% of the NK cells and 0 . 4–2% of the NKT cells , DCs , CD4+ and CD8+ T cells were GFP+ . In order to assess cell populations producing high amounts of IL-10 on a population level instead of on a per cell basis , the percentages of different cell populations of the splenocytes were determined on days 2 and 5 post infection . NK cells constituted 0 . 96±0 . 23% of the splenocytes on day 2 and 1 . 81±0 . 13% on day 5 post infection while NKT cells accounted for 0 . 11±0 . 05% on day 2 and 0 . 2±0 . 03% on day 5 . B cells were present at 28 . 7±4 . 3% on day 2 and at 56 . 6±2 . 18% on day 5 while CD4+ T cells constituted 7 . 92±1 . 47% on day 2 and 8 . 26±1 . 35% on day 5 and CD8+ T cells accounted for 4 . 97±1 . 26% on day 2 and 11 . 17±0 . 87% on day 5 . DCs represented 0 . 75±0 . 15% on day 2 and 0 . 92±0 . 18% on day 5 while F4/80+ macrophages formed 1 . 43±0 . 55% on day 2 and 3 . 42±0 . 17% on day 5 ( data not shown ) . To integrate the different abundance of IL-10 producing cell populations , C57BL/6 mice were depleted of specific cell populations and overall IL-10 mRNA levels in splenocytes were compared to undepleted control mice on day 2 post infection . Depletion of neutrophils , CD4+ or CD8+ T cells , CD25+ T cells and B cells did not lead to a decrease in the overall IL-10 mRNA levels ( Figure 2A and B ) . Unexpectedly , depletion of NK-like cells did not result in overall decreased IL-10 mRNA levels ( Figure 2B ) despite the high IL-10 mRNA expression level in purified cells ( Figure 1D ) . In contrast , depletion of CD11c+ cells ( DCs , alveolar macrophages and marginal zone macrophages; Figure 2C ) led to significantly decreased overall IL-10 mRNA expression . In addition , clodronate liposome meditated depletion of splenic ( red pulp , marginal zone and metallophilic [27] , but not CD68+ [28] ) macrophages led to decreased expression of IL-10 mRNA compared to undepleted controls ( Figure 2D ) . Thus , DCs and monocytes/macrophages account for a significant proportion of IL-10 mRNA on day 2 post infection . Analysis on day 5 post infection revealed that CD8+ T cell depleted mice expressed similar overall levels of IL-10 mRNA as undepleted controls ( Figure 2E ) , while CD4+ T cell depleted , clodronate liposome treated mice and DC specific IL-10 knockout mice ( Il-10fl/flxCD11c-Cre+ mice ) exhibited reduced levels ( Figure 2F–H ) . We chose to use Il-10fl/flxCD11c-Cre+ mice instead of DT-treated CD11c-DTR+ mice for the analysis of IL-10 mRNA levels on day 5 post infection due to the described toxicity of DT treatment for 5 days [29] and to avoid secondary consequences on viral control which might be caused due to inefficient T cell priming in absence of DCs [30] . Thus , DCs and monocytes/macrophages account for a first wave of IL-10 production , followed by a second wave of IL-10 production by DCs , CD4+ T cells and monocytes/macrophages on day 5 post infection . Even though different cell populations expressed IL-10 at different time points after LCMV infection , it is unclear if IL-10 secreted by any particular cell population is promoting viral chronicity and T cell exhaustion . To assess the effects of cell-type specific IL-10 deficiency on T cell effector functions and the resulting decision between chronicity versus clearance of the viral infection , we made use of conditional IL-10 knockout mice . These mice feature a floxed IL-10 gene and express the Cre recombinase under control of different cell-type specific promoters . While Il-10−/− mice were able to clear LCMV Clone 13 infection by day 12 , Il-10fl/flxM5-Cre+ mice , harboring a mast cell specific IL-10 deficiency , were unable to control the infection akin their Cre negative littermate controls and C57BL/6 mice ( Figure S3A–D ) . Single target amplification of the Il-10 locus of single FACS sorted mast cells from Il-10fl/flxM5-Cre+ mice previously confirmed that selectively the deleted version , but never the loxP-flanked product was amplified , while the opposite was true for non-mast cells and Cre-negative control cells [31] . We chose to assess T cell functionality by assessing TNF-α production as a surrogate for polyfunctionality of T cells as it is well established that virtually all TNF-α producing T cells are co-producers of IFN-γ [32] , [33] ( representative raw data are depicted in Figure S4 ) . We decided to use stimulation with gp33–41 ( gp33 ) as the frequencies of cytokine producing CD8+ T cells are higher than after stimulation with np396–404 and therefore not as prone to unspecific variances close to the detection limit ( as the percentage of TNF-α producing T cells is fairly low in exhausted T cells; [34] ) . TNF-α production by antiviral CD4+ T cells was measured after stimulation with the viral peptide gp61–80 ( p13 ) . In line with the fact that mast cell derived IL-10 was dispensable for the development of a chronic infection , the antiviral CD8+ and CD4+ T cells of Il-10fl/flxM5-Cre+ mice showed an exhausted phenotype in lung and spleen ( identified by reduced TNF-α production ) unlike Il-10−/− mice ( Figure S3E–J ) . In addition , the number of IFN-γ+ CD8+ T cells was not significantly different in Il-10fl/flxM5-Cre+ and Il-10fl/flxM5-Cre− mice in the spleen ( Figure S5A ) . Thus , mast cell specific IL-10 deficiency neither improves T cell mediated cytokine production nor virus control . Il-10fl/flxCD19-Cre+ mice were previously shown to delete IL-10 specifically in B cells and not or only insignificantly in splenic CD3+ T cells , peritoneal macrophages and tail biopsies [35] . Similar to mast cell specific IL-10 knockout mice , B cell specific IL-10 knockout mice displayed comparable virus titers ( Figure 3A–D ) and TNF-α production by CD4+ and CD8+ T cells in spleens and lungs to their Cre negative littermates and to C57BL/6 controls ( Figure 3E–J; FACS plots of a representative staining are depicted in Figure S4 ) . In addition , the number of IFN-γ+ CD8+ T cells was similar Il-10fl/flxCD19-Cre+ and Il-10fl/flxCD19-Cre− mice in the spleen ( Figure S5B; FACS plots of a representative staining are depicted in Figure S4 ) . NK cells produced high amounts of IL-10 on a per cell basis on day 2 post infection ( Figure 1D ) . To examine if NK cell derived IL-10 is decisive for Clone 13 infection to become chronic , Il-10fl/flxNcr1-Cre+ mice were derived from Il-10fl/fl and Ncr1-Cre+ mice . This mouse strain deleted the Il-10 gene in 63% of the NK cells ( Figure S6A ) . Il-10fl/flxNcr1-Cre+ mice were infected with Clone 13 and analyzed on day 12 post infection . In agreement with the observation that depletion of NK-like cells did not influence total IL-10 mRNA levels ( Figure 2B ) , Il-10fl/flxNcr1-Cre+ mice displayed high virus titers like their Cre negative littermates and like C57BL/6 mice ( Figure 4A–D ) . In addition , the antiviral CD8+ and CD4+ T cells showed both low percentages and total numbers of TNF-α+ cells and were as exhausted as in the Cre negative littermate controls and as in C57BL/6 mice ( Figure 4E–J ) . Furthermore , the number of IFN-γ+ CD8+ T cells was similar in Il-10fl/flxNcr1-Cre+ and Il-10fl/flxNcr1-Cre− mice in the spleen ( Figure S5C ) . Thus , the high amount of IL-10 mRNA on a per cell basis did not translate into a biologically relevant role of NK cell-derived IL-10 for the decision between chronicity and clearance of LCMV infection . CD8− DCs and pDCs expressed IL-10 mRNA on day 2 post infection ( Figure 1D ) . In accordance with the literature [24] , absence of CD11c+ cell-derived IL-10 resulted in strongly decreased overall IL-10 mRNA levels on day 2 and 5 post infection in the spleen ( Figure 2C and G ) . Il-10fl/flxCD11c-Cre+ mice proved to delete the Il-10 gene specifically in DCs ( Figure S6B ) . Unexpectedly , LCMV Clone 13 infection was not controlled in Il-10fl/flxCD11c-Cre+ mice ( Figure 5A–D ) and resulted in low percentages and total numbers of TNF-α producing antiviral CD8+ and CD4+ T cells comparable to littermate and C57BL/6 controls ( Figure 5E–H ) . Furthermore , the number of IFN-γ+ CD8+ T cells was similar in the spleens of Il-10fl/flxNcr1-Cre+ and Il-10fl/flxNcr1-Cre− mice ( Figure S5D ) . Thus , even though CD11c+ cells contribute substantially to the early overall IL-10 production during LCMV infection ( Figure 1D , 2C and G ) , CD11c+ cell derived IL-10 does not seem to play a role for the outcome of the infection and the quality of the antiviral T cell response . As neutrophils and macrophages produce IL-10 on day 2 and 5 post infection ( Figure 1D and E ) and as monocyte/macrophage depletion reduced the overall IL-10 mRNA levels ( Figure 2D and H ) , Il-10fl/flxLysM-Cre+ mice , in which selectively myeloid cells lack IL-10 production , were infected with Clone 13 . Il-10fl/flxLysM-Cre+ mice were previously shown to delete IL-10 specifically in neutrophils and macrophages and not or only insignificantly in dendritic cells , T cells and B cells [36] . Interestingly , 50% of the Il-10fl/flxLysM-Cre+ mice succeeded in significantly reducing LCMV Clone 13 titers like Il-10−/− mice and unlike the Cre negative littermates and C57BL/6 mice ( Figure 6A–D ) . In line with this , a higher percentage and total number of TNF-α producing T cells was observed in the spleens and lungs of Il-10fl/flxLysM-Cre+ mice which could control the infection compared to Cre negative littermate controls , the Il-10fl/flxLysM-Cre+ mice which could not control the infection and C57BL/6 mice ( Figure 6E–J ) . In addition , the number of IFN-γ+ CD8+ T cells was higher in the Il-10fl/flxLysM-Cre+ which could reduce the virus load than in Il-10fl/flxLysM-Cre− mice in the spleen ( Figure S5E ) . Thus , myeloid cell derived IL-10 contributes significantly to the biologically relevant effects of IL-10 without accounting for the full phenotype . As depletion of monocytes/macrophages decreased the levels of IL-10 mRNA on days 2 and 5 post infection , while αLy6G depletion failed to do so ( Figure 2A ) , it is likely that monocytes/macrophages are the relevant IL-10 source rather than neutrophils . Based on the finding that CD4+ and CD8+ T cells produced IL-10 after Clone 13 infection ( Figure 1D ) and that CD4+ T cell depletion decreased IL-10 mRNA levels on day 5 post infection ( Figure 2F ) , we investigated if T cell derived IL-10 complements macrophage derived IL-10 to exert it's biological effects . Il-10fl/flxCD4-Cre+ mice were shown previously to delete the Il-10 gene specifically in T cells [37] . Interestingly , almost all Il-10fl/flxCD4-Cre+ mice succeeded in clearing Clone 13 within 12 days comparable to Il-10−/− mice and opposed to Cre negative littermate controls and C57BL/6 mice ( Figure 7A–D ) . In line with these data , frequencies and numbers of TNF-α producing T cells were increased in mice with selective IL-10 deficiency in T cells compared to Cre negative littermate controls and C57BL/6 mice ( Figure 7E–J ) . Likewise , the number of IFN-γ+ CD8+ T cells was higher in the spleens of Il-10fl/flxCD4-Cre+ than in Il-10fl/flxCD4-Cre− mice ( Figure S5F ) . However , the percentage and total number of TNF-α producing antiviral CD4+ and CD8+ T cells was lower in Il-10fl/flxCD4-Cre+ mice compared to Il-10−/− mice , underlining the fact that macrophages and potentially also other cell populations contribute to the overall relevant IL-10 production . It should be noted that a slight increase in the viral inoculum led to increased incidence of viral chronicity in Il-10fl/flxCD4-Cre+ mice while most Il-10−/− mice were still able to clear the infection ( data not shown ) , suggesting that IL-10 derived from T cells and myeloid cells act together to exert its full biological activity . To further assess which CD4+ T cell subpopulation might be mainly responsible for the IL-10 production , TIGER IL-10 reporter mice were analyzed for GFP expression on day 10 post infection in TFH and TH1 cells using PSGL1 and Ly6C as surrogate markers ( with PSGL1+ Ly6C+ CD4+ T cells representing TH1 cells and PSGL1− Ly6C− cells representing TFH cells; Figure S7A–F ) . GFP+ cells could be detected in both subpopulations . Likewise , GFP expression could be detected in a subpopulation of Neuropilin+ natural Tregs . However , the biggest increase compared to the naïve situation could be detected in Neuropilin− induced Tregs ( Figure S7G–I ) . The finding that induced Tregs as opposed to TH1 cells are probably responsible for the majority of the CD4+ T cell derived IL-10 is supported by the finding that only very few IFN-γ+ CD4+ T cells co-produced IL-10 ( Figure S7J–K ) .
Several cell populations are capable of producing IL-10 [10] , [38] . DCs , B cells , macrophages , CD4+ and CD8+ T cells were previously shown to produce IL-10 after LCMV infection [2] , [3] , [22] , [24] . However , data on functional consequences of IL-10 deficiency in defined cell types were lacking so far . In agreement with previous studies [2] , [3] , [22] , [24] we demonstrate here that DCs , B cells , macrophages , CD4+ and CD8+ T cells , but also neutrophils , NK cells and NKT cells produce IL-10 in the early phase of high dose LCMV Clone 13 infection . These diverse IL-10 producing cell types potentially all play important and distinct roles in suppressing the host's immune response . However , we clearly show that myeloid and T cell derived IL-10 is critical in promoting viral chronicity and T cell exhaustion , while IL-10 derived from other sources , including DCs , only contributes marginally - if at all - to the suppression of the antiviral immune response . Since depletion of monocytes/macrophages but not neutrophils led to markedly reduced IL-10 mRNA levels , we conclude that monocyte/macrophage derived IL-10 accounts for the reduced virus titer and increased cytokine production by antiviral T cells occurring in the myeloid cell specific IL-10 knockout mice . IL-10 drives viral persistence in a number of chronic infections . Elevated IL-10 levels were found to correlate with HIV replication in humans and chronicity during HCV and HBV infection [4] , [5] , [6] . Likewise , disruption of IL-10 receptor signaling enhances CD8+ effector functions of T cells isolated from HIV , HCV and HBV infected individuals [6] , [7] , [8] , [9] like after LCMV infection [2] , [3] . The dominant IL-10 producing cell population varies with different infections , likely reflecting differences in pathogen-specific responses and tissue-specific immune regulation . Usually , IL-10 is produced by multiple cell populations [10] . During HIV infection , IL-10 production by T cells , B cells , NK cells and monocytes was reported [4] with monocytes/macrophages being the subset producing highest levels of IL-10 in the peripheral blood [39] , [40] , [41] . In contrast , B cell derived IL-10 is important during HBV infection [6] . During murine cytomegalovirus ( MCMV ) infection , IL-10 can be produced by CD4+ T cells , B cells , inflammatory macrophages , dendritic cells and NK cells [19] , [35] , [42] , [43] . The biologically relevant in vivo sources of IL-10 in MCMV infection were identified as CD11c+ cells and to some extent macrophages . IL-10 from these cellular sources led to the suppression of NK/DC cross-talk and consequently to curtailed CD4+ T cell priming , thereby promoting prolonged lytic MCMV [19] . In addition , B cell produced IL-10 was shown to suppress the CD8+ T cell response in the spleen during MCMV infection [35] . During Leishmania major infection anti-IL-10R treatment leads to sterile cure of infection with complete elimination of the parasite . In this infection model , CD4+ and CD8+ T cell derived IL-10 plays an essential role for the establishment of pathogen persistence [44] . Thus , the different importance of certain cell populations serving as IL-10 source might depend on the specific tissue , the cell population affected and the time after infection . It was recently shown that myeloid suppressor cells inhibit CD8+ T cell proliferation during chronic LCMV infection [45] . However , the fact that in vitro co-cultivated myeloid suppressor cells could still inhibit CD8+ T cell proliferation in presence of an anti-IL-10R antibody [45] renders it rather unlikely that myeloid suppressor cells are a relevant source of IL-10 , assuming that IL-10 acts directly on the CD8+ T cells . CD4+ T cell depletion led to a decrease in the overall IL-10 mRNA levels on day 5 post infection with LCMV while depletion of CD8+ T cells had no effect ( Figure 2 ) . Therefore , it is likely that CD4+ T cell derived IL-10 is more important than CD8+ T cell derived IL-10 . IL-10 was initially identified as TH2 cytokine [46] , but can also be produced by TH1 , TH17 , TFH , natural and induced Tregs [47] . As αCD25 depletion did not decrease IL-10 mRNA levels on day 2 post infection , natural Tregs do not seem to be relevant IL-10 producers during LCMV infection . These data were supported by the finding that IL-10+ Neuropilin+ natural Tregs did not accumulate upon infection . In contrast to natural Tregs , IL-10-producing Neuropilin− induced Tregs accumulated by day 10 after LCMV infection and likely constitute a subpopulation of CD4+ T cells that is responsible to promote chronicity of the infection , at least on day 10 post infection when LCMV chronicity is already established . Accumulation of induced Tregs during established chronic LCMV infection is in agreement with previous studies [48] , [49] . Interestingly , the depletion of monocytes/macrophages decreased IL-10 mRNA levels on days 2 and 5 post infection while CD4+ T cell depletion led to reduced concentrations of IL-10 mRNA only on day 5 . This implies a cooperation of macrophages and CD4+ T cells in a way that the macrophages provide a first wave of IL-10 until , in a second wave , different subtypes of CD4+ T cells ( TH1 , TFH , natural Tregs and induced Tregs with induced Tregs exhibiting the highest percentage of IL-10 producing cells ) start to significantly contribute to the overall IL-10 production . Together with the fact that all analyzed CD4+ T cell subpopulations contribute to IL-10 production , it seems as if multiple cell types share the task of IL-10 secretion in a time-dependent manner . It is possible that IL-10 production by monocytes/macrophages is preceded by a first wave of IL-10 production by NK cells . However , absence of NK cell derived IL-10 did not promote control of high dose LCMV infection , suggesting that IL-10 production by monocytes/macrophages plays a more dominant role . It should be noted , however , that the NK cell-specific Il-10 deletion efficacy was only 63% ( Figure S6 and [50] ) , but in combination with the observation that NK cell depletion did not lead to an overall reduction in IL-10 mRNA on day 2 post infection it is rather unlikely that NK cell produced IL-10 plays a decisive role in promoting LCMV chronicity . The observation that dendritic cells produce IL-10 and that the IL-10 mRNA levels decreased dramatically in absence of CD11c+ cells ( Figure 2C ) is consistent with the finding that Il-10fl/flxCD11c-Cre mice expressed considerably less IL-10 ( Figure 2G and [24] ) . Importantly and to some extend surprisingly , we extend this observation by showing that IL-10 produced by CD11c+ cells lacks functional consequences with respect to antiviral T cell function or the establishment of viral chronicity ( Figure 5 ) . It is likely that the precise anatomical localization and kinetics of the IL-10 secreting cells in relation to the responding cell is of considerable importance . This would be one possible explanation for the observation that IL-10 produced by CD11c+ cells does not influence the outcome of the infection even though IL-10 promoter activity could be detected in DCs of IL-10 reporter mice and even though depletion of CD11c+ cells dramatically decreased the overall amount of IL-10 mRNA . On the other hand , post-transcriptional regulation of IL-10 may vary in different cell types , such that quantitative differences evidenced by mRNA analyses might not translate to the protein level [51] , [52] . One potential caveat of IL-10 mRNA analysis after in vivo depletion of certain cell populations for several days into LCMV infection is that the absence of specific cell types may alter virus titers which might in turn affect the IL-10 production by other cells types . In addition , depletion of CD11c+ cells by DT administration in CD11c-DTR+ mice was shown to cause neutrophilia [53] . So , if neutrophils contributed to IL-10 production significantly , their increased abundance might have masked the decrease in IL-10 mRNA levels in diphtheria toxin treated CD11c-DTR mice . Our data do not permit a conclusion on whether the more potent antiviral T cell responses observed in absence of T cell or myeloid-derived IL-10 are a cause or a consequence of the concomitantly reduced virus titers , but they underscore an important role of IL-10 in shaping the mutual relationship between prolonged high antigen load and decreased T cell function [26] . It has to be noted that only a defined LCMV inoculum is controlled in the T specific IL-10 knockout mice , while a wider range of inocula is controlled in full IL-10 deficient mice . This is likely one reason why the full IL-10 knockout mice show more robust T cell responses compared to the cell-specific knockout mice , yet these small increases might be sufficient under these experimental settings to allow for virus control . At present it is unclear on which cell types the T cell or myeloid cell derived IL-10 acts to promote viral chronicity but it is likely that multiple immune cell subsets will be targeted . It was recently shown that splenic DCs treated in vitro with type I IFN produce IL-10 [24] and that anti-IFNR1 blockade led to reduced numbers of IL-10+ PD-L1+ immunoregulatory DCs which in turn stimulated elevated numbers of IFN-γ+ CD4+ T cells and a decrease in viral titers [54] . It is likely that type I IFN also directly induces IL-10 production in macrophages at the early time points after infection . However , as the amounts of IFN decline at later time points [55] , this might indicate that CD4+ T cell derived IL-10 is induced by other mechanisms . In this context we have measured type I IFNs in the serum of C57BL/6 and Il-10−/− mice and we have observed increased concentrations of type I IFN in Il-10−/− mice on day 3 post infection ( data not shown ) . As Type I IFNs are an essential signal 3 in LCMV to drive T cell expansion [56] , [57] and effector cell differentiation [58] , it is possible that increased peak levels of type I IFNs might contribute to more effective T cell responses which in turn would accelerate virus control . It is also possible that increased levels of type I IFNs in Il-10−/− mice might have a direct effect on LCMV control , but this is rather unlikely as we do not see any changes in viral load in the first 5 days post infection . As we determined that IL-10 has to act within the first 2 days it is very likely that initially innate immune cells or non-hematopoietic cells are targeted by IL-10 . Indeed , we have preliminary data that IL-10 does not directly act on CD4+ or CD8+ T cells and on DCs ( data nor shown ) . The innate target cells could be macrophages themselves ( being infected by LCMV ) which are modulated by IL-10 such that they would better sustain viral replication , potentially by inhibition of NO production as it was shown for parasite infection [59] or by inhibition of TNF-α and IFN-γ production [60] . We believe that a direct effect of IL-10 on virus production is rather unlikely as we have seen in our in vivo kinetics of viral loads that differences between C57BL/6 and Il-10−/− mice only become apparent after day 6 of infection ( Fig . 1 ) . Alternatively , NK cells might be possible IL-10 target cells as they are mainly situated in the red pulp of the spleen where they might be targeted by macrophage-secreted IL-10 . In MCMV infection it has been shown that NK cells are over-activated in absence of IL-10 and are undergoing increased levels of activation-induced cell death [43] . As it has been shown in the LCMV system that activated NK cells can reduce the number of activated virus-specific CD4+ T cells [61] , it is conceivable that potentially reduced numbers of NK cells in the absence of IL-10 would translate into better preserved virus-specific CD4+ ( and CD8+ ) T cell responses . With respect to the secondary requirement of CD4+ T cell derived IL-10 driving viral chronicity we hypothesize that by day 4–5 post infection , activated CD4+ T cells will not only be in secondary lymphoid organs but also in ( infected ) peripheral tissues where their IL-10 secretion might affect local immune control . Based on our observation that IL-10 does not directly act on T cells and based on additional observations that the per cell killing efficiency of CD8+ T cells ( when using optimally peptide-loaded APCs ) is not altered in presence or absence of IL-10 ( data not shown ) , we postulate that either the increased frequencies of T cells in peripheral tissues lead to better virus control or that in addition presence of ( CD4+ T cell secreted ) IL-10 would impact on the antigen presentation on infected cells . IL-10 is known to down-modulate MHC expression and it could thereby reduce sensitivity towards LCMV infected cells . In summary , our data clearly indicate a non-redundant role of T cell and myeloid cell derived IL-10 in facilitating viral chronicity and T cell exhaustion after LCMV Clone 13 infection as opposed to DCs , which were so far proclaimed to exert this function .
LCMV Clone 13 was propagated and titrated as described [62] . The viral peptides gp33–41 ( gp33; KAVYNFATM ) , np396–404 ( np396; FQPQNGQFI ) and gp61–80 ( p13; LNGPDIYKGVYQFKSVEFD ) were purchased from NeoMPS ( Strasbourg , France ) . Animal experiments were performed according to the guidelines of the animal experimentation law ( SR 455 . 163; TVV ) of the Swiss Federal Government . The protocol was approved by the Cantonal Veterinary Office ( animal experimentation number 146/2008 and 127/2011 ) . Ncr1-Cre mice [50] were crossed to Il-10fl/fl mice [37] yielding Il-10fl/flxNcr1-Cre mice . C57BL/6 ( Janvier Elevage , Le Genest Staint Isle , France ) , Il-10fl/flxCD4-Cre [37] , Il-10fl/flxLysM-Cre [36] , Il-10fl/flxCD11c-Cre [24] , Il-10fl/flxCD19-Cre [35] , Il-10fl/flxM5-Cre [31] , Il-10−/− [63] , TIGER [64] and CD11c-DTR/GFP [29] mice were kept under specific pathogen-free conditions and were intravenously ( i . v . ) infected with LCMV Clone 13 . Each virus stock was titrated in vivo and injected at doses that led to a chronic infection in C57BL/6 mice and to reduced virus titers by day 10 to 12 in most of the Il-10−/− mice ( between 3×105 and 106 ffu per mouse ) . Antibodies used for flow cytometry were purchased from BD Pharmingen ( Allschwil , Switzerland ) , Biolegend ( LucernaChem , Luzern , Switzerland ) and R&D Systems ( Abingdon , England ) . For the depletion of neutrophils , mice were i . p . injected with 500 µg αLy6G ( clone 1A8 , BioXCell , West Lebanon , NH , USA ) on day -2 and one hour before infection ( 99 . 8% reduction in the percentage of Ly6G+ cells ) . For the depletion of CD4+ T cells , mice were i . p . injected with 200 µg αCD4 ( YTS 191 . 1 ) on day -2 and one hour before infection ( depletion efficacy of 97% while the percentage of CD8+ T cells increased slightly ) . For the depletion of B cells , mice were i . p . injected with 250 µg αCD20 ( 18B12; IgG2a; [65] ) on day -2 and one hour before infection ( depletion efficacy of 98% while the percentage of CD8+ T cells and CD4+ T cells increased compared to mice treated with an isotpye control ) . For the depletion of CD8+ T cells , mice were i . p . injected with 200 µg αCD8 ( YTS 169 . 4 ) on day -2 and one hour before infection ( depletion efficacy of 99% while the CD4+ T cell population stayed unaffected ) . Depletion of NK-like cells was performed by i . p . administration of 300 µg αNK1 . 1 ( PK136 , BioXCell ) on days -3 and -1 ( depletion efficacy of 96% while the percentage of CD3+ cells was not affected ) . For the depletion of regulatory CD4+ T cells mice were i . p . injected with 200 µg of αCD25 ( PC-61 . 5 . 3 , BioXCell ) 3 and 1 day before infection ( depletion efficacy of 70% ) . For depletion of CD11c+ cells , CD11c-DTR+ and CD11c-DTR− mice were i . p . treated with 200 ng diphtheria toxin ( Sigma-Aldrich , Buchs , Switzerland ) on days -2 , -1 and 0 ( depletion efficacy of 95% while the percentage of B cells was not affected ) . For depletion of monocytes/macrophages mice were i . p . treated with liposomes containing 1 mg clodronate [27] or equal amounts of empty control liposomes on days -7 , -5 , -2 and 0 ( depletion efficacy of 96% while the percentage of B cells and T cells was not affected ) . The depleted mice were analyzed on day 2 post infection . When analyzed on day 5 post infection , the depleting agent was continuously administered every second to third day . Splenocytes and mononuclear cells from lungs were prepared , stimulated and stained as previously described [26] . Phorbol 12-myristate 13-acetate ( PMA ) and Ionomycin were purchased from Sigma and used at 50 ng/ml and 500 ng/ml , respectively . Multiparameter flow cytometric analysis was performed using a FACS LSRII flow cytometer ( BD Biosciences , Allschwil , Switzerland ) with FACSDiva software . Analysis was performed using FlowJo software ( Tree Star , San Carlos , CA ) . Sorting was performed using a FACS Aria flow cytometer ( BD Biosciences , Allschwil , Switzerland ) . The sorted cell populations were defined as follows: CD4+ T cells: CD4+ CD25−; CD25+ CD4+ T cells: CD4+ CD25+; CD8+ T cells: CD8+ CD4−; B cells: CD4− CD19+; pDCs: B220+ PDCA-1+ Ly6C+; DCs: CD11c+ I-A/I-E+; CD8− DCs: CD11c+ I-A/I-E+ CD8−; CD8+ DCs: CD11c+ I-A/I-E+ CD8+; inflammatory monocytes: CD11b+ Ly6C+ Ly6G−; neutrophils: CD11b+ Ly6C− Ly6G+; macrophages: CD11bint F4/80+; NKT cells: CD3+ NK1 . 1+; NK cells: CD3− NK1 . 1+ . NK cells were isolated from splenocytes on day 7 after i . p . infection with 5×104 pfu MCMV Smith by magnetic negative selection of CD3+ and CD19+ cells and subsequent FACS sorting of CD3− NK1 . 1+ cells . DCs were isolated from splenocytes on day 7 after s . c . injection with 106 B16-Flt3L cells by magnetic negative selection of CD3+ and CD19+ cells and subsequent FACS sorting of CD11c+ MHCII+ cells . Southern Blot was performed as described [37] . Total RNA was isolated with Trizol ( Invitrogen , Basel , Switzerland ) according to the manufacturer's instructions . For isolation of RNA from whole organs , the organ was homogenized with a stainless steel bead for 1 . 5 min at 25 Hz in a Qiagen Tissue Lyser . 1 µg of RNA was reverse transcribed with M-MLV Reverse Transcriptase RNase , H Minus ( Promega , Dübendorf , Switzerland ) . Real-time PCR was performed using a Rotorgene 3000 instrument ( Corbett Research , Eight Miles Plains , Australia ) to measure SYBR green ( Sensi Mix , R&D Systems ) incorporation . The following primer sets were used: IL-10: 5′-ATGCTGCCTGCTCTTACTGACTG-3′ and 5′-CCCAAGTAACCCTTAAATCCTGC-3′ [66]; β-actin: 5′-CCCTGAAGTACCCCATTGAAC-3′ and 5′-CTTTTCACGGTTGGCCTTAG-3′ . The amount of IL-10 mRNA was normalized to β-actin RNA levels for each sample and was expressed relative to the levels of uninfected control mice . The two-tailed unpaired t test was applied for statistical analysis when the distribution was Gaussian . Otherwise , the Mann-Whitney test was performed . * p<0 . 05; ** p<0 . 01; ***p<0 . 001; **** p<0 . 0001; n . s . not significant . | Chronic viral infections like Hepatitis B and C Virus ( HBV and HCV ) and Human Immunodeficiency Virus ( HIV ) in humans affect more than 500 million people worldwide . While a robust T cell response is a hallmark of many acute infections one hurdle inhibiting the clearance of chronic viral infections is that the immune-suppressive cytokine IL-10 modulates the virus-host balance towards induction of T cell dysfunction . IL-10 is produced by several cell types during chronic Lymphocytic Choriomeningitis Virus ( LCMV ) infection but it is currently unclear which cellular sources are responsible to promote viral chronicity . Here , we demonstrate that T cell responses improved markedly , and that normally chronic LCMV Clone 13 infection could be cleared when either myeloid cells or T cells lacked IL-10 production . Furthermore , mice depleted of monocytes/macrophages or CD4+ T cells exhibited reduced overall levels of IL-10 mRNA . These data suggest that the decision whether LCMV infection becomes chronic or can be cleared critically depends on CD4+ T cell and monocyte/macrophage produced IL-10 early during the establishment of viral chronicity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2013 | Macrophage and T Cell Produced IL-10 Promotes Viral Chronicity |
The SV40 small t antigen ( ST ) is a potent oncoprotein that perturbs the function of protein phosphatase 2A ( PP2A ) . ST directly interacts with the PP2A scaffolding A subunit and alters PP2A activity by displacing regulatory B subunits from the A subunit . We have determined the crystal structure of full-length ST in complex with PP2A A subunit at 3 . 1 Å resolution . ST consists of an N-terminal J domain and a C-terminal unique domain that contains two zinc-binding motifs . Both the J domain and second zinc-binding motif interact with the intra-HEAT-repeat loops of HEAT repeats 3–7 of the A subunit , which overlaps with the binding site of the PP2A B56 subunit . Intriguingly , the first zinc-binding motif is in a position that may allow it to directly interact with and inhibit the phosphatase activity of the PP2A catalytic C subunit . These observations provide a structural basis for understanding the oncogenic functions of ST .
Simian virus 40 ( SV40 ) is a DNA tumor virus in the polyomavirus family . SV40 may play a role in a subset of human cancers , and the study of transformation induced by SV40 has led to many insights into the pathways involved in spontaneously arising cancers [1] . The Early Region of SV40 is essential for transformation and encodes two oncoproteins , the large T antigen ( LT ) and small t antigen ( ST ) , through alternative splicing . LT binds to a number of host proteins including the retinoblastoma and p53 tumor suppressors . ST , which shares its N terminus with LT but has a unique C-terminal end , is also a potent oncoprotein that plays a critical role in the transformation of several human cell types [2 , 3] . For example , the cointroduction of LT , ST , the telomerase catalytic subunit hTERT ( human telomerase reverse transcriptase ) , and an oncogenic allele of H-RAS imparts a tumorigenic phenotype to a wide range of primary human cells [4–6] . The tumorigenic activity of ST is strictly dependent on its interaction with protein phosphatase 2A ( PP2A ) , since mutant versions of ST that are unable to bind PP2A fail to induce tumorigenic activity [5] . PP2A is a large family of heterotrimeric enzymes that accounts for the majority of total Ser/Thr phosphatase activity in most tissues and cells [7–10] . Although a dimer comprised of a ~65 kDa scaffolding A subunit and a ~36 kDa catalytic C subunit constitutes the core enzymatic activity of PP2A , the binding of a third regulatory B subunit to the AC core enzyme regulates PP2A activity , cellular localization , and substrate specificity . PP2A can exist in cells as either the AC core complex or an ABC heterotrimeric holoenzyme that is the dominant form in most cells . The scaffold A subunit is composed of 15 HEAT ( huntingtin , elongation factor 3 , A subunit of protein phosphatase 2A , and target of rapamycin ) repeats [11] , while the C subunit contains a catalytic domain that shares sequence homology with other Ser/Thr phosphatases such as protein phosphatase 1 and protein phosphatase 2B ( calcineurin ) . In humans there are at least 18 B subunits that can be classified into B ( B55 ) , B′ ( B56 ) , B′′ , and B′′′ families based on sequence similarity . PP2A regulates a wide array of cellular processes . In particular , several lines of evidence implicate PP2A as a tumor suppressor gene . Specifically , PP2A inhibitors such as microcystin and okadaic acid are potent tumor promoters and the ST antigen acts as a potent oncogene [12–16] . Consistent with its role as a tumor suppressor , low frequency mutations of the Aα and Aβ subunits are found in breast , lung , and colorectal cancers [12 , 17–20] . Mutations in PP2A Aα result in functional haploinsufficiency that depletes complexes containing B56γ , thereby leading to cell transformation [16 , 21] . In contrast , mutations in Aβ disrupt the interaction of Aβ with the small GTPase RalA , leading to constitutive RalA phosphorylation and activity [22] . ST interacts with the PP2A AC core dimer through the direct interaction between ST and the PP2A scaffolding A subunit [23 , 24] . Prior work has suggested that ST associates with the PP2A C subunit through interactions with the A subunit . While the C subunit interacts with HEAT repeats 11–15 of A subunit , ST interacts with HEAT repeats 3–7 , which is also the binding site for PP2A B subunits [25 , 26] . The ability of ST to displace multiple B subunits from the A subunit has been demonstrated both in vitro [27 , 28] and in vivo [16 , 29] . The displacement of B subunits by ST inhibits PP2A activity towards multiple substrates , but increases phosphatase activity towards histone H1 [22 , 30] . Therefore , ST can be considered as a viral B subunit that changes the biochemical properties of PP2A . The 174 amino acid ST shares its N-terminal 78 residues with the N-terminal sequence of LT , including a region that exhibits J domain ( also termed the DnaJ domain ) function [3 , 31 , 32] . The unique C-terminal 96 residues of ST antigen interact with the PP2A A subunit . The C-terminal unique domain is rich in Cys residues , six of which are organized into two CxCxxC clusters and bind two zinc ions , providing structural stability to ST [33 , 34] . The N-terminal J domain is not required for the interaction of ST with the A subunit of PP2A but may contribute to the high affinity binding , since the removal of the first 51 residues from ST caused a 140-fold reduction in inhibition towards the phosphatase activity of PP2A AC core complex [35] . The N-terminal J domain shares sequence homology with the DnaJ family of molecular co-chaperones , which promote the ATPase and chaperone activities of heat shock protein 70 ( Hsp70 ) , an important chaperone in the cell [36 , 37] . Hsp70 binding region has been mapped to the surface formed by J domain helices 2–3 [38 , 39] . The DnaJ-like domain in ST and the DnaJ domain in E . coli are interchangable without a loss in co-chaperone activity [40] . While the structure of the J domain can be predicted from prior structural determinations of LT , the overall structure of ST remains to be unraveled and it is unclear how ST may interact with PP2A and regulate PP2A activities . We have determined the crystal structure of full-length SV40 ST in complex with the full-length Aα subunit of PP2A . This structure reveals two novel zinc-binding motifs formed by the unique C-terminal domain , the structural linkage of the J and unique domain of ST , and the interaction site of ST with the structural A subunit . Together with our biochemical data , we provide a structural basis for understanding the tumorigenic activity of ST protein .
The protein complex containing full-length SV40 ST and full length murine PP2A Aα subunit ( A-ST complex ) were co-expressed in E . coli and purified to homogeneity . Crystal structure of the complex was determined by a combination of molecular replacement , using the PP2A A subunit structure as the searching model , and single-wavelength anomalous dispersion of intrinsic zinc atoms in ST , and was refined at 3 . 1 Å resolution ( Table 1 ) . Four complexes were found in each asymmetric unit . In each complex , the scaffolding A subunit contains 15 HEAT repeats that forms a horseshoe shape . The four A-ST complexes in the asymmetric unit have essentially the same structure , except HEAT repeats 11–15 that show substantial conformational variation ( see below ) . ST contains an N-terminal J domain and a C-terminal unique domain . These two domains sit on the concave and convex sides of the ridge of the A subunit horseshoe structure , respectively , by interacting with intra-repeat loops of the Aα subunit HEAT repeats 3–7 ( Figure 1 ) , which is also the binding site for B56γ1 in the A-B56γ1-C trimeric PP2A holoenzyme structure [25 , 26] . SV40 ST exhibits an all α-helix structure with two zinc-binding sites in the unique domain ( Figure 2A ) . The J domain contains three helices and has a structure similar to the previously solved crystal structure of the J domain of SV40 LT and a NMR structure of polyomavirus DnaJ-like domain [41 , 42] , with Cα root mean square deviations ( RMSDs ) of 1 . 84 and 1 . 83 Å , respectively . The DnaJ domain in the E . coli DnaJ protein forms a complex with Hsp70 in the flexible L3 region [38 , 39] . Interestingly , the expected Hsp70 binding region of the J domain is not involved in the interaction with PP2A A subunit and appears widely accessible to Hsp70 . In contrast , the Hsp70 binding region of J domain in LT is involved in the binding of retinoblastoma protein [41] . The linker between helices 2 and 3 , which has been implicated in Hsp70 interaction , as well as the linker between J and unique domains ( residues 72–86 ) , is not visible in any of the four complexes in the asymmetric unit , suggesting structural flexibility of these two linkers in the A-ST complex . The ST unique domain is composed of four helices . Three of these helices are directly involved in the interaction with two zinc ions . The first zinc-binding site is positioned between two roughly parallel helices ( α4 and α5 ) and the L5 loop , coordinated by four conserved cysteine residues ( Cys103 , Cys111 , Cys113 , and Cys116 ) . Cys103 stems from the C-terminal end of α4 , while Cys113 and Cys116 are positioned in the N-terminal end of α5 ( Figure 2A ) . The second zinc is located between two roughly perpendicular helices ( α5 and α6 ) and coordinated by three conserved cysteines in the second cysteine cluster ( Cys138 , Cys140 , and Cys143 ) , with Cys140 and Cys143 from the N-terminal end of α6 and the conserved His122 from the middle of α5 ( Figure 2A ) . Although these two zinc ions are located in discrete positions , both of them coordinate with α5 . These two zinc-binding motifs are completely distinct from the previously proposed GAL4-type zinc cluster for ST and are also different from classical zinc finger structures [33 , 34 , 43] . The J and unique domains have an interface that is mostly hydrophobic ( Figure 2B ) . These two domains of SV40 ST interact with each other by several hydrogen-bonding and hydrophobic interactions . The indole rings of Trp135 and Trp147 make hydrogen bonds with the backbone carbonyl group of Gly25 and Trp24 , respectively . The hydroxyl group of Tyr139 forms a hydrogen bond with the imidazole ring of His70 . Pro28 and Ile163 make a hydrophobic interaction between two domains ( Figure 2B ) . It should be noted that both the N and C termini of ST are located at the interface of the J and unique domains . In particular , the N terminus is located in the joint point of J domain and the A subunit , and directly interacts with the A subunit . The interface and relative orientation of these two domains are essentially identical in all four A-ST complexes in the asymmetric unit . The Cα RMSDs of ST among these four complexes are between 0 . 95 Å and 1 . 03 Å . Therefore , the J and unique domains of ST are structurally coupled and ST consists of one globular fold . Since residues in the A-ST interface are highly conserved among ST proteins ( Figure 2C ) , this domain organization should be conserved among ST proteins in the polyomavirus family . The interactions between the Aα subunit and SV40 ST are formed through the intra-loop regions of repeats 3 to 7 in the Aα subunit ( Figures 1 and 3 ) . Detailed interactions are summarized in Figure 3 . ST uses both its J domain and the second zinc-binding motif ( in particular the L6 loop ) for interacting with PP2A Aα subunit , while the first zinc finger is not directly involved in interactions with PP2A Aα subunit , but may be involved in interactions with PP2A catalytic C subunit ( see below ) . One notable feature of this interaction is that five of eight residues in the ST unique domain ( Figures 2C and 3 ) in the A-ST interface interact with the Aα subunit through the backbone of those residues . The backbone conformations of those ST residues ( Leu133 , Met146 , Trp147 , Phe148 , and Gly149 ) are largely determined by the second zinc-binding site , demonstrating the importance of the second zinc-binding site in the Aα subunit interaction . In contrast , all of the Aα subunit residues interact with ST through their side chains , and some of those residues were identified by previous A subunit mutagenesis studies [44] . When the A-ST complex structure is superimposed with the previously reported A-B56-C PP2A holoenzyme structure [25 , 26] , it is obvious that ST and the B56 subunit interact with the same region of PP2A A subunit and have very similar “footprints” on the scaffold A subunit ( Figure 4A ) . The majority of Aα residues involved in the A-ST interaction—including Glu100 , Trp140 , Phe141 , Asp177 , Pro179 , and Met180—are also involved in the A-B56 interaction ( Figure 4B ) . The side chain conformations of those Aα residues are quite similar in both structures , except Trp140 . The position of the indole ring of Trp140 is flipped between these two complex structures , mostly due to the intercalation of Pro132 of ST between Trp140 and Phe141 by forming a hydrophobic core ( Figures 3 and 4B ) . The largely overlapping Aα binding sites of ST and B56 explains how ST competes with B56 for the binding of the Aα subunit [25 , 26] . In the A-B56-C PP2A holoenzyme structure , both B56 and C subunits sit on the same side ( intra-repeat loop side ) of the horseshoe shape formed by the scaffolding A subunit [25 , 26] . When HEAT repeats 2–10 in the A-ST complex are superimposed with corresponding regions of the A-B56-C complex , the first zinc-binding motif , in particular helix α4 and its flanking region , is in close proximity to the C subunit active site ( Figure 5 ) . The closest atoms between ST and C subunit in this superposition are within van der Waals distances . It should be noted that , when the A-ST complex structure is superimposed with the PP2A AC core complex , the distance between the ST first zinc-binding motif and C subunit appears further apart than in the A-B56-C holoenzyme . However , since the PP2A A subunit has substantial structural flexibility ( see below ) , it is plausible that ST may interact directly with the C subunit near its catalytic active site , just like the B56 subunit that interacts with both A and C subunits . The A-ST interface and the structure of HEAT repeats 2–10 of the A subunit are quite rigid , since these structures can be well superimposed in all four A-ST complexes in the asymmetric unit . In contrast , HEAT repeats 11–15 apparently have highly variable structures , and the first HEAT repeat is not well folded in some cases ( Figure 6A and 6B ) . While the RMSDs of all 15 HEAT repeats among four A-ST complexes in the asymmetric unit range from 1 . 29 Å to 3 . 02 Å , the RMSD of HEAT repeats 2–10 are on average 1 . 09 Å . The conformational variation of the Aα subunit results mostly from conformational flexibility in HEAT repeats 10–13 , since HEAT repeats 13–15 among the four complexes are easily superimposed , with an average RMSD of 1 . 15 Å ( Figure 6C ) . Structural comparison of the A-ST crystal structure with previously reported crystal structures of A , AC complex , and A-B56-C complexes [25 , 26 , 45 , 46] also support the conclusion that HEAT repeats 2–10 and HEAT repeats 13–15 form two relatively rigid blocks . However , there is substantial structural flexibility between these two structural blocks , due to the result of accumulative conformational changes in HEAT repeats 10–13 ( Figures S1 and S2 ) . To determine whether the amino acid residues observed in our structure to be interaction points between ST and Aα , we generated a set of ST ( R7A , R21A , P132A , and W147A ) and Aα ( D177A , R183A , E216A , Q217A , and R258A ) mutants and performed in vitro binding assays . We found that substitution of alanine for residues Arg7 , Arg21 , or Pro132 of ST abrogated interaction between ST and wild-type Aα ( Figure 7 ) . In addition , we found that the W147A ST mutant showed reduced binding to Aα compared to wild-type ST . Analysis of Aα mutants indicated that single alanine substitutions at position Glu216 disrupted PP2A A-ST interaction ( Figure 7 ) . These observations provide strong support to the A-ST interface observed in our crystal structure .
Here we report the crystal structure of SV40 ST in complex with the murine PP2A A subunit . It is striking that all four A-ST complexes in the asymmetric unit of our crystal lattice have essentially the same structure , except the flexible HEAT repeats 11–15 of the A subunit that are not involved in the A-ST interaction . This observation argues strongly that the ST structure as well as the A-ST interface observed in our crystal structure are independent of crystal packing and should be physiologically relevant . Since the human and murine Aα subunits are identical except for one residue that is distant from the ST binding site ( Ser324 in human , Thr324 in mouse ) , this structure is likely representative of the interaction of ST with the human PP2A . Moreover , since most ST residues in the structural core and the A-ST interface are conserved among ST and middle t ( MT ) proteins in the polyomavirus family , this structure provides a structural basis for understanding the oncogenic activities of each of the ST and MT proteins in the polyomavirus family . We found that the N-terminal J and the C-terminal unique domains of ST are structurally coupled . The conserved hydrophobic interface between the J and unique domains may be important for the structural integrity of ST . Previous studies have been shown that each ST molecule contains two zinc ions [33 , 34] . Since ST contains two cysteine cluster motifs ( CXCXXC ) that are absolutely conserved , it was proposed that ST may resemble GAL4 , which contains a Zn ( II ) 2Cys6 binuclear cluster [33 , 34] . In our ST crystal structure , instead of forming a binuclear cluster , these two zinc ions are located in two separate positions and form two novel zinc-binding motifs . The first and second cysteine clusters ( residues 111–116 , 138–143 , respectively ) form two zinc-binding motifs together with the conserved Cys103 and His122 , respectively . Both zinc ions interact with helix α5 and stabilize the structure of the C-terminal unique domain . The Aα subunit of PP2A consists of 15 HEAT repeats , with each repeat containing two antiparallel helices . Overall the PP2A Aα subunit forms a horseshoe shape structure . SV40 ST “rides” on the structural ridge formed by intra-HEAT-repeat loops of PP2A Aα subunit ( Figure 1 ) . Previous work has demonstrated that the unique domain of SV40 ST is the essential binding domain for PP2A AC core complex [5] . In our structure the second zinc-binding motif , in particular the L6 loop , forms extensive interactions with the PP2A Aα subunit . In support of our structural observations , in vitro structure-directed mutagenesis studies demonstrate that the interface between the second zinc-binding motif , in particular Pro132 that intercalates between Trp140 and Phe141 of PP2A A subunit , is essential for A-ST interactions ( Figure 7 ) . Point mutations of two residues in the ST J domain ( R7A and R21A ) , or mutation of an A subunit residue that forms a hydrogen bond with ST Arg21 ( E216A , Figure 3 ) , all abolish the A-ST interaction ( Figure 7 ) , confirming the J domain is directly involved in A subunit interaction . In addition , this structure is completely consistent with previous mutagenesis studies of PP2A A subunit [44] , since essentially all A subunit mutations that have weaker ST binding activities map to the A-ST interface in our structure . Previous work has indicated that a region of the unique domain encompassing the first cysteine cluster is necessary for the binding of ST to PP2A , and its N-terminal flanking region ( residues 97–103 ) are also important for PP2A interactions [35 , 47] . However , the first zinc-binding motif does not interact with the A subunit of PP2A in the crystal structure . Instead , the crystal structure suggests that the first zinc-binding motif may directly interact with the C subunit near its active site , since the first zinc-binding motif is spatially close to the active site of the PP2A C subunit in the structural superposition of PP2A and A-ST complexes ( Figure 5 ) . This hypothesis is supported by structural flexibility of the scaffold A subunit . Our structural comparison indicates that HEAT repeats 10–13 of PP2A Aα subunit have substantial structural flexibility ( Figure 6 ) . This flexibility of the Aα subunit may allow for the accommodation of different types of regulatory B subunits or B-like proteins , such as ST , into the AC core enzyme that interact with both A and C subunit . Consistent with prior reports [23 , 24] , we failed to observe stable direct interaction between purified ST and PP2A C subunit using a glutathione S-transferase ( GST ) pull-down assay ( unpublished data ) . It is possible that ST directly interacts with the C subunit via the formation of a stable complex between ST and the A subunit . Therefore , ST may form trivalent interactions with the PP2A AC complex—two of them ( via the J domain and the second zinc-binding motif ) interacting with the A subunit , and the third one ( via the first zinc-binding motif ) binding to the C subunit . This hypothesis explains why the first zinc-binding motif does not directly interact with PP2A A subunit , yet is required for the interaction with and the inhibition of phosphatase activity of PP2A AC core complex [35] . This model is also supported by the observation that ST fragments containing both J domain and the first but not the second zinc-binding motif interact with PP2A AC complex and inhibit PP2A AC dimer phosphatase activity [29 , 35] . ST may interfere with substrate binding via its interaction near the active site of the C subunit . Therefore , in addition to competing with the PP2A B subunit for PP2A A subunit binding , ST may directly modulate the phosphatase activity of the AC core complex , which accounts for substantial proportion of PP2A enzyme in the cell . Future work will be needed to understand if and how ST may directly interact with the C subunit . Prior studies have demonstrated that the unique domain but not the J domain was sufficient for interaction with PP2A AC complex . Although not essential for A-ST interaction , the deletion of the J domain significantly decreased the inhibitory activity of ST on the PP2A AC core dimer [35] , suggesting that the J domain enhances the binding of ST to the PP2A A subunit . Consistent with this view , mutation of either Arg7 or Arg21 , two residues on the J domain surface involved in PP2A A interaction , disrupts the interaction between ST and A subunit . Alternatively , the J domain may play a role in stabilizing the spatial position of the first zinc-binding motif by allowing its efficient interaction with the C subunit . This interaction may be particularly important in the AC-ST complex formation , because the structural flexibility of HEAT repeats 10–13 of the A subunit may not permit the C subunit to stay in a fixed position and interact with ST efficiently . Indeed , although the unique domain of ST binds to PP2A A , this binding fails to inhibit PP2A AC phopshatase activity [35] . In this regard , we note that the second zinc-binding motif binds to the A subunit primarily through loop–loop interactions and on the concave side of the A subunit structure only , while the J domain interacts with the convex side of A subunit . While the second zinc-binding motif may be the primary docking site , the J domain may fix the relative orientation between ST unique domain and the A subunit , with the N-terminal J and C-terminal unique domains sitting on the convex and concave side of the horseshoe shape , respectively . This orientation stabilization may be important for the first zinc-binding motif to effectively inhibit the phosphatase activity of the PP2A C subunit . In addition to the inhibition of the phosphatase activity of the PP2A AC dimer , the J domain may also play a role in the oncogenic activity of ST by providing an additional binding site for Hsp70 , even when in complex with the PP2A AC complex , as suggested by our crystal structure . The potential simultaneous interaction with PP2A and Hsp70 may couple these two functions of ST . For example , ST may bring PP2A and Hsp70 together to allow for the dephosphorylation of protein ( s ) bound to Hsp70 . In summary , our structural and biochemical studies reveal the structure of the ST family and define the interaction between ST and the A subunit of PP2A . In addition , our work suggests that ST may directly interact and regulate the activity or substrate specificity of the PP2A catalytic C subunit , and Hsp70 may bind to PP2A-bound ST and thus define PP2A activity and/or substrate specificity . Taken together , our work provides a structural basis for the oncogenic activity of ST and MT antigens in the polyomavirus family . Since ST binds to PP2A Aα in a manner similar to that used by the regulatory B subunits , these findings provide not only new insights into the regulation of PP2A but may also provide a foundation for the development of small molecules that alter the function of PP2A .
Full-length SV40 ST ( strain VA45‐54‐2 ) and full-length mouse PP2A Aα subunit with tobacco etch virus protease ( TEV ) cleavage sites were cloned into pGEX4T1 vector ( Amersham Biotech , http://www . gelifesciences . com ) and pET28a vector ( Novagen , http://www . emdbiosciences . com/html/NVG/home . html ) , respectively , and co-transformed into E . coli strain BL21 ( star ) ( Invitrogen , http://www . invitrogen . com ) . Mouse and human Aα subunits are identical in protein sequence , except for one residue that is distant from the ST binding site ( Ser324 in human , Thr324 in mouse ) . Coexpression of Aα subunit and SV40 ST was induced by the addition of 0 . 1 mM IPTG at OD600 = 0 . 6 upon shifting the temperature from 37 °C to 18 °C , and cells were grown for an additional 18 h . Cells were then collected by centrifugation and resuspended with lysis buffer ( 30 mM Tris-HCl [pH 8 . 0] , 50 mM NaCl , 5 mM β-mercaptoethanol ) including protease inhibitors ( PMSF , leupeptin , and benzamidine ) . Resuspended cells were lysed by sonication , and cell debris removed by centrifugation at 26 , 000 g for 1 h . Soluble fractions were filtered with 0 . 8 μm syringe filters and applied into a Ni-NTA affinity column pre-equilibrated with 30 mM Tris-HCl ( pH 8 . 0 ) , 50 mM NaCl , 5 mM β-mercaptoethanol . Target protein complexes ( the Aα subunit with GST-tag and SV40 ST with His-tag ) were eluted with elution buffer ( 30 mM Tris-HCl [pH 8 . 0] , 50 mM NaCl , 300 mM imidazole , 5 mM β-mercaptoethanol ) and dialyzed overnight at 4 °C in 30 mM Tris-HCl ( pH 8 . 0 ) , 50 mM NaCl , 5 mM DTT . Dialyzed protein was applied to a GST affinity column to remove free SV40 ST , and on-column cleavage with TEV protease was performed at 4 °C overnight . The flow-through fraction of the GST column was reapplied into the Ni-NTA column to remove cleaved His-tag and TEV protease . The flow-through fraction of the Ni-NTA column was concentrated and applied into a Superdex 200 size-exclusion column ( Amersham Biotech ) pre-equilibrated with 30 mM Tris-HCl ( pH 8 . 0 ) , 50 mM NaCl , 5 mM DTT . Fractions of the heterodimeric complex of full-length Aα subunit and full-length SV40 ST were pooled and concentrated up to 10 mg/ml and used for the crystallization trial . Crystals of the A-ST complex were obtained at room temperature with the hanging drop vapor diffusion method . After optimizing initial conditions and extensively searching for additives , combined with microseeding , the optimal crystallization condition was 16% PEG 3350 , 0 . 2 M ammonium formate , 30 mM spermine , 6% 6-aminocaproic acid , 10 mM DTT . Crystals were cryoprotected by gradually increasing the PEG 3350 concentration up to 26% and frozen with liquid nitrogen . Data collection was performed at Advanced Light Source beamline 5 . 0 . 1 . The best dataset has a resolution of 3 . 1 Å , with high redundancy . The space group was P1 , with the unit cell dimensions a = 94 . 69 Å , b = 105 . 50 Å , c = 111 . 98 Å , α = 115 . 64° , β = 109 . 57° , γ = 94 . 10° . These datasets were integrated and scaled using HKL2000 and SCALEPACK [48] . There were four heterodimers in one asymmetric unit with 54% solvent content . Initial molecular replacement was performed with Phaser [49] by using the Aα subunit structure in the trimeric PP2A complex [25] as the search model . The program Phaser found all four Aα subunits with pseudo D2 symmetry . However , both Rfree and Rwork were over 55% after overall rigid-body refinement , suggesting substantial conformational changes in the Aα subunit . Therefore , rigid body refinement was performed in Refmac5 , with every single HEAT repeat as a separate group [50] . After this rigid body refinement , Rfree dropped to about 40% . Since SV40 ST was demonstrated to contain two zinc-binding sites [34] that gave significant anomalous signal at the experimental wavelength ( λ = 1 . 000 Å ) , the same dataset was scaled as an anomalous dataset using HKL2000 . The anomalous difference map , calculated using phases derived from the refined molecular replacement solution , gave the positions of eight zinc sites ( for four ST molecules in the asymmetric unit ) without any ambiguity . Zinc sites were further refined with Sharp [51] , and phases were calculated together with the partial model of the Aα subunit . The calculated map showed clear density of SV40 ST after the density modification with DM . Since the conformation of each four Aα subunit was different , noncrystallographic symmetry averaging was helpful only to build SV40 ST . Further model building was done with Xtalview [52] and COOT [53] . TLS refinement was performed using Refmac5 in the CCP4 suite [50] . Because of the conformational variations in the A subunit , noncrystallographic symmetry averaging was not applied during refinement . The final model has an Rfree of 30 . 5% , and an Rwork of 24 . 4% . Zinc sites are confirmed by anomalous difference map . None of the nonglycine residues are in the disallowed region of the Ramanchandran plot . The final structural models of ST contain residues 1–44 , 48–78 , 85–172 ( chain e ) ; 1–40 , 49–74 , 87–172 ( chain f ) ; 1–41 , 51–70 , 92–172 ( chain g ) ; and 1–41 , 51–73 , 94–172 ( chain h ) . GST-tagged ST and Aα mutants were generated using the QuickChange Site-Directed Mutagenesis Kit ( Stratagene ) and expressed in E . coli strain BL21 ( star ) . Wild-type Aα and Aα mutants were isolated with glutathione-sepharose ( Amersham Biosciences ) GST-tags were removed from wild-type ST and ST mutants and equal amounts of wild-type ST or ST mutants were added to Aα-glutathione-sepharose precipitates . Binding assays were performed in 30 mM Tris ( pH 8 . 0 ) , 50 mM NaCl , 5 mM DTT , 0 . 2% NP-40 buffer for 4 h . The beads were washed five times , and the proteins were eluted with reduced glutathione , followed by SDS-PAGE and immunoblotting . For immunoblotting , we used affinity-purified polyclonal antibodies against SV40 ST [16] and monoclonal antibodies ( clone 6F9 ) against Aα ( Abcam , http://www . abcam . com ) .
Coordinates and structural factors have been deposited in the Protein Data Bank ( PDB , http://www . rcsb . org/pdb ) with the accession code 2PF4 . | The study of how DNA tumor viruses induce malignant transformation has led to the identification of key pathways that also play a role in spontaneously arising cancers . One such virus , simian virus 40 ( SV40 ) , produces two proteins , the large T and small t antigens , that bind and inactivate tumor suppressor genes important for cell transformation . Specifically , SV40 small t antigen ( ST ) binds to and perturbs the function of the abundant protein phosphatase 2A ( PP2A ) . PP2A is a family of heterotrimeric enzymes , composed of a structural A subunit , a catalytic C subunit , and one of several regulatory B subunits . Here we have determined the structure of SV40 ST in complex with the PP2A structural subunit Aα . SV40 ST consists of an N-terminal J domain and a C-terminal unique domain that contains two separate zinc-binding motifs . SV40 ST binds to the same region of PP2A as the regulatory subunit B56 , which provides a structural explanation for the displacement of regulatory B subunits by SV40 ST . Taken together , these observations provide a structural basis for understanding the oncogenic functions of ST . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
]
| [
"biochemistry",
"infectious",
"diseases",
"virology",
"in",
"vitro",
"chemistry",
"chemical",
"biology"
]
| 2007 | Structural Basis of PP2A Inhibition by Small t Antigen |
Phlebotomine sand flies are vectors of human leishmaniases , important neglected tropical diseases . In this study , we investigated diel patterns of oviposition behavior , effects of visual cues on oviposition-site selection , and whether these affect the attraction of gravid Phlebotomus papatasi ( Scopoli ) , the vector of old-world cutaneous leishmaniasis , to olfactory cues from oviposition sites . To evaluate these questions , we conducted a series of experiments using attraction and oviposition assays within free-flight test chambers containing gravid females entrained under a 14:10 hrs light:dark photoperiod . By replacing sticky-screens or moist filter papers every three hours , we showed that oviposition site search occurs mainly in the latest part of the night whereas peak oviposition occurs during the early part of the night . Behavioral responses to olfactory oviposition cues are regulated by time-of-day and can be disrupted by transient exposure to a constant darkness photoperiod . Gravid females , but not any other stage , age , or sex , were attracted to dark , round oviposition jars , possibly resembling rodent burrow openings . This visual attraction disappeared in the absence of an illumination source . Egg deposition rate was not affected by jar color . Olfactory cues had the strongest effect when the visual cues were minimal . Our study showed , for the first time , that visual cues in the form of oviposition-site color , lighting level , and photoperiod are important in guiding the oviposition behavior of phlebotomine sand flies . Furthermore , such visual cues could modify the flies’ sensitivity to olfactory oviposition cues . Our results suggest that chemosensory and visual cues are complementary , with visual cues used to orient gravid females towards oviposition sites , possibly at long- to medium-ranges during crepuscular periods , while olfactory cues are used to approach the burrow in darkness and assess its suitability at close-range . Implications to sand fly control are discussed .
Phlebotomine sand flies ( Diptera: Psychodidae ) transmit protozoan parasites ( Leishmania spp . ) , as well as pathogenic bacteria ( Bartonella bacilliformis ) and viruses [1–3] . Most significant are the human leishmaniases that , following malaria and dengue , are the most pervasive vector-borne diseases . Leishamaniases are found worldwide in warm tropical , semi-arid , and arid environments . The population at risk is estimated at approximately 350 million people with annual incidence estimated at 1–1 . 5 million cases of cutaneous leishmaniasis , 500 , 000 cases of visceral leishmaniasis , and an estimated Disability-Adjusted Life Year ( DALY ) of 2 . 3 million [4 , 5] . Leishmaniasis is distributed in the poorest regions of the world , and considered a neglected disease in terms of under-reporting , under-funding of research , and inadequate health care [6–8] . With no vaccine to protect against the etiologic agent , reduction of exposure to sand fly bites using insecticides is the most effective prevention measure [6 , 9 , 10] . This strategy is seriously constrained , however , by the evolution of insecticide resistance , which renders residual sprays less effective [10 , 11] . Also , given the limited knowledge of sand fly breeding sites [12 , 13] , source reduction using biolarvicides has not been an effective approach [10 , 14] . Hence , a more focused , targeted , and efficient control method is urgently needed [15] . An alternative approach to delivery of the insecticide to the vector is to bring the vector to the insecticide using attractants [16–18] . Oviposition-site attractants can provide the basis for a novel control and surveillance approach targeting gravid females that are typically responsible for pathogen transmission and population amplification . Oviposition traps have been used for the control of mosquitoes [19–23] , but no such tool yet exists for the control of sand flies . Unlike other biting Diptera , sand flies develop in terrestrial rather than aquatic microhabitats [1 , 3] . Eggs are typically laid in soil rich in organic material on which the larvae feed and develop through four instars before pupation and adult emergence [1 , 3 , 24 , 25] . For both new- and old-world sand fly species , olfactory cues originating from organic matter of various sources have been identified as important sources of oviposition attractants and stimulants [26–29] , and some of these cues were shown to be produced by microbes [30 , 31] . Conspecific eggs are also attractive to sand flies [32–34] , with dodecanoic acid identified as an oviposition stimulant for Lutzomyia longipalpis ( Lutz and Neiva ) [35] . Insects orient to oviposition sites using multi-modal sensory systems . Chemosensory cues play a central role as oviposition attractants and as lures in oviposition traps , especially for nocturnal insects [18 , 20 , 21] . However , visual cues may also play an important role , especially in diurnal and crepuscular insects [19] . For example , Aedes albopictus ( Skuse ) mosquitoes prefer to lay eggs in black plastic cups rather than in white , blue , or orange cups , or cups with black-and-white contrasting patterns [36] , and Aedes triseriatus ( Say ) and Anopheles spp . visually prefer to oviposit in dark water [37 , 38] . Although sand flies were shown to respond to visual cues [39–44] , nothing is known regarding their visual assessment of oviposition sites . Many sand fly species , including our study species Phlebotomus papatasi ( Scopoli ) , are tightly associated with animal burrows within which they find shelter , blood-feed , lay eggs and where coprophagic larvae feed on decaying fecal matter [13 , 45] . Therefore , it is reasonable to hypothesize that gravid females would be visually attracted to dark round objects resembling burrow openings . The timing of trap deployment may also affect the efficacy of oviposition traps because different species prefer to lay eggs at different times . The circadian oviposition behavior of mosquitoes has been relatively well studied , but not so for sand flies . For example , Ae . aegypti , a day-time biter , consistently laid eggs in the afternoon irrespective of ambient temperatures , and when maintained on a 12:12 hrs photoperiod it laid eggs just before light-out [46] . Given the nocturnal or crepuscular nature of sand flies [3 , 47 , 48] it is often assumed that they oviposit at night [49 , 50] . Indeed , gravid females of Ph . papatasi , which typically breed in rodent burrows , were observed to exit the burrows in the early part of the night , suggesting that oviposition-site search activity may start early in the night [51] . In contrast , the majority of Phlebotomus orientalis ( Parrot ) gravid females were caught after midnight , suggesting that oviposition-related activity occurs mostly in the later part of the night [52] . However , both studies relied on indirect evidence based on locomotor activity , and to date nothing is known about the timing of oviposition-related activities in sand flies . An important observation with mosquitoes was their integration of sensory inputs from olfactory cues and visual cues [53–55] , and that it may vary during the diel cycle [56–59] . With sand flies , evidence for such interactions is scant . By studying circadian clock gene expression of Lu . longipalpis , researchers observed an evening peak that anticipated the lights-off cue [60] . With Ph . papatasi , adding a humidity source to a CDC trap increased capture only for white or clear traps but not for black traps [40] . To our knowledge , no studies with either mosquitoes or sand flies have documented such interactions in the context of oviposition behavior . In this work , we used Ph . papatasi , a vector of Leishmania major , the etiological agent of zoonotic cutaneous leishmaniasis ( ZCL ) in the Middle East [3 , 61] . We investigated its diel pattern of attraction to oviposition sites and timing of egg deposition , the effects of visual cues on oviposition-site selection and evaluated whether visual cues and photoperiod affect the attraction of gravid females to known olfactory oviposition cues . Specifically , our study focused on the following four aims: ( 1 ) Evaluate if time-of-day affects: ( A ) the timing of oviposition-site search activity , ( B ) timing of egg deposition , and ( C ) attraction to known olfactory cues; ( 2 ) Evaluate if transient changes in the photoperiod affect the attraction of sand flies to known olfactory oviposition cues; ( 3 ) Evaluate if visual cues , in the form of oviposition site color and illumination level , affect oviposition-site selection; and ( 4 ) Evaluate if visual cues affect the attraction of sand flies to known olfactory oviposition cues .
As part of the sand fly colony maintenance , sand fly blood-feeding on ICR white mice was conducted at the SoBran research facility , Gateway University Research Park in North Carolina under SoBran IACUC Protocol number UNC-002-2016 dated 12/12/2016 . This protocol was approved by the SoBran Gateway Animal Care and Use Committee; Chaired by Stephanie Strahan . This institution is USDA-registered and PHS-assured , therefore , SoBran’s animal care and use standards adhere to the Animal Welfare Act ( USC 7 , Sections 2131–2159 , 1966 , as amended ) and Animal Welfare Regulations ( 9 CFR , Chapter 1 , Parts 1–4 ) ; the PHS Policy on Humane Care and Use of Laboratory Animals USHHS , NIH , 2015 ) ; and The Guide for the Care and Use of Laboratory Animals ( NRC , 8th edition ) , among others . In experiment #1 , we evaluated the change in the number of flies caught on a sticky screen trap or the number of eggs laid on a filter paper over a diel cycle . We analyzed these longitudinal count data using a random-intercept negative-binomial regression model , with cage as the clustering factor [62] . We used paired Student’s t-test to compare the effect of treatment versus control within a cage . In order to compare the relative attraction of sand flies for the treatment , among treatment levels ( e . g . , illumination level , time window , jar color ) we calculated the proportion of flies caught in the treatment jar compared with total flies trapped in the two jars combined ( hereafter , preference ) . Values above 0 . 5 indicated preference for the treatment and values below 0 . 5 indicated aversion to it . Given that “preference” is a proportion , we analyzed these data using weighted logistic regression , with number of flies making a behavioral choice used as the weighting factor [63] . Analyses were conducted using Stata ( StataCorp . , College Station , TX ) .
In this experiment , sand flies that were reared under a 14:10 light:dark photoperiod were tested for their olfactory preference to a known oviposition attractant ( larval frass ) under different photoperiod conditions: the 14:10 photoperiod to which they were entrained or constant darkness . Flies exposed to the entrainment photoperiod exhibited stronger attraction to the rearing medium compared with flies switched transiently to constant dark conditions with an average preference rate of 60 . 1 ± 3% and 49 . 0 ± 4% , respectively ( t = 2 . 15 , df = 59 , P = 0 . 03 ) ( Table 1 ) . On average , under the standard 14:10 photoperiod , significantly more flies were trapped in the frass-baited jar than in the sand only control ( 4 . 52 ± 0 . 3 versus 3 . 06 ± 0 . 3 flies , respectively; paired-t = 3 . 32 , df = 30 , P = 0 . 001 ) , whereas in assays conducted under constant dark conditions ( 0:24 light:dark ) no significant difference was found between the two jar types ( 4 . 1 ± 0 . 5 versus 3 . 7 ± 0 . 3 , respectively; paired-t = 0 . 612 , df = 29 , P = 0 . 27 ) . This pattern was consistent in the two rooms in which the assays were conducted ( Table 1 ) . Also , no significant difference was found with respect to the response rates , with an average response rate of 45 . 3 ± 3% and 46 . 3 ± 3% under these respective conditions ( t = 0 . 238 , df = 59 , P = 0 . 81 ) . Here too , results were consistent in the two rooms in which the experiment was conducted ( Table 1 ) . We conducted two-choice assays , using free-flight cages containing a pair of sticky traps of the same color with one jar containing larval frass ( olfactory cues ) and the other jar serving as control ( moist sand only ) . Under complete darkness , gravid females exhibited clear olfactory attraction to the treatment jar irrespective of jar color ( Fig 4 ) . As light level increased to low-level ( a 120-fold increase in brightness ) , olfactory attraction to the frass-containing jars decreased and became non-significant for all jar colors ( Fig 4 ) . This decrease was particularly steep and statistically significant for the black jars where attraction disappeared altogether ( Fig 4 ) . For other jar colors , preference showed a non-significant decreasing trend . No significant difference was found among jar colors ( deviance = 2 . 89 , df = 2 , P = 0 . 23 ) . With further increase in luminance ( dim lighting level , 2-fold brighter than low-light ) , olfactory attraction to frass-baited jars remained low and non-significant for both black and white jars . For transparent jars , however , preference for the frass-baited jars increased and became significantly greater than for the black jars ( paired-t = 2 . 73 , P = 0 . 014 ) ( Fig 4 ) .
Essentially , nothing was known about the timing of oviposition-related behaviors in sand flies . With sand flies being predominantly nocturnal or crepuscular [3 , 47 , 48] , it was assumed that they oviposit during the dark phase [49 , 50] . However , the questions as to when they search for oviposition sites and when they actually deposit their eggs have remained unanswered . With specially-designed exit-entrance traps , Yuval and Schlein ( 1986 ) observed that gravid Ph . papatasi females constituted a large fraction of flies exiting rodent burrows in the early part of the night , suggesting that oviposition-site search activity is more likely to occur early in the night [51] . In contrast , our results show a clear peak in activity related to oviposition-site search activity in the last three hrs of the scotophase that seems to also spill-over into the dawn and early hours of the morning . Furthermore , our results indicate that this time-period ( the last three hours of the scotophase ) is also when olfactory attraction to known oviposition attractants peaks . Obviously , our experiments should be repeated using field-caught or freshly produced lab-reared flies under field conditions . However , assuming our findings extend to natural sand fly populations , these results suggest circadian regulation of oviposition site-search behavior that affects the sand fly’s activity level and possibly sensory responsiveness to olfactory and visual oviposition cues . Why oviposition site-search activity is delayed until the later part of the night is yet to be determined . It is possible that , initially , gravid females need to search for a sugar-meal in order to replenish their energy reserves before switching to an oviposition site-search mode . This speculation is consistent with an observation by Yuval and Schlein ( 1986 ) showing that a high proportion of females entering burrows at the later part of the night were sugar fed , compared with those exiting burrows in the earlier part of the night [51] . With respect to timing of egg-laying , we observed a distinct peak of oviposition during the first three hours of the scotophase ( which also included 1 hr of dusk ) and lower oviposition rates before or after . These results highlight the adaptive value of circadian organization , which enables the coordination of metabolic processes and behaviors , including the expression of sensory receptors , relative to local entrainment conditions . The circadian organization of Ph . papatasi gravid females may be as follows: gravid females initially search for a sugar-meal during the night , then they switch to actively seeking a suitable oviposition site ( often a rodent burrow ) late in the night , they shelter within the burrow during the day , and then , possibly triggered by declining lights , other environmental changes such as relative humidity , or activity of the rodent host , they initiate egg-deposition at the beginning of the following night . Such a pattern is consistent with the behavior of many nocturnal mosquito species that tend to oviposit during dusk and early evening [19] . It is possible that a resting period within the burrow , before oviposition , enables females to complete egg maturation , mate , or even take supplementary blood meals from the resting host within the burrow . We cannot , however , discount the possibility that gravid females might also be laying eggs during the day within the dark burrow , and this possibility will be evaluated experimentally in the future . Attraction to a known olfactory substrate ( larval frass ) was shown to vary along the diel cycle with few flies attracted during the photophase or early in the scotophase , a slight increase towards midnight , and many flies attracted to frass during the last three hours of the night . Furthermore , our results indicate that the level of olfactory attraction of gravid sand flies to an attractive substrate is affected by the photoperiod regime to which they are exposed . Specifically , sand flies entrained under a natural photoperiod ( 14:10 hrs light:dark , including 1 crepuscular hr at the start and end of the photophase ) exhibited greater attraction to frass under these conditions than when switched to constant darkness ( 0:24 hrs ) for the 6-hr assay . On the other hand , activity levels , as expressed in terms of the flies’ response rate , did not differ between the two photoperiod regimes ( Table 1 ) . These results suggest that visual cues associated with the dawn crepuscular period and early photophase may affect the flies’ oviposition site-selection behavior by modifying their sensory responses to volatile olfactory cues . We validated experimentally that preference for the black jars was due to visual cues and not due to olfactory attraction to potential volatile components of the black paper , because flies did not differentiate between jars containing concealed black or white papers . Gravid females clearly preferred black jars over transparent ( clear ) jars or white jars , but they failed to differentiate between white and clear jars . To further explore the role of visual cues , we performed the black-versus-clear-jar choice experiment at three illumination levels: complete darkness , low light ( simulating dark moonless night conditions ) , and dim light ( simulating moon-lit night conditions ) . In the absence of any light , jars of different colors should not be visually differentiable , and indeed , sand flies did not exhibit a preference for black or clear jars . In contrast , in low and dim light gravid females clearly preferred the black jar over the clear jar . Likewise , gravid females preferred the black jar over the white jar in dim light , but showed no preference between clear and white jars . We also discovered that the effect of the visual cue pertains only to the oviposition site-search phase but not to the actual egg-laying phase . When provided with the choice of laying eggs on a wet black versus a wet white paper no difference in egg numbers was found . This observation is consistent with previous studies on both old- and new-world sand flies indicating that oviposition stimulation is triggered by chemosensory cues of fecal or conspecific origin [26 , 27 , 29–31] . Most striking was the fact that attraction to black jars over white jars was only exhibited by mature gravid blood-fed females . In contrast , non-blood-fed mature females , 74% of which were autogenous , did not exhibit such visual attraction ( although a slight trend in that direction was suggested ) . This suggests that the level of female gravidity ( % gravid ) and fecundity ( per-capita rate of egg production ) , which were shown to be higher in blood-fed females , are associated with the sensory acuteness of these females to visual oviposition cues . Visual attraction of adult males to black jars could be expected to mirror that of gravid females if mate-seeking males would await incoming gravid females in burrow entrances . The non-significant attraction trend we observed might be consistent with such a trend , but more study on that is obviously required . We also observed that activity level of young flies of both genders was considerably lower than that of mature flies , suggesting that activity of young flies may be stimulated by other cues such as those indicating presence of sugar- or blood-meal that are more relevant to flies at this stage [17 , 51] . Overall , this stage-specificity of responses to visual cues suggests that the black jars are perceived by gravid females as a visual oviposition cue and not as a generic cue . This result somewhat differs from findings that host-seeking Ph . papatasi flies were attracted to CDC traps where the collection cup was wrapped in black paper [40] . Attraction of gravid flies to black round jars that might be perceived as rodent burrow openings is consistent with our knowledge of the biology of Ph . papatasi that in arid areas are highly dependent on rodent-host burrows ( typically Psammoys obesus or other Meriones spp . ) as suitable oviposition and shelter sites [64 , 65] . The slight , non-significant affinity of sand flies of both genders and age/stages to black jars suggests a certain generality of this cue . Yet , the strong attraction of gravid females to a dark and round object suggests that they have evolved to become highly attuned to such a visual cue , which would facilitate the detection of sparsely distributed oviposition sites in arid environments . In complete darkness , gravid females exhibited significant olfactory attraction to used larval frass , a known source of volatile oviposition attractants [26 , 27] , irrespective of oviposition jar color . As the light level increased , however , the degree of attraction to the frass-containing jar decreased or even disappeared altogether as visual cues became more apparent . This effect was particularly evident with black jars , which , as described above , provided the strongest visual cue to gravid females . Our results suggest that although visual and olfactory cues are integrated in the central nervous system to drive behavior , different sensory modalities take precedence during the orientation sequence . Such effects have been observed with mosquitoes [19] . In the confines of the relatively small , dimly lit , arena under our experimental conditions , visual cues appear to prevail over olfactory cues . Under natural conditions , however , we suspect that a cryptic dark burrow entrance on the surface of a relatively poorly lit ground might not be as apparent as a jar with a black vertical silhouette against a light-colored background , as used in our assays . We further suspect that chemosensory and visual cues are complementary , with visual cues used to orient gravid females towards potential oviposition sites at long-to-medium ranges during the crepuscular periods , while olfactory cues are used to approach the burrow in darkness and assess its suitability at close-range . Direct evaluation of the complementary effect of these cues would require a separate factorial-design experiment , including manipulation of the visual apparency ( signal-to-noise ) of the burrow entrance . Results of this study add important novel insights for our understanding of the basic biology of Phlebotomine sand flies and have important implications for management of sand fly populations . First , in the context of the development of an oviposition trap , it is clearly necessary to take visual cues into account . Such a trap should be black and round resembling a rodent burrow entrance . Optimizing trap shape and design is an important next step , and of particular importance are its vertical aspect and apparency relative to the ground terrain . Then , visual cues should be complemented with chemosensory cues that include olfactory stimuli to guide the flies to the burrow and contact and gustatory cues to arrest females and stimulate oviposition . Our results show that visual cues might transcend olfactory cues , but this might not be the case under natural light and landscape conditions . Our findings that oviposition site search activity occurs in the late scotophase would suggest that traps should be activated in the field during the second part of the night and early morning . This would require that traps either be deployed in the later part of the night or that the olfactory lure is operational throughout the night . Nevertheless , these results require confirmation in the field with traps or rodent-burrow monitors under various natural-light conditions . Finally , results of this study contribute three important features and conditions for optimizing bioassays for further studies of sand fly oviposition site-selection . First , when screening olfactory attractants , visual cues should be minimized and probably transparent jars would serve this purpose best . Second , attraction bioassays with gravid sand flies should be implemented in the later part of the scotophase under minimal light , conditions under which chemosensory information takes precedence over visual cues . Finally , flies should be reared and maintained using a natural photoperiod representative of summer conditions to optimize the sensitivity of bioassays . These recommendations are already implemented in current studies in our lab . | Sand flies are vectors of human leishmaniases , an important neglected tropical disease . An alternative approach to the conventional delivery of an insecticide to the vector is to bring the vector to the insecticide using oviposition ( egg-laying ) -site attractants . Olfactory cues originating from organic matter have been identified as important sources of oviposition attractants . However , nothing is known regarding visual assessment of oviposition sites by sand flies . Also , little is understood about diel egg-laying patterns of sand flies . Finally , it is not known if visual cues and time-of-day may affect their sensitivity to olfactory cues . In this study , we investigated these questions in a series of lab experiments using free-flight cage arenas , with Phlebotomus papatasi ( vector of cutaneous leishmaniasis ) . We showed that peak oviposition-site search and sensitivity to olfactory cues occurs mainly in the latest part of the night whereas peak oviposition occurs during the early part of the night . We demonstrated that only gravid females , but not any other stage or sex , were attracted to dark , round oviposition jars resembling burrow openings . Finally , we showed that sensitivity to olfactory cues is reduced in the presence of strong visual cues and in the absence of natural diel photoperiod . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"laboratory",
"equipment",
"engineering",
"and",
"technology",
"social",
"sciences",
"sand",
"flies",
"light",
"neuroscience",
"animals",
"electromagnetic",
"radiation",
"reproductive",
"physiology",
"artificial",
"light",
"filter",
"paper",
"vision",
"insect",
"vectors",
"infectious",
"diseases",
"disease",
"vectors",
"insects",
"arthropoda",
"physics",
"mosquitoes",
"psychology",
"sensory",
"cues",
"eukaryota",
"equipment",
"physiology",
"oviposition",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"sensory",
"perception",
"physical",
"sciences",
"organisms"
]
| 2019 | Diel periodicity and visual cues guide oviposition behavior in Phlebotomus papatasi, vector of old-world cutaneous leishmaniasis |
Lactase persistence ( LP ) is common among people of European ancestry , but with the exception of some African , Middle Eastern and southern Asian groups , is rare or absent elsewhere in the world . Lactase gene haplotype conservation around a polymorphism strongly associated with LP in Europeans ( −13 , 910 C/T ) indicates that the derived allele is recent in origin and has been subject to strong positive selection . Furthermore , ancient DNA work has shown that the −13 , 910*T ( derived ) allele was very rare or absent in early Neolithic central Europeans . It is unlikely that LP would provide a selective advantage without a supply of fresh milk , and this has lead to a gene-culture coevolutionary model where lactase persistence is only favoured in cultures practicing dairying , and dairying is more favoured in lactase persistent populations . We have developed a flexible demic computer simulation model to explore the spread of lactase persistence , dairying , other subsistence practices and unlinked genetic markers in Europe and western Asia's geographic space . Using data on −13 , 910*T allele frequency and farming arrival dates across Europe , and approximate Bayesian computation to estimate parameters of interest , we infer that the −13 , 910*T allele first underwent selection among dairying farmers around 7 , 500 years ago in a region between the central Balkans and central Europe , possibly in association with the dissemination of the Neolithic Linearbandkeramik culture over Central Europe . Furthermore , our results suggest that natural selection favouring a lactase persistence allele was not higher in northern latitudes through an increased requirement for dietary vitamin D . Our results provide a coherent and spatially explicit picture of the coevolution of lactase persistence and dairying in Europe .
Lactase persistence ( LP ) is an autosomal dominant trait enabling the continued production of the enzyme lactase throughout adult life . Lactase non-persistence is the ancestral condition for humans , and indeed for all mammals [1] . Production of lactase in the gut is essential for the digestion of the milk sugar lactose . LP is common in northern and western Europeans as well as in many African , Middle Eastern and southern Asian pastoralist groups , but is rare or absent elsewhere in the world [1]–[4] . In Europeans LP is strongly associated with a single C to T transition in the MCM6 gene ( −13 , 910*T ) , located 13 . 91 kb upstream from the lactase gene [5] . Furthermore , in vitro studies have indicated that the −13 , 910*T allele can directly affect LCT gene promoter activity [6] . The −13 , 910*T allele ranges frequency from 6%–36% in eastern and southern Europe , 56%–67% in Central and western Europe , to 73%–95% in the British Isles and Scandinavia [7] , [8] while LP ranges in frequency from 15%–54% in eastern and southern Europe , 62%–86% in Central and western Europe , to 89%–96% in the British Isles and Scandinavia [9] . This makes the −13 , 910*T allele a good candidate for predicting LP in Europe . However , genotype/phenotype frequency comparisons have shown that the −13 , 910*T allele cannot account for LP frequencies in most African [3] and Middle Eastern populations [10] . Instead , different LP-associated alleles occurring in the same genomic region have been reported , indicating convergent evolution [2] , [4] , [10] , [11] . Using long-range haplotype conservation [8] and variation in closely linked microsatellites [12] as proxies for allelic age , the −13 , 910*T variant has been estimated to be between 2 , 188 and 20 , 650 years old and between 7 , 450 and 12 , 300 years old , respectively . These recent age estimates , when considered in conjunction with modern allele frequencies , indicate that −13 , 910*T has been subjected to very strong natural selection ( s = 0 . 014–0 . 19; [8] ) . It is interesting to note that similar estimates for the strength of selection have been obtained for one of the major African LP variants [4] . It is unlikely that lactase persistence would provide a selective advantage without a supply of fresh milk and this has lead to a gene-culture co-evolutionary model where lactase persistence is only favoured in cultures practicing dairying [13]–[16] , and dairying is more favoured in lactase persistent populations [14] , [17]–[19] . The reasons why LP , in conjunction with dairying , should confer such a strong selective advantage remain open to speculation . Flatz and Rotthauwe [20] proposed the calcium assimilation hypothesis , whereby a lactase persistence allele is favoured in high-latitude regions because reduced levels of sunlight do not allow sufficient synthesis of vitamin-D in the skin . Vitamin D is required for calcium absorption and milk provides a good dietary source of both nutrients . Additional factors are likely to include the ability to consume a calorie and protein-rich food source , the relative constancy in the supply of milk ( in contrast to the boom-and-bust of seasonal crops ) , and the value of fresh milk as a source of uncontaminated fluids . It is likely that the relative advantages conferred by these various factors differ in Europe and Africa . Estimates of the age of the −13 , 910*T correspond well with estimates of the onset of dairying in Europe . Slaughtering age profiles in sheep , goats and cattle suggest dairying was present in south-eastern Europe at the onset of the Neolithic [21] , [22] , while residual milk proteins preserved in ceramic vessels provide evidence for dairying in present day Romania and Hungary 7 , 900–7 , 450 years BP [23] . Furthermore , residual analyses of fats indicate dairying at the onset of the Neolithic in England , some 6 , 100 years BP [24] , [25] , and after to 8 , 500 BP in the western parts of present day Turkey [26] . Allelic age estimates are also consistent with the results of a recent ancient DNA study [27] which showed that the −13 , 910*T allele was rare or absent among early farmers from Central and Eastern Europe . These observations lend support to the view that −13 , 910*T , and thus LP , rose rapidly in frequency only after the onset of dairying , as opposed to the ‘reverse-cause’ hypothesis [14] , [17]–[19] , whereby dairying developed in response to the evolution of LP . Important questions remain regarding the location of the earliest −13 , 910*T-carrying dairying groups and the demographic and gene-culture co-evolutionary processes that shaped the modern distribution of LP in Europe . The present-day distribution of the −13 , 910*T allele might be taken to indicate an origin in Northwest Europe . However , the earliest archaeozoological and residual lipid and protein evidence for dairying is found in the Near East , in Southeast Europe and in Mediterranean Europe [21] , [26] , [28] . While these observations can seem contradictory , forward computer simulations have shown that the centre of distribution of an allele can be far removed from its location of origin when a population expands along a wave front [29] , [30] . Assuming that the −13 , 910*T-allele was only subjected to strong natural selection in dairying groups , it is likely that −13 , 910*T-carrying dairyers underwent demographic expansion to a greater extent than non-dairying groups . While gene flow between dairying and non-dairying groups would ultimately lead to genetic homogeneity , under conditions of limited gene flow between cultural groups , it is plausible that the earliest LP peoples would have made a higher contribution to the European gene pool than their non-LP neighbours . In this study we use demic forward computer simulations to examine potential scenarios for the spread of LP in Europe . We simulate three interacting cultural groups ( hunter gatherers , non-dairying farmers and dairying farmers ) and track the spread of an allele that is selected only in one group ( dairying farmers ) . We also track the expected proportion of genetic ancestry from the geographic region where LP/dairying coevolution began . We parameterize intrademic gene flow between cultural groups , interdemic gene flow , sporadic longer-distance migration , the cultural diffusion of subsistence practices and selection favouring lactase persistent dairyers . We compare the predicted frequency of a LP allele and arrival dates of farmers – from simulation outcomes – to known frequencies of the −13 , 910*T allele [3] , [8] and carbon-14 based estimates of the arrival dates of farmers [31] at different locations throughout Europe . We employ approximate Bayesian computation ( ABC ) , a set of methods that allow the estimation of parameters under models too complex for a full-likelihood approach [32] . By comparing summary statistics on the observed data with those computed on our simulated datasets , ABC enables us to estimate the key demographic and evolutionary parameters including the region where LP-dairying coevolution in began in Europe .
The simulation model we have employed here is relatively complex compared to related human demographic/evolutionary models reported [33]–[36] . The inclusion of a selected allele and three distinct but interbreeding cultural groups is necessary for the type of questions we are addressing . But the inclusion of four parameters related to sporadic migration activity , namely the proportion of individuals available for sporadic long-distance migration and the sporadic mobility of each of the 3 cultural groups ( modeled separately as a Gaussian random walk process ) both allows us to tackle the problem of migration overseas and adds , in our view , an extra level of realism to the model . However , as with any simulation model of population history , many simplifying assumptions have to be made and the extent to which these assumptions may lead to erroneous conclusions remains unknown . For example , we have not considered the ‘reverse-cause’ hypothesis [14] , [17]–[19] – which proposes that dairying first arose in populations that were already LP – because both ancient DNA evidence [27] and data from lipid residues on pots [26] are inconsistent with this view . However , this does not mean that once LP-dairying gene-culture coevolution was established , conversion to the culture of dairying was more likely in high LP frequency populations . Such a process is captured in our model to an extent , in that ‘cultural’ conversion is determined by the frequency of the receiving cultural group ( see equation 4 ) , and LP is unlikely to rise to high frequencies anywhere without the presence of dairying . Nonetheless , a more explicit treatment of this process may lead to different conclusions . Some parameters , such as those relating to the effects of climate zone/elevation , and the logistic growth rate , are fixed based on realistic assumptions [37]–[39] . For those parameters that are allowed to vary within a range we note that an important shortcoming is that in any single simulation their value is constant over the 360-generation duration of the run . This may be a particular issue for selection acting on an LP allele in Fd ( see below ) . Since we identify ‘good’ simulations using their fit to only two data sets ( arrival time of farming and LP allele frequency , both at a range of geographic locations ) it is unsurprising that our analysis is relatively uninformative for some parameters . However , inclusion of these parameters does serve to reflect uncertainty in their values . Estimates of the arrival dates for farming the 11 locations we consider here were calculated as local weighted averages of calibrated carbon-14 dates [31] from a Gaussian sampling region . We set the standard deviation of this region at the average nearest neighbour distance to ensure that most of the carbon-14 data was used . However , the geographic density of carbon-14 dates is highly uneven across Europe and so the number of such dates that are informative for farming arrival time at any of the 11 locations will vary . Also , there appears to be a considerable amount of noise in the dates for the first farmers . For example , the earliest carbon-14 date for farming in Ireland predates those for Great Britain , the Low Countries and Denmark . To test if these concerns had a major effect on our results we reanalysed our simulation date by setting the target farming arrival dates as those inferred by assuming a constant rate of spread of farming ( estimated at 0 . 9 km/year [31] ) and calculating the great circle distance from Anatolia to each sampling location . The results of this reanalysis were very similar to those presented above ( see Supplementary Figures S3 , S4 , S5 , and S6 ) , We are well aware that the spread of the Neolithic over Europe was not as constant as our model assumes . After the arrival of the Neolithic in the Balkans , there is a pause of approximately 800 years before it starts to spread to Central Europe , and there is another pause of 1 , 000 years before it spreads further into the northern German lowlands and other parts of the northern Europe . Clearly , the carbon-14 dates we have used to estimate the farming arrival times will not fully reflect the complex history of neolithisation in all parts of the continent . The list of parameters for which the marginal posterior distributions are notably narrower than their corresponding prior ranges ( selective advantage , intrademic gene flow , the sporadic migration distance of Fd and Fnd , and the geographic origin location of LP/dairying co-evolution ) – which we interpret as those parameters for which our analysis is informative – is an unsurprising one since we would expect these parameters to have the greatest influence on the spread of an LP allele and farming in Europe . Likewise , it is unsurprising that the proportion available for sporadic migration and the sporadic mobility of ( a ) dairying farmers , and ( b ) non-dairying farmers are both strongly negatively correlated ( Figure 2A and 2B ) since we would expect these parameters to be confounded in influencing the arrival time of farmers at different locations . The estimated selective advantage conferred by a LP allele ( mode = 0 . 0953; 95% CI = 0 . 0518–0 . 159 ) is in good agreement with previous estimates for Europeans ( 0 . 014–0 . 15 [8] ) . However , it should be noted that ( 1 ) this estimate is for selection only in dairying farmers , who make up just under half of the population that we simulate , and ( 2 ) we assume that selection is constant over time . It is possible that selection favouring LP has in fact been episodic and possibly spatially structured in different climate zones [20] , [40]–[44] . Episodic selection would be difficult to model without additional information on when those episodes were likely to have occurred . But we reason that constant selection strength is a more parsimonious assumption in the absence of evidence to the contrary . If , as modelled here , dairying farmers made up less than half of the European post-Neolithic population then we would expect the real continent-wide selection values for LP to average less that half of what we estimate here . Such a range of selection values are , however , still consistent with previous estimates based on haplotype decay [8] . Perhaps the most interesting result presented here is our estimation of the geographic and temporal origins of LP-dairying co-evolution . We find the highest posterior probabilities for a region between the central Balkans and central Europe ( see Figure 3 ) . At first sight such a location of origin may seem counter intuitive since it is far-removed from Northwest Europe , where the −13 , 910*T allele is found at highest frequency . However , previous simulations have shown that the geographic centroid of allele can be offset from its location of origin , particularly when it occurs on the wave front of a demographic expansion [29] , [30] . The lactase-dairying coevolution origin region inferred here is consistent with a number of archaeologically attested patterns concerning the emergence and spread of dairying . Recent carbon isotope ratios from lipids extracted from archaeological sherds show the presence of milk fats in present-day western Turkey and connect these findings to an increased importance of cattle herding [26] , [45]–[48] . In general , the spread of the Neolithic lifestyle from the Aegean to Central Europe goes hand in hand with the decline of the importance of sheep and goat and the rise in frequency of cattle bones in archaeological assemblages . While the Balkans at the beginning of the Neolithic still shows a variety of subsistence strategies [49] , the middle Neolithic in SE-Europe and the earliest Neolithic in Central Europe after 7 , 500 BP show a clear preponderance of cattle . Benecke [50] gives the following averaged rates for the respective domestic species: cattle 55 . 2% , sheep and goat 32 . 6% , pig 12% . The proportion of cattle in Central Europe increases during the following centuries to an average of 73% and then stays ( with a few exceptions ) stable for most prehistoric periods of Middle and northern Europe . Thereby , cattle herding is in most cases connected with kill-of profiles indicative for dairying [22] , [50]–[55] . Milk consumption and dairying have been proposed to be as early as the Pre-Pottery Neolithic B of the Near East and may even be a reason for domestication [56] , [57] . Without doubt , it was a common cultural practice during all phases and regions of the European Neolithic , especially for goat and cattle . However , a fully developed dairying-based farming economy emerges first during the late Neolithic in Southeast Europe and the Middle Neolithic Cultures following the Linearbandkeramik ( LBK ) in Central Europe , and is connected mainly to cattle and partly also to goat ( for the Rössen culture see [50] , [55] ) . In the Mediterranean , milking of cattle occurs episodically [28] and sheep and goat remain the dominant domestics , as they were earlier in Anatolia and the Aegean . It is very likely that the goat and sheep , and to a lesser extent cattle , based economies of the Mediterranean used processed milk in the form of yoghurt , cheese and other milk-derived products instead of fresh milk . The nutritional and agricultural differences between southern Europe , the Mediterranean and central and northern Europe , as well as historic reports , point to this . For instance , the Romans used goat and sheep milk for the production of cheese , and cattle as a draught animal . In contrast the Germanic peoples and other inhabitants of central and northern Europe practised cattle dairying and drank fresh milk in significant amounts . Strabo reports in his Geography [58]: “Their [sc . “the men of Britain”] habits are in part like those of the Celti , but in part more simple and barbaric - so much so that , on account of their inexperience , some of them , although well supplied with milk , make no cheese; and they have no experience in gardening or other agricultural pursuits . ” Overall , by considering the results from our simulations and archaeological , archaeozoological , and archaeometric findings , it seems very plausible to connect the geographic origin of the spread of LP to the increasing emergence of a cattle-based dairying economy during the 6th millennium BC . The geographic region of origin of the LBK – in modern day Northwest Hungary and Southwest Slovakia [59] , [60] – certainly correlates well with our results ( see Supplementary Figure S7 ) . The date of origin of LP-dairying coevolution estimated here ( mode = 7 , 441 years BP; 95% CI = 6 , 256 to 8 , 683 years BP; see Figure 4A and Supplementary Table S2 ) also fits well with dates for the early LBK in Central Europe ( ∼7 , 500 years BP ) and its proposed main predecessor , the Starčevo culture of the northern Balkan Peninsula and south of Lake Balaton ( 8 , 100 to 7 , 500 years BP; [61] ) . However , as explained above , our date estimate is conditioned by farming arrival dates in the estimated LP-dairying coevolution origin region . As a result , our date and location estimates are not independently derived . Nonetheless , a role for LP-dairying coevolution in the later rapid spread of LBK culture – from its origins in the Carpathian Basin – into central and Northwest Europe would be consistent with the significantly higher sporadic migration distances we infer for of Fd when compared to Fnd . This is also consistent with the rapid dissemination of the LBK culture over a territory of 2 , 000 km width and approximately one million square kilometres within less than 500 years [62] . Contrary to our expectations , we did not find that the presence of a positively selected LP allele in early dairying groups increases the unlinked genetic contribution of people living in the region where LP-dairying coevolution started to the modern European gene pool , when using demographic parameter values estimated here . The main reason for this is likely to be the relatively high inferred rates of intra- and interdemic gene flow between dairying and non-dairying farmers and between neighbouring demes , respectively , leading to a rapid erosion of any demographic ‘hitchhiking’ of unlinked genomic regions . Additionally , we only track the genetic contribution of people living in and around the deme of LP/dairying coevolution from the inception of this process . Since it takes some time for the LP allele to rise to appreciable frequencies , any demographic ‘hitchhiking’ effect may become important only after the allele centroid has moved some distance away from its origin deme . Another notable result was obtained when we compared the range of expected 13 , 910*T allele frequencies at different European locations – from simulations accepted at the 0 . 5% tolerance level – to those observed . While all observed values were within the 95% equal tail probability interval of the simulated values , many were somewhat offset from the modes . We interpret this as indicating that our model does not fully explain the distribution of the 13 , 910*T allele in Europe . One possible explanation for this is that migration activity – as modeled here by interdemic gene flow and sporadic unidirectional migration – has increased subsequent to the expansion of farming into the northwestern reaches of Europe . In this scenario the farming expansion phase , occurring 9 , 000 to 5 , 500 years BP , would be mainly responsible for generating the 13 , 910*T allele frequency cline in Europe but higher migration activity following this period would then have a homogenizing effect in LP allele frequencies . Intriguingly , a general pattern can be seen ( Supplementary Figures 1 ) whereby observed frequencies are lower than expected in northern Europe and higher than expected in southern Europe . Such a pattern is the opposite of what we would expect if selection for LP was higher in northern latitudes through a greater requirement for dietary vitamin D and calcium because low-sunlight conditions reduce UV-mediated vitamin D production in the skin [20] . This frequently cited mechanism [9] , [41] , [42] , [44] , [63]–[65] was not included in our model and thus would seem to have negative explanatory power . Thus our simulations indicate that geographically and temporally homogeneous selection in combination with well-attested underlying demographic processes are sufficient to explain , indeed , to over-explain , the LP/latitude correlation in Europe . However , it should be noted that since we have not explicitly included a parameterised latitudinal effect on selection in our model , there may be scenarios where such an effect could also explain patterns of LP in Europe . As inferred here , the spread of a LP allele in Europe was shaped not only by selection but also by underlying demographic processes; in this case the spread of farmers from the Balkans into the rest of Europe . We propose that this combination of factors could also explain the apparent homogeneity of LP-associated mutations in Europe . In Africa there are at least four known LP-associated alleles , including three that are likely to be of African origin [2] , [4] as well as −13 , 910*T , which is likely to be of European origin [3] , [12] . The greater apparent diversity of LP-associated mutations in Africa may reflect a greater genetic diversity in general , leading to the availability of more mutations upon which selection can act following the advent of dairying . However , we suggest that this diversity is the result of an ‘imposition’ of dairying culture on a pre-existing farming people , rather than the spread of dairying being tied to the spread of dairyers . Such a model would require the availability of a number of , albeit low-frequency , LP-causing mutations; either through a high mutation rate or a large number of potential LP-causing sites . It is therefore possible that , in the absence of the spread of dairying being linked to a major demographic expansion , high LP-allele diversity will also be found in the Indian subcontinent . We accept that the model we have used does not accommodate all data ( both genetic and archaeological ) that is potentially informative on the coevolution of LP and dairying in Europe . Future improvements can be made by adding more ‘realism’ to the model and by increasing the number of data types that are used in the ABC analysis , leading to more integrative inference . The former should include both adding more fixed parameter information ( such as the effects of past vegetation , climate variation and other geographic features on migration parameters and carrying capacities [66]–[68] ) and estimating currently fixed parameters such as the ratio of dairying to non-dairying farmers . The latter could be achieved by writing the simulation model so that it generates expectations for other data types . For example , including the movement of domestic cattle could be used to generate expectations on patterns of ancient and modern cattle genetic diversity , for which considerable data is available [69]–[73] . Finally , it should be possible to extend the approach we have used to study the evolution of LP and dairying in other parts of the world . We infer that the coevolution of European LP and dairying originated in a region between central Europe and the northern Balkans around 6 , 256 to 8 , 683 years BP . We propose the following scenario: after the arrival of the Neolithic in south-eastern Europe and the increasing importance of cattle herding and dairying , natural selection started to act on a few LP individuals of the early Neolithic cultures of the northern Balkans . After the initial slow increase of LP frequency in those populations and the onset of the Central European LBK culture around 7 , 500 BP , LP frequencies rose more rapidly in a gene-culture co-evolutionary process and on the wave front of a demographic expansion ( see Supplementary Videos S1 , S2 and S3 ) , leading to the establishment of highly developed cattle- ( and partly also goat- ) based dairying economies during the Middle Neolithic of central Europe around 6 , 500 BP . A latitudinal effect on selection for LP , through an increased requirement for dietary vitamin D [20] , is unnecessary to explain the high frequencies found in northern Europe .
Our simulation approach is motivated by a previous demic computer simulation study [33] and has features in common with more recent applications of this approach [34]–[36] . Geographic space is modelled as a series of rectangular demes arranged to approximate the European landmass ( 2375 land demes and 1511 sea demes ) . Each deme has attributes of elevation , area ( which varies due to the curvature of Earth and is calculated accordingly for each individual deme ) , and a climate ( Mediterranean , Temperate , or Cold/Desert – see Supplementary Figures S8 and S9 ) . A maximum total population size is specified for each land deme taking into account its area , and assuming that lower elevation and mild Mediterranean climate results in a greater potential population size , while harsher conditions , such as high elevations and cold/desert climates , result in a smaller potential population size [38] . The ratio for the relative contribution coefficients of climate and elevation factors to the population size is fixed at 1∶4 in this study; meaning that elevation has a more dramatic effect than climate on population size . The maximum deme population size ( carrying capacity , Kdeme , Supplementary Figure S10 ) is calculated by: ( 1 ) where cl and el are the climatic and relative elevation factors , respectively; cl having values of 1 for Mediterranean , 2/3 for Temperate , and 1/3 for cold/desert climates [38] ( see Supplementary Figure S9 ) , and el being calculated as: ( 1 . 1 ) So el ranges between 0 at the highest elevation and 1 at sea level ( see Supplementary Figure S8 ) . Dmax is the maximum population density and is fixed at 5 individuals per km2 ( i . e . in a sea level Mediterranean climate deme [39] ) , and Ademe is the area of the deme in km2 . Each deme contains three distinct cultural groups: non-dairying farmers ( Fnd ) , dairying farmers ( Fd ) , and hunter-gatherers ( HG ) . The ratios of ceiling population size for Fnd , Fd , and HG ( as a proportion of the total maximum population size for the deme , Kdeme ) are 50∶50∶1 respectively [37] , [39] . Each cultural group in each deme is assigned a frequency for an allele that is subjected to genetic drift ( modelled by intergenerational binomial sampling ) and an allele at an unlinked locus that is not ( as explained below ) . Initially the frequency of both ‘alleles’ is set at zero . The former represents a LP allele and is subject to selection of intensity s , only in the Fd group . The latter , here termed the GB ( genetic background ) ‘allele’ , is used to track the general genetic ancestry component from the region where the LP allele is first found among dairying farmers . It will be used to infer the expected proportion of genes that originate from this region . The two alleles are assumed to be unlinked and are modelled separately . We treat s as an unknown but bounded parameter , and choose random values ranging from 0 to 0 . 2 in simulations [8] . The LP and GB ‘allele’ frequency dynamics are determined in each generation by five processes: ( 1 ) intrademic bidirectional geneflow between cultural groups; ( 2 ) bidirectional geneflow between demes ( interdemic ) within the same cultural groups; ( 3 ) sporadic unidirectional migration within the same cultural groups; ( 4 ) cultural diffusion ( CD ) ; and ( 5 ) selection operating on LP allele-carrying individuals within the Fd group . Hardy-Weinberg equilibrium within each cultural group within each deme is assumed . Population size increase for each cultural group in each deme is modelled by logistic growth , limited by the carrying capacity of each group within each deme . We fixed the growth rate to r = 1 . 3 per generation , a value estimated from data of world population growth rate over the last 10 , 000 years , excluding the post-Industrial Revolution population boom ( US Census Bureau: www . census . gov ) . In addition , the Fd group is allowed to increase in size as a function of the selective advantage of the LP allele , s , by considering the number of LP individuals and the selective advantage to being a LP dairyer ( see equation 7 ) . We define intrademic bidirectional geneflow as the exchange of individuals between different cultural groups within a deme ( see Supplementary Figure S11 ) . A proportion of individuals in each cultural group , Pc , are deemed ‘available to change group’ . The actual number of individuals that are exchanged between cultural groups i and j , Bi↔j , is determined as follows: ( 2 ) Where Ni and Nj are the total number of individuals belonging to each cultural group . We treat Pc as an unknown but bounded parameter , and choose random values ranging from 0 to 0 . 2 in simulations [74] , [75] . We define interdemic bidirectional geneflow as the exchange of individuals between the same cultural groups in neighbouring demes ( see Supplementary Figure S11 ) . A proportion of individuals in each cultural group , Pd , are deemed ‘available to change deme’ . The actual number exchanged is determined using the same formula as for intrademic bidirectional geneflow ( equation 2 ) , except we substitute Pd for Pc , and Ni and Nj are the total number of individuals belonging to each cultural group in each neighbouring deme . In each generation , each cultural group in each deme undergoes bidirectional geneflow with one neighbouring deme , randomly chosen from the 8 possible . We define sporadic unidirectional migration as the movement of some individuals in a particular cultural group and deme to the same cultural group in a different deme ( see Supplementary Figure S11 ) . A proportion of individuals in each cultural group , Ps , are deemed ‘available to migrate’ . The actual number of individuals that migrate , Nmig , is dependent on the ‘pressure’ to leave the current deme and the availability of unoccupied carrying capacity in the destination deme ( ‘attractiveness’ ) , and is determined as follows: ( 3 ) Where Kcurr and Kdest are the carrying capacities , and Ncurr and Ndest are the number of people in the cultural group , in the current home and destination demes respectively . We treat Ps as an unknown but bounded parameter , and choose random values ranging from 0 to 0 . 2 in simulations . The destination deme is chosen by a Gaussian random-walk process , which takes into account the mobility of the cultural group and the topography of the home deme . The Gaussian distribution is centred on the home deme; and its standard deviation is the product of the mobility of the cultural group , Mi , and the relative mobility factor of the home deme , Mcurr . We treat Mi as a separate unknown but bounded parameter for each of the three cultural groups , and choose random values ranging from 0 to 3 ( demes ) in simulations . Mcurr is determined for each deme by its elevation , allowing greater mobility at lower elevations [76] , [77] , with fixed values of 0 . 5 ( demes ) at mountainous terrain ( above 1100 meters ) , 1 . 0 at lowlands ( below 1100 meters ) , and 1 . 5 at coastal demes . The sporadic unidirectional migration function allows movement overseas , but whenever a sea deme is identified as a non-realistic destination deme the nearest neighbouring coastal deme is chosen instead . This feature , together with the attractiveness of low elevation land and the higher Mcurr value for coastal demes , creates the realistic tendency of a faster spread of farming along coastlines , consistent with archaeological data [78] . We define Cultural Diffusion ( CD ) as the spread of culture and technology by learning through exposure rather than by migration ( see Supplementary Figure S11 ) . In our simulations a proportion of individuals in each cultural group , Pdif , are deemed ‘available to convert’ from one cultural group to another . The number of individual that convert from cultural group i to cultural group j , Ni→j , is determined by this parameter and the proportion of the carrying capacity ( K ) of the home deme ( deme 0 ) and in the 8 neighbouring demes ( demes 1 to 8 ) that is taken up by cultural group j , as follows: ( 4 ) where b is the relative influences of the home deme and the 8 neighbouring demes ( fixed to 0 . 75 ) . We treat Pdif as an unknown but bounded parameter , and choose a random value ranging from 0 to 0 . 2 in each simulation . That value is then applied to ‘conversions’ between all 3 cultural groups . The geographic location where LP/dairying gene-culture coevolution starts is chosen at random from all land demes . This LP mutation is initialized at a frequency of 0 . 1 in Fd when their population size reaches a critical size in the chosen start deme , set to a minimum of 20 individuals per deme in simulations . While we would expect any de novo mutation to always have an initial frequency of 1/2N , we also expect that it will have a high probability of extinction unless selection is very strong [79] . Indeed , in preliminary simulations this was observed ( data not shown ) . Thus , for computational efficiency we condition on the LP mutation having already reached a frequency of 0 . 1 in Fd in the deme of origin . However , such a starting frequency means that little more than four LP alleles are initialized in simulations . Selection acting on the LP allele , p , increases its frequency in Fd only , as follows [80]: ( 5 ) where s is the selection coefficient for p , and p′ is the new LP allele frequency . In addition , selection acting on the LP allele increases the number , N , in Fd as follows: ( 6 ) where N′ is the new number of Fd in a particular deme . All simulations were run for 360 generations which , assuming a generation time of 25 years [81] , [82] , corresponds to the 9 , 000-year history of farming in Europe . We performed 200 , 000 simulations in total . The genetic contribution of the population living in the region of origin of LP/dairying gene-culture coevolution to the overall European population is tracked over generations by calculating the GB ‘allele’ frequency over all demes in all 3 cultural groups . In the generation when the LP allele is initialized , all cultural groups in the origin deme and 8 neighbouring demes are assigned the unlinked GB ‘allele’ at a frequency of 1 . The GB ‘allele’ is subjected to the same intra- and inter-deme geneflow and migration processes as described above , but is not subject to drift , as modelled by binomial sampling , or to selection . At the end of each simulation this GB allele is taken to represent the general genetic contribution of the population living in the region of origin of LP to the modern European population . The ancestry component of Europeans , at any generation , that originates from people living in the region of origin of the LP allele ( FGB ) is calculated as follows: ( 7 ) where n is the number of land demes , Ni is the total number of people in deme i , and pGBij and Nij are the frequency of the GB ‘allele’ and the population size in deme i/cultural group j , respectively . To estimate parameters of interest we use an ABC approach , following [32] . By comparing summary statistics computed on each simulated dataset to those from the observed data , we are able to accept only those simulations with summary statistics sufficiently close to the target ( i . e . the observed summary statistics ) and reject the remainder . We then perform a weighted local-linear regression on these retained parameter sets , with weight determined by the “distance” between the simulation summary statistics and the target ( all details below ) . This generates approximate marginal posterior probability distributions for each parameter of interest , from which we derive our modal point estimates . Our chosen summary statistics , U , are the frequencies of the −13 , 910*T allele at 12 different sample locations around Europe , the Near East and western Asia [7] , [8] . In addition we include as summary statistics the times to arrival of farming at 11 of the same locations ( the Anatolia location is excluded as the simulation model is initialized with this as the origin of the spread of farming into Europe ) . We recognize that these are not summary statistics sensu stricto but are parameters in the model for which we have independent estimates . However , the simulations , being stochastic , generate a distribution of arrival times , and we need to condition on those that are consistent with the known archaeological evidence . The most straightforward way to do this is to place a point prior on the arrival dates , and then condition on these using the ABC machinery , as if they are summary statistics . The point priors for the arrival dates of farming at 11 of the 12 sampling locations considered ( Anatolia was set to 9 , 000 years as the simulations begin 360 generations ago in ‘an Anatolia’ populated by farmers ) were calculated as follows: ( 1 ) The average nearest-neighbour distance ( ANND ) between each sampling location was calculated ( 557 . 13 km ) . ( 2 ) A 2-D Gaussian sampling region was constructed around each of the 11 sampling locations , of standard deviation = ANND/1 . 96 ( this ensures that 95% of each Gaussian sampling region will be within the ANND ) . ( 3 ) A weighted average of all dates within 3 standard deviations of the sampling location was calculated using all calibrated carbon-14 earliest farming arrival dates from Pinhasi et al . [31] , and weighting using the distance from the sampling location and the standard probability density function for a Gaussian distribution . Assuming a generation time of 25 years [81] , [82] these observed dates are converted to generations from the start of the simulation , which was set at 9 , 000 years BP or 360 generations ago ( see Table 1 ) . We also include two Spearman's rank-order correlation coefficients , calculated separately for the 12 T-allele frequencies and the 11 times to arrival of farming , giving a total of 25 summary statistics . When calculating these statistics for the simulated data: we take LP frequencies in the final generation of the simulation at the 12 corresponding geographic locations; and the time to arrival of farming is defined as the simulation generation at which either Fd or Fnd reach 1% of their carrying capacity within each of the 11 corresponding location demes . All time to arrival of farming statistics are scaled to the interval [0 , 1] by dividing by the total number of simulated generations ( 360 ) . Parameters of interest , φ , are: the east-west and north-south coordinates of the location where the LP-allele first undergoes selection among Fd; the generation at which this selection starts; the selective advantage of LP within the Fd group , s; the proportion available for interdemic bidirectional geneflow , Pd; the proportion available for intrademic bidirectional geneflow among cultural groups , Pc; the rate of cultural diffusion , Pdif; the proportion of people available for sporadic migration , Ps; the mobility of each of the three cultural groups , Mi; and the contribution of people living in the deme where LP-dairying gene-culture coevolution began and its 8 surrounding demes , FGB , to the modern European gene-pool . The uniform prior distributions for each parameter are given in Supplementary Tables S2 and S3 . Our full ABC algorithm is as follows: ( 1 ) choose the summary statistics U as outlined above and calculate their values , u , for the observed data ( these are given in Table 1 ) , ( 2 ) choose a tolerance level δ ( as suggested we pre-define a proportion of the best fitting simulations , Pδ , to accept and from this calculate an implicit tolerance level δ ) , ( 3 ) sample a parameter set φi from the pre-determined prior distribution of φ , ( 4 ) simulate forward under our model using parameter set φi , ( 5 ) in the final generation of our simulation we calculate the summary statistics , ui , for this simulated data , ( 6 ) If ∥ui−u∥≤δ ( where ∥ . ∥ is the Euclidean norm between the two vectors ) we accept parameter set φi , ( 7 ) steps 3 to 6 are repeated until we have a sufficient number of retained parameter sets , ( 8 ) A local-linear standard multiple regression is then performed to adjust the φi , with each φi weighted according to the size of ∥ui−u∥ using the Epanechnikov kernel function Kδ ( t ) ( see [32] for details ) , ( 9 ) The resulting fitted parameter sets φi* form a random sample from the approximate joint posterior distribution P ( φ|U = u ) . All retained parameters – except for the two coordinate values and the generation at which the co-evolutionary process starts – were log transformed prior to the regression step , and subsequently back-transformed to produce the fitted parameter sets φi* , as suggested by Beaumont et al . [32] . The simulation and ABC analysis procedures were written in the Python Programming Language ( URL: http://www . python . org/ ) employing the numarray and Numpy array handling libraries . Maps and animations were generated using the Python library PyNGL . Post-ABC analysis data was processed and visualised using the statistical package ‘R’ ( URL: http://www . R-project . org/ ) . | Most adults worldwide do not produce the enzyme lactase and so are unable to digest the milk sugar lactose . However , most people in Europe and many from other populations continue to produce lactase throughout their life ( lactase persistence ) . In Europe , a single genetic variant , −13 , 910*T , is strongly associated with lactase persistence and appears to have been favoured by natural selection in the last 10 , 000 years . Since adult consumption of fresh milk was only possible after the domestication of animals , it is likely that lactase persistence coevolved with the cultural practice of dairying , although it is not known when lactase persistence first arose in Europe or what factors drove its rapid spread . To address these questions , we have developed a simulation model of the spread of lactase persistence , dairying , and farmers in Europe , and have integrated genetic and archaeological data using newly developed statistical approaches . We infer that lactase persistence/dairying coevolution began around 7 , 500 years ago between the central Balkans and central Europe , probably among people of the Linearbandkeramik culture . We also find that lactase persistence was not more favoured in northern latitudes through an increased requirement for dietary vitamin D . Our results illustrate the possibility of integrating genetic and archaeological data to address important questions on human evolution . | [
"Abstract",
"Introduction",
"Discussion",
"Material",
"and",
"Methods"
]
| [
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics",
"evolutionary",
"biology/human",
"evolution",
"nutrition",
"computational",
"biology/molecular",
"genetics",
"computational",
"biology/evolutionary",
"modeling"
]
| 2009 | The Origins of Lactase Persistence in Europe |
Cohesin is crucial for genome stability , cell division , transcription and chromatin organization . Its functions critically depend on NIPBL , the cohesin-loader protein that is found to be mutated in >60% of the cases of Cornelia de Lange syndrome ( CdLS ) . Other mutations are described in the cohesin subunits SMC1A , RAD21 , SMC3 and the HDAC8 protein . In 25–30% of CdLS cases no mutation in the known CdLS genes is detected . Until now , functional elements in the noncoding genome were not characterized in the molecular etiology of CdLS and therefore are excluded from mutation screening , although the impact of such mutations has now been recognized for a wide range of diseases . We have identified different elements of the noncoding genome involved in regulation of the NIPBL gene . NIPBL-AS1 is a long non-coding RNA transcribed upstream and antisense to NIPBL . By knockdown and transcription blocking experiments , we could show that not the NIPBL-AS1 gene product , but its actual transcription is important to regulate NIPBL expression levels . This reveals a possibility to boost the transcriptional activity of the NIPBL gene by interfering with the NIPBL-AS1 lncRNA . Further , we have identified a novel distal enhancer regulating both NIPBL and NIPBL-AS1 . Deletion of the enhancer using CRISPR genome editing in HEK293T cells reduces expression of NIPBL , NIPBL-AS1 as well as genes found to be dysregulated in CdLS .
Genetic information needs to be inherited without any changes over numerous cell generations and from parents to offspring . This process crucially depends on the cohesin complex that ensures genome stability during cell divisions , DNA damage repair and is involved in the three-dimensional organisation of the chromatin fibre in the cell nucleus [1 , 2 , 3] . The association of cohesin with DNA critically depends on the cohesin-loading factor NIPBL [4 , 5 , 6] . However , NIPBL is also involved in gene regulation , independent on its role for cohesin [7 , 8] . NIPBL is the gene that is most frequently ( >60% of cases , OMIM 122470 ) found to be mutated in the human developmental disorder Cornelia de Lange syndrome ( CdLS , 1 of 10 , 000–30 , 000 live births ) [9 , 10] [11 , 12] . This syndrome is characterized by craniofacial anomalies , upper limb malformations , growth and mental retardation , hirsutism , and other system abnormalities [13 , 14] . Mutations in the cohesin subunits SMC1A , SMC3 , RAD21 and the cohesin regulator HDAC8 [15 , 16 , 17 , 18] account for ~10% of the more moderately affected patients . The genetic causes of 20–25% of CdLS patients are still unknown . The actual NIPBL expression levels seem to be critical for developmental processes in human and mouse . In a cohort of severely affected CdLS patients only ~65% of NIPBL expression is observed while mildly affected cases showed ~75% expression; one case report describes a mild CdLS phenotype with as little as 15% reduction of the NIPBL transcript [19 , 20 , 21] . In Nipbl heterozygous knockout mice with a CdLS reminiscent phenotype , the levels of Nipbl expression are reduced by only 25–30% , suggesting a compensation by the intact allele [22] . Further , high NIPBL levels have been found to confer poor prognosis in non-small cell lung cancer [23] . In this study we aimed at gaining insight into the regulation of the NIPBL gene by identifying regulatory elements in the noncoding genome . Different studies have shown that a large number of potentially disease-causing mutations/variants in the genome are located outside coding regions in regulatory non-coding areas , highlighting the importance of the identification of these elements for the study and diagnosis of human genetic diseases [24] [25] [26] . The potential relevance of such regions for CdLS is illustrated by mutations in the 5’ untranslated region of NIPBL ( NIPBL:c . -316_-315delinsA; [19] and NIPBL:c . -94C>T; [10] ) that affect NIPBL expression levels and are disease causing without affecting the NIPBL protein sequence . Gene regulatory elements controlling CdLS genes have not yet been identified and analysed for mutations at all . During the course of our project , a study aiming at analysing non-coding RNAs overlapping with known autism-related genes identified a NIPBL promoter-associated antisense transcript of 5 . 3 kb , named NIPBL-AS1 , in brain tissues from patients affected by Autism Spectrum Disorders ( ASD ) [27] . Since NIPBL-AS1 and NIPBL showed a concordant expression in their hands as well as ours , we hypothesized that NIPBL-AS1 and NIPBL might be interconnected and that this could represent a gene regulatory mechanism . Long non-coding RNAs ( lncRNA ) were shown to control genes located in their vicinity in cis but also in distant domains in trans ( reviewed in [28] ) . These processes can be mediated by the transcription of the lncRNA per se , which affects nucleosome positioning or histone modifications at gene promoters [29 , 30 , 31 , 32] or creates a permissive chromatin environment [33] . Alternatively , the lncRNA transcript itself functions as a scaffold to mediate silencing or activation of a target gene by binding to chromatin modifiers , transcription factors or proteins that are part of the transcription preinitiation complex [34 , 35 , 36 , 37 , 38 , 39 , 40] . Thus , we aimed at understanding whether and how the lncRNA NIPBL-AS1 influences NIPBL transcription . Other crucial gene regulatory elements are enhancers . In mammals , they have been found up to several hundred kilobases from the target transcription start site , as for instance the Sonic Hedgehog ( SHH ) enhancer that is located 1MB upstream of the gene [41] . Enhancers are “open” chromatin regions marked by H3 . 3/H2A . Z histone variants and enriched for histone modifications such as mono- and di-methylated lysine 4 of histone H3 ( H3K4me1/H3K4me2 ) and acetylated lysine 27 of histone H3 ( H3K27ac ) ( reviewed in [42] ) . Enhancers for the NIPBL gene were so far unknown . Here we identified an enhancer that stimulates expression of NIPBL as well as the lncRNA NIPBL-AS1 .
NIPBL-AS1 is located upstream of NIPBL in a head-to-head orientation and encodes for a 5 . 3 kb lncRNA ( Fig 1A ) . Considering the emerging roles of lncRNA in genome regulation , we asked whether the NIPBL-AS1 transcript acts in NIPBL expression regulation . Since long noncoding RNAs can function in the nucleus ( MALAT1 , XIST ) or in the cytoplasm ( linc-MD1 , NORAD ) [43] we addressed the localization of NIPBL-AS1 by fractionation of HEK293T cells into cytoplasmic and nucleoplasmic RNA . The RT-qPCR analysis of the cell fractions showed that NIPBL-AS1 is not retained in the cell nucleus as much as lncRNAs with nuclear functions like MALAT1 or XIST [43] . However a fraction of NIPBL-AS1 is still present in the nucleus ( S1 Fig ) . To test whether NIPBL-AS1 acts directly on NIPBL expression , we depleted NIPBL-AS1 in HEK293T cells using Antisense Oligonucleotides ( ASO ) , which activate the RNAseH pathway by forming DNA/RNA hybrids [44] , leading to degradation of the targeted NIPBL-AS1 RNA . ASOs are in particular suited to knockdown nuclear lncRNA but are also efficient for cytoplasmic lncRNA , so both pools can be efficiently depleted [45 , 46 , 47 , 48] . We used two ASOs to target either the 5’ end or the 3’ end of the NIPBL-AS1 gene ( respectively ASO2 and ASO3 ) and one non-targeting control ( ASO C ) . The transcript levels of NIPBL-AS1 ( P132 ) and NIPBL ( P138; P42 ) were assessed by RT-PCR/qPCR at 48 hours after transfection . Two intron-spanning primer pairs in the middle ( P138 ) and at the 3’-end ( P42 ) of the NIPBL gene that cannot discriminate between the different reported NIPBL splice variants [49] were used ( Fig 1A ) . NIPBL-AS1 transcripts were significantly reduced by both ASOs ( Fig 1B ) , yielding depletion efficiencies for NIPBL-AS1 as high as 80% for ASO2 and 70% for ASO3 . NIPBL transcript levels were not affected when we used a primer pair positioned more in the front part of the NIPBL gene ( primer P138 ) . When we used a primer spanning two of the last exons ( P42 ) we observed a small but significant difference ( Fig 1C ) . To validate our results we repeated our experiment in a different cell line , HB2 cells—a model cell line for normal breast endothelium , that can be easily transfected ( S1B and S1C Fig ) . Here we obtained the same result for primer 138 but the change that we observed for primer 42 is not significant . If the lncRNA would affect NIPBL transcription directly we would expect a significant and robust change detectable by both primer pairs . Therefore we concluded that the RNA product originating from NIPBL-AS1 does not remarkably influence NIPBL expression levels . NIPBL and NIPBL-AS1 are transcribed from a shared bidirectional promoter that is GC rich and lacks TATA elements . The annotated transcription start sites ( TSS ) of both genes are separated by 77 bp ( Fig 2A ) . The promoter is characterized by a DNAse hypersensitive region flanked by mirrored H3K4me3 marks ( Fig 2A ) . Only little is known about the regulation of bidirectional promoters and about the relationship between the two transcribed genes . Since the NIPBL-AS1 transcript seems to have no relevant function for NIPBL regulation , we hypothesized that the actual transcription of NIPBL-AS1 is important for NIPBL expression . To investigate this , we used the CRISPR/Cas9 system to interfere with the transcription process of NIPBL-AS1 and NIPBL using different specific guide RNA ( CRISPRi ) ( Fig 2A and 2B ) . It has already been reported that the DNA-binding of the rather bulky complex of catalytically inactive Cas9 ( dCas9 ) and guide RNA ( gRNA ) is sufficiently stable to block the progression of RNA polymerase II , leading to silencing of the target gene [50] . We designed gRNA for three different dCAS9 blocks ( Block I- targeting NIPBL-AS1 , Block II—targeting NIPBL , Block III—placed in the NIPBL gene but not blocking ) as well as a control gRNA ( GC ) in distant gene desert on chromosome 5 . Since it was shown that gRNAs targeting the non-template DNA strand have a higher gene silencing effect compared to those targeting the template strand [50] , we recruited dCas9 to the 5’ end of NIPBL-AS1 using two guide RNAs ( Block I , G1 and G2 ) targeting the non-template strand of NIPBL-AS1 ( Block I , Fig 2A and 2B ) . To block transcription of NIPBL , we used three guide RNAs ( G3 , G4 and G5 –Block II ) targeting the non-template strand of the gene ( Block II , Fig 2A and 2B ) . To verify the specificity of the transcription block we used two gRNAs ( Block III , G6 and G7 ) designed complementary to the template strand of NIPBL ( Block III , Fig 2A and 2B ) since gRNAs targeting the template strand do not efficiently block gene transcription . The gRNAs for each block were mixed and co-transfected with a construct coding for catalytically inactive Cas9 ( dCas9 ) into HEK293T cells . First , we analysed the wider promoter region of NIPBL-AS1 and NIPBL by ChIP-qPCR ( for primer positions see S2A Fig ) for the coverage of the initiating Polymerase II , marked by phosphorylation of the C-terminal Ser5 residue ( RNA PolII Ser5 ) [51 , 52 , 53 , 54] , and the total RNA Polymerase II ( RNA Pol II ) . We observed an increased accumulation of RNA PolII Ser5 upstream of the guide RNA target sites of Block I compared to the control block ( Block III ) ( see S2B Fig , primer AS3 ) . Vice versa we observed in Block II a similar accumulation of RNA PolII Ser5 upstream of the guide RNA target sites ( S2D Fig , primer AS7 ) . For both blocks we also observed a slight increase in total RNA Pol II signals upstream of the guide RNA target sites ( S2C Fig , primer AS3 and S2E Fig , primer AS7 ) . RNA PolII and RNA PolII Ser5 signal increments tend to spread on the respective gene that was not blocked ( NIPBL for Block I and NIPBL-AS1 for Block II ) . We then analysed the transcript levels of NIPBL and NIPBL-AS1 under the different blocks . When we transfected the cells with Block I , only 30% of the NIPBL-AS1 transcript was still detectable compared to the control guide RNA transfection ( GC ) , indicating that NIPBL-AS1 transcription was significantly blocked ( see Block I , Fig 2C ) . NIPBL mRNA levels after blockage of NIPBL-AS1 transcription were increased ( Fig 2D ) . By using a primer pair detecting the unspliced NIPBL transcript ( Fig 2E ) we confirmed that the increase in NIPBL transcripts originates from de-novo transcription and not from increased stability of the mRNA . In cells transfected with Block II we could achieve a significant reduction of NIPBL transcription , with only 30% of the transcript still detectable ( Fig 2D ) . We observe now an upregulation of the NIPBL-AS1 transcript ( Fig 2C ) , similar to the observations for the NIPBL transcript in Block I . In cells transfected with Block III we did not observe a relevant reduction of NIPBL or NIPBL-AS1 transcription ( Fig 2C and 2D ) . In summary , we could achieve an efficient block of the transcription of NIPBL and NIPBL-AS1 ( Fig 2 ) . This is probably due to the accumulation of RNA PolII and RNA PolII Ser5 at the block sites ( see primer AS3 and AS7 in S2 Fig ) . The accumulation of PolII Ser5 is more evident than that of RNA PolII; this is consistent with the presence of a stalled RNA PolII , since RNA PolII Ser 5 is also a mark for stalling . [51 , 52 , 53 , 54] . This leads to a reduced expression of the gene involved in the blocking but also a significant upregulation of the corresponding gene that is not blocked . We have therefore demonstrated that the transcription of NIPBL-AS1 and NIPBL is indeed interconnected and that by reducing the transcription activity of NIPBL-AS1 we increased the transcription at the NIPBL gene and vice versa . We have also uncovered an interesting approach to eventually manipulate the expression of the NIPBL gene without interfering with the NIPBL gene itself . We aimed at identifying the distal regulatory elements of NIPBL and NIPBL-AS1 using chromosome conformation capturing ( 3C-seq , a derivative of 4C ) [55] , initially in the HB2 cell line but also in HEK293T cells since these cells are better suitable for CRISPR genome editing . Using BglII as 3C-seq restriction enzyme and a viewpoint located at the NIPBL promoter ( VP1 ) , we observed contacts to two intergenic regions , namely R1 and R2 , respectively at 130 kb and 160 kb upstream of the NIPBL promoter ( S3A Fig ) . We focused on these regions since the analysis of ChIP sequencing data for different histone marks and DNaseI hypersensitivity tracks for six cell lines ( GM1287 , K562 , HeLa-S3 , HMEC , HUVEC , HSMM ) , available from the ENCODE project [56] , revealed that the R1 region highly correlates with open chromatin ( DNAseI ) and histone variants/marks found at enhancers and active transcription ( H2A . Z , H3K4me1 , H3K4me2 , H3K4me3 ) [57] in different cell lines ( Fig 3C , S3C and S3D Fig ) . Region R2 correlates only in GM12878 cells with enhancer marks . Thus , we hypothesized that R1 is the best candidate for a NIPBL distal enhancer . The 3C-seq analysis also revealed contacts from the NIPBL promoter into the NIPBL gene body , covering nearly all of the 47 exons ( 188 kb ) , also confirmed by two additional viewpoints inside NIPBL ( VP2-3 ) ( S3A Fig ) . However , since these contacts do not involve regions with characteristic enhancer marks we omitted them from our search for NIPBL/NIPBL-AS1 enhancers . To confirm the observed long-range interactions of the NIPBL promoter , we performed a higher resolution 3C-seq ( using ApoI as restriction enzyme ) using one viewpoint at the NIPBL promoter ( VP4 ) and two viewpoints at R1 ( VP5 ) and R2 ( VP6 ) in HEK293T cells ( Fig 3A ) and HB2 cells ( S3B Fig ) . These experiments confirmed the contact between the NIPBL promoter and the distal intragenic region R1 ( Fig 3A , S3B Fig , blue tracks , VP4 ) . Vice versa , the viewpoint located at R1 ( VP5 ) showed contacts between this potential enhancer and the promoter of NIPBL and NIPBL-AS1 ( or 5’end of NIPBL-AS1 ) ( Fig 3A , S3B Fig , green tracks , VP5 ) . These interactions involved two CTCF binding sites at R1 and two within NIPBL-AS1 respectively ( see CTCF ChIP-sequencing peaks in HEK293 cells in Fig 3B or S3C Fig ) , that are positioned in the forward-reverse motif orientation that favours long-range interactions ( Fig 3B ) [58 , 59 , 60] . The interaction between R2 ( VP6 ) and the NIPBL promoter is no longer observed ( Fig 3A , S3B Fig , red tracks , VP6 ) . Therefore we consider from now on only R1 as candidate enhancer . Consistent with our observations , recently published high resolution Hi-C data showed contacts between the NIPBL promoter and the distal region R1 in seven different human cell lines ( GM12878 , HMEC , NHEK , KBM7 , HUVEC , K562 , IMR90 ) [59] ( S4A and S4G Fig ) . This indicates conservation of this long-range interaction between different cell types , eventually also between species since Hi-C data from mouse CH12 cells also showed long-range contacts of the NIPBL promoter to a potential distal enhancer next to the Slc1a3 gene ( S4H Fig ) . Since we identified R1 as potential distal regulatory element , we wanted to test whether this element displays activity typical for enhancers with respect to NIPBL and NIPBL-AS1 . For this we deleted R1 using CRISPR/Cas9 genome editing in HEK293T cells . The R1 region comprises two fragments , which we termed R1_1 and R1_2 , enriched in histone marks typical for enhancers and actively transcribed regions as well as CTCF sites ( CTCF#1 in R1_1 and CTCF#2 in R1_2 ) ( Fig 4A ) . To dissect the contribution of R1_1 and R1_2 in the regulation of NIPBL and NIPBL-AS1 we designed guide RNA that specifically delete either a region including R1_1 ( D1 , gRNA2 and gRNA3 ) or R1_1 and R1_2 at the same time ( D2 , gRNA1 and gRNA3 ) ( Fig 4A ) . In total we obtained four clones ( D1_89 , D1_38 , D1_42 , D1_63 ) for the smaller deletion of 5 kb ( D1 ) and five clones ( D2_35 , D2_18 , D2_25 , D2_33 , D2_103 ) for the 12 kb deletion ( D2 ) . The homozygous targeting was confirmed by PCR where we obtained an amplification only for primers designed to detect the deletion , but not for primers designed to detect the intact genomic region ( see S5D and S5H Fig ) . We assessed the effects of these two deletions on the expression of NIPBL . RT-qPCR analysis showed a reduction of NIPBL transcript levels to in average 60% for both depletions and transcription of NIPBL-AS1 was also reduced ( Fig 4B , S6 Fig ) . The reduction of NIPBL expression can also be detected at the protein level by western blotting ( Fig 4C ) . This supports that R1 acts as distal enhancer of NIPBL . We concluded that the actual enhancer localizes in the R1_1 fragment since the larger deletion including also R1_2 does not lead to more significant changes in NIPBL and NIPBL-AS1 transcript levels . In order to validate the enhancer activity of R1_1 and R1_2 in an independent experimental setup , we cloned the two fragments in plasmids carrying a luciferase gene under control of the NIPBL promoter ( Fig 4D ) . Two similar sized random DNA fragments were used as controls ( Fig 4D , cont1 , cont2 ) . The constructs were transfected into HEK293T cells and the luciferase activity analysed . In comparison to the control , DNA fragment R1_1 leads to a clear 2 . 5 fold increase of the luciferase activity but not R1_2 ( Fig 4E ) . Taken together , these results point to R1_1 as the enhancer for NIPBL and NIPBL-AS1 . To investigate how the deletions D1 ( including R1_1 ) and D2 ( including R1_1 and R1_2 ) affect the long-range interactions of the NIPBL promoter , we performed 3C-seq with one clone from D1 ( D1_C89 ) and one clone from D2 ( D2_C35 ) ( Fig 4F ) using the NIPBL promoter ( VP4 ) , R1 ( VP5 ) and R2 ( VP6 ) as viewpoints ( Fig 4F ) . The D1 deletion removed the CTCF#1 site ( S2A Fig and Fig 4D ) but the promoter remains in contact with the CTCF#2 site ( see Fig 4F , VP4 and VP5 panels for D1 ) . When both CTCF sites ( CTCF#1 and CTCF#2 ) are deleted ( D2 ) only very little interactions of the promoter region ( Fig 4F , VP4 panel for D2 ) into the area surrounding the deleted enhancer regions are detectable , as well as the contacts into the NIPBL gene body . We conclude that the NIPBL/NIPBL-AS1 enhancer overlaps with the R1_1 region but both CTCF sites in R1 ( CTCF#1 and CTCF#2 ) are required for the long-range interactions between R1 and NIPBL/NIPBL-AS1 . To demonstrate the importance of the R1_1 enhancer we asked whether the deletion of the enhancer affects genes that were found to be dysregulated in CdLS patients [20] and that we previously confirmed to be regulated by NIPBL [8] . RT-qPCR analyses revealed that all three analysed genes ( BBX , ZNF695 , GLCCI1 ) are downregulated in the larger deletion ( D2 ) and two genes in the smaller deletion ( D1 ) ( Fig 5A and S7 Fig ) . This further supports the relevance of the identified enhancer region for the regulation of NIPBL and also of the NIPBL downstream targets which are dysregulated in CdLS . NIPBL transcript levels were found to be reduced in CdLS patients with NIPBL mutations but the cause of this downregulation was unclear [8 , 20] . We were curious to check NIPBL and NIPBL-AS1 transcript levels in cells from CdLS patients with a heterozygous truncation mutation of NIPBL [20 , 21] . We assessed levels of NIPBL transcripts in three patient LCLs with early truncating mutations in NIPBL ( PT1-3 ) as well as in four control LCLs ( C1-4 ) ( described in [8 , 20] and S8A Fig ) . Consistent with previous reports [20 , 21] , the NIPBL mRNA levels in the patient cells were reduced to 60–70% ( Fig 5B and S8B Fig ) . Explanations for this imply either a downregulation of the NIPBL gene and/or a degradation of the NIPBL transcript by the nonsense-mediated mRNA degradation pathway , as has been previously hypothesized [61] . In the first case we might expect a misregulation of NIPBL-AS1 since our data show that the transcription of NIPBL and NIPBL-AS1 are interconnected . In the second case we would not observe an effect on NIPBL-AS1 . Assessing the NIPBL-AS1 in the patient cells showed that the transcript levels of NIPBL-AS1 were not significantly affected ( Fig 5B , S8B Fig ) . Next we analysed by pyrosequencing the fraction of the mRNA originating from wt and mutant alleles . Here we observed 38% ( PT1 ) or 24% ( PT3 ) of mutant transcripts ( S8C Fig ) which increase up to 46–48% when blocking nonsense-mediated mRNA decay by cycloheximide , indicating at least a partial degradation of mutant NIPBL transcripts in patient cells , while both alleles remain actively transcribed . This explains why NIPBL-AS1 transcription is not altered and supports our finding that the NIPBL-AS1 transcript is not involved in NIPBL transcription and vice versa ( Fig 1 ) .
NIPBL encodes for a protein that is required to load the cohesin complex onto DNA but has also a role as transcription factor [8] . It is also the most frequently mutated gene in Cornelia de Lange syndrome ( >60% of the cases ) . However , in about 25–30% of the cases the genetic cause of the disease is unknown . Some of these cases might be explained by mosaicism [62 , 63] , but the disease-causing mutations might also reside in gene-regulatory regions that are unknown and therefore not covered in diagnostics . Interestingly , in heterozygous Nipbl knock-out mouse tissues the Nipbl transcript was found to be only reduced by 25–30% and a compensatory regulatory mechanism for Nipbl was suggested [22] . Therefore , the gene regulation of NIPBL is of great interest for eventual approaches to manipulate the expression of the gene . Here we characterized in depth the regulation of the NIPBL gene by a bidirectional promoter controlling the transcription of NIPBL and of the 5 . 3 kb lncRNA NIPBL-AS1 . We investigated the role of the lncRNA for NIPBL expression and identified and characterized a conserved distal enhancer that stimulates NIPBL and NIPBL-AS1 expression . It has been shown that lncRNAs can function either through their RNA product or through their active transcription . To test this systematically we first depleted NIPBL-AS1 by antisense oligonucleotides ( ASO ) without observing a significant change in NIPBL expression ( Fig 1B and 1C ) . Similarly , reduction of the NIPBL transcript in CdLS patient LCLs with heterozygous NIPBL truncation mutations , at least in part due to nonsense-mediated mRNA decay , does not alter the level of lncRNA transcription ( Fig 5B and S8 Fig ) . Therefore the RNA products of both genes seem to do not influence each other . This is consistent with the observation that the NIPBL-AS1 sequence is not conserved among mammals . To test whether the transcription of the lncRNA from the bidirectional promoter is important for NIPBL expression , we blocked the transcription of NIPBL-AS1 using CRISPRi [50] and observed an upregulation of the NIPBL transcript . Vice versa , blockage of the transcription of NIPBL lead to an upregulation of the lncRNA ( Fig 2C and 2D ) . This underlines an interesting mechanism that could be potentially explored for the manipulation of NIPBL expression levels in clinical applications . Antisense transcription has been observed for a large number of promoters in the human genome [64 , 65 , 66] . The NIPBL-AS1 is a rather long unspliced noncoding RNA ( 5 . 3 kb ) , transcribed from a bidirectional promoter shared with NIPBL . Both genes seem to be transcribed to comparable levels according to the RNA-seq signals ( Fig 1A ) and our transcript analysis in LCLs ( Fig 5B ) . How the transcription at such bidirectional promoters is regulated is still unclear ( for a review see [67] ) ; therefore our observation that the transcripts are coupled in such manner that blocking the transcription of one gene leads to increased transcription of the other gene may represent a novel mechanism in fine tuning of gene expression . We also identified one distal enhancer of NIPBL and NIPBL-AS1 located 130 kb upstream of the promoter within a region that we call R1 ( Fig 3 ) . This region contains two CTCF sites: the one facing the NIPBL gene ( CTCF #1 ) correlates strongly with active chromatin marks ( e . g . H3K4me1 , H3K4me2 , H3K4me3 ) while the second one ( CTCF #2 ) shows only little correlation with those histone marks ( Fig 4A ) . The deletion of 5 kb ( D1 ) including a region ( R1_1 ) enriched in enhancer marks and including the CTCF#1 site led to downregulation of both NIPBL and NIPBL-AS1 expression . A larger deletion of 12 kb ( D2 , comprising R1_1 and R1_2 ) that included CTCF#1 and CTCF#2 did not lead to further downregulation of expression . The enhancer of NIPBL/NIPBL-AS1 resides therefore in R1_1 and overlaps with CTCF#1 . Interestingly , the 3C-seq experiments reveal that two CTCF binding sites close to NIPBL/NIPBL-AS1 bidirectional promoter strongly contact the two CTCF binding sites within the R1 region . The motif orientation of these sites is consistent with the preferential motif orientation for loop formation . Both sites in R1 seem to be required for the long-range interactions of the NIPBL/NIPBL-AS1 promoter . This promoter-enhancer interaction seems to be conserved between different tissues and also across species ( S4A and S4H Fig ) . We propose that the interactions between these CTCF sites are important for the recruitment of the enhancer to the NIPBL/NIPBL-AS1 promoter . It remains to be shown whether this enhancer-promoter loop is a permanent loop scaffold , as suggested for developmental genes in fly [68] and TNF-α-responsive genes in human [69] . Due to the close distance of the two TSS , the looped enhancer might stimulate alternatively the transcription at the TSS of NIPBL or NIPBL-AS1 , suggesting a competition between both genes for the enhancer activity . This is supported by our observation that blocking the transcription of one gene , with RNA Pol II Ser5 accumulating at the block site , leads to increased transcription of the other gene . An interesting observation concerning the NIPBL gene are the intensive contacts of the NIPBL promoter all over the NIPBL gene covering nearly all of the 47 exons ( 188 kb ) ( S3 Fig ) . Since the NIPBL intragenic viewpoints ( VP2-3 ) contact the promoter and the enhancer and vice versa the enhancer contacts the NIPBL gene body ( VP5 ) , we might eventually observe here a tracking of the promoter that is still in contact with the enhancer along the NIPBL gene together with elongating RNA Pol II , as recently described by Lee et al . [70] , although this remains to be tested in greater detail . In summary , we have discovered very important features of the genetic context of the NIPBL gene and obtained functional insight on the regulation of the NIPBL gene expression . Although we did not observe a direct effect of the lncRNA for the NIPBL gene , we found that the transcription of the two genes is interconnected , suggesting a mechanism for the fine-tuning of NIPBL expression . With these experiments we have also demonstrated one possibility to boost the transcriptional activity of the NIPBL gene by interfering with the NIPBL-AS1 lncRNA . This could be further explored for other genes with a similar arrangement of lncRNA and gene , and eventually also be used to manipulate gene expression in patient cells . Moreover , we identified a distal enhancer controlling sense and antisense transcription at the bidirectional promoter of NIPBL/NIPBL-AS1 . Given that even a modest reduction of NIPBL expression dramatically impacts on development , we suggest to include this non-coding genomic element into molecular diagnostics for CdLS .
HEK293T cell line was cultured in DMEM supplemented with 0 . 2mM L-glutamine , 100 units/ml penicillin , 100 mg/ml streptomycin and 10% FCS and was grown at 37°C and 5% CO2 . HB2 cells ( 1-7HB2 , a clonal derivative of the human mammary luminal epithelial cell line MTSV1-7 , [71] ) were cultured in DMEM supplemented with 0 . 2 mM L-glutamine , 100 units/ml penicillin , 100 mg/ml streptomycin , 10% FCS , 5 μg/ml hydroxycortisone and 10 μg/ml human insulin . Lymphoblastoid cell lines derived from controls and Cornelia de Lange syndrome patients [20] were obtained from Ian Krantz ( The Children’s Hospital of Philadelphia , Philadelphia , Pennsylvania , United States of America ) and cultured in RPMI medium supplemented with 0 . 2mM L-glutamine , 100 units per ml penicillin , 100 mg per ml streptomycin , 20% FCS . Cells were harvested and total RNA was prepared using Trizol Reagent ( Invitrogen ) . After chloroform extraction and isopropanol precipitation , pellets were dissolved in DEPC water . cDNA was generated by reverse transcription using oligo ( dT ) 18 primer ( Invitrogen ) , Superscript II Reverse Transcriptase ( RT ) ( Invitrogen ) and RNaseOUT Recombinant Ribonuclease Inhibitor ( Invitrogen ) according to the manufacturer’s instructions . The amounts of the different transcripts were compared by qPCR using SYBR Green and Platinum Taq Polymerase ( Invitrogen ) in CFX96 light cycler ( BioRad ) and specific primers . ΔΔCt method was used to calculate the fold change in gene expression using the housekeeping gene SNAPIN and the control sample for normalization . Anti-Sense Oligonucleotides ( ASO ) were designed followed the guidelines from Integrated DNA Technologies ( IDT ) ( https://eu . idtdna . com/Scitools/Applications/AntiSense/Antisense . aspx ? source=menu ) . Two ASO were designed against the 5’ and 3’ end of NIBPL-AS1 and one were designed against a non-targeting sequence to use as control . All the ASOs were modified with a phosphorothioate ( PS ) linkages that confer nuclease resistance . The following ASOs were used: The position of the phosphorothioate ( PS ) linkages is indicated as * . HB2 or HEK293T cells we transfected with 400 pmol of either ASO2 , ASO3 or Control ASO ( IDT ) using Lipofectamine 2000 ( Invitrogen ) according to the manufacture’s instruction . Cells were harvested 48 hours after transfection and cDNA was prepared as described above . ΔΔCt method was used to calculate the fold change in gene expression using the housekeeping gene SNAPIN and the control sample for normalization . The expression vectors for active Cas9 and catalytically inactive Cas9 were obtained from Addgene . The guide RNA were designed using the Cas9 design tool ( http://cas9 . cbi . pku . edu . cn/CasDesign ) [72] and inserted in the gRNA Cloning Vector ( Addgene #41824 ) [73] following the protocol deposited together with the vector . The sequences and positions of the guide RNA are listed in the S1 Table . To block the transcription of NIPBL-AS1 , two gRNAs ( G1 and G2; Block I ) targeting the 5’ end of NIPBL-AS1 were used . To block the transcription of NIPBL , three gRNAs ( G3 , G4 , G5; Block II ) targeting the 5’ end of NIPBL were used . As controls we used either a combination of guide RNAs on NIPBL that are not effective for RNA PolII blocking ( G6 and G7; Block III ) or one guide RNA localizing outside of the locus ( GC ) . The gRNA were designed using the web tool http://cas9 . cbi . pku . edu . cn/CasDesign and cloned into a pCR-Blunt II-TOPO vector ( Plasmid #41824 from AddGene ) by Gibson Assembly ( New England BioLabs ) according to the manufacture’s instruction . HEK293T cells were transfected with the gRNAs vectors and the catalytically inactive Cas9 vector ( dCas9 ) [74] ( Plasmid #47948 from AddGene ) using Lipofectamine 2000 ( Invitrogen ) according to the manufacture’s instruction . Cells were harvested 48 hours after transfection and cDNA was prepared as described above . ΔΔCt method was used to calculate the fold change in gene expression using the housekeeping gene SNAPIN and the Control sample for normalization . Chromosome conformation capture sequencing was performed as previously described in [55] . Briefly , cells were crosslinked with 1% ( w/v ) formaldehyde for 10 minutes and quenched with 120mM glycine . Crosslinked-cells were resuspended in lysis buffer ( 50mM Tris-HCl pH 8 . 0 , 0 . 5% NP-40 , 50mM NaCl and Complete protease inhibitor ( Roche ) ) and subjected to enzymatic digestion using 400 units of BglII ( Roche ) or ApoI ( New England Biolabs ) for the higher resolution protocol . Digested chromatin was then diluted and ligated using 5 units of T4 DNA ligase ( Promega ) under conditions favoring intramolecular ligation events . After reversing the crosslink at 65°C over night , the digested and ligated chromatin was subjected to a second enzymatic digest using NlaIII ( New England Biolabs ) to produce smaller DNA fragment . The resulting digested DNA underwent a second ligation using 10 units of T4 DNA ligase ( Promega ) under conditions favoring self-ligation events that produce circular DNA molecules . The unknown DNA fragment , ligated to the fragment of interest ( called viewpoint ) , was amplified by inverse-PCR using specific primer design in the outer part of the restriction site of the viewpoints , linked with the Illumina adapter sequences . The samples were then single-read sequenced using the Illumina Genome Analyzer II generating 76bp reads . The reads were trimmed to remove the Illumina adapter sequences and mapped against human genome ( hg18 and hg19 ) . Analysis was performed as previously described [1 , 75] . To deplete the candidate enhancers ( R1_1 and R1_2 ) in region R1 , three gRNAs ( gRNA_1 , gRNA2 and gRNA_3 ) targeting the region were designed and cloned as described above . To delete R1_1 ( 5 kb deletion in R1 region ) , HEK 293T cells were transfected with 2ug of gRNA2 , gRNA3 and the vector coding for the catalytically active Cas9 ( Plasmid #48139 , AddGene ) using Lipofectamine 2000 ( Invitrogen ) according to the manufacture’s instruction . To delete the entire R1 region including R1_1 and R1_2 ( 12 kb deletion ) , HEK293T cells were transfected with 2 μg of gRNA1 , gRNA3 plasmids and the catalytically active Cas9 . Cell were cultured in a puromycin-containing medium and single clones were picked and PCR analyzed for the presence of the corresponding deletion . Cells from different clones were harvested and lysed with lysis buffer ( 100mN Tris-HCl pH8 . 0 , 5mM EDTA , 0 . 2% SDS , 50mM NaCl and proteinase K ) . Genomic DNA was isolate using isopropanol and samples were analyzed by PCR . Primer sequences are listed in the Supplemental information . Clones positive for deletions D1 and D2 were grown , cells were harvested and cDNA was prepared as described above . HEK293T cells were used as control . ΔΔCt method was used to calculate the fold change in gene expression using the housekeeping gene SNAPIN and the Control sample for normalization . Fractionation of cells was performed as described [76] with addition of RNAse inhibitors ( RNAseout , Invitrogen ) according to the manufacturer’s instructions . Preparation of RNA with from the fractions using Trizol and the cDNA preparation and subsequent qPCR analysis are described above . To test whether the candidate enhancer has indeed an effect on the NIPBL promoter we cloned the luciferase gene under control of the NIPBL promoter ( hg19 , pos . chr5:36 , 875 , 913–36 , 876 , 915 ) together with the putative enhancer elements ( R1_1— ( hg 19 ) chr5:36743873–36745877 , R1_2 – ( hg 19 ) chr5:36738465–36742009 ) or a similar sized control region in the pGL4 . 10 vector ( see scheme Fig 4A ) . Cloning primer sequences are available upon request . HEK293 cells were transfected using FuGENE-HD ( Promega , Madison , USA ) with one of the following constructs: ( 1 ) empty pGL4 . 10 vector; ( 2 ) pGL4 . 10 vector containing R1_1 and NIPBL promoter; ( 3 ) ( 2 ) pGL4 . 10 vector containing R1_2 and NIPBL promoter; ( 4 ) pGL4 . 10 vector containing genomic DNA of the same size as the regulatory element . After 24 hours the cells were lysed and the activity of firefly and renilla luciferase determined with the Dual Luciferase Reporter Assay ( Promega ) in a TriStar2 LB Multidetection Microplate Reader ( Berthold , Bad Wildbad , Germany ) . Genomic regions and primer sequences are given in the table . All measurements were verified in a minimum of three independent experiments and as triplicates in each experiment . The chromatin immunoprecipitation with anti-RNA PolII Ser5 and anti-PolII was performed as described [1] . A DNA fragment of interest was PCR-amplified with a biotinylated primer as described [77] . After denaturation , the biotinylated single-stranded PCR amplicon was hybridized with the sequencing primer , specific for the analyzed position . The allelic dosage was quantified on PyroMark Q24 instrument , with PyroMark Gold Q24 Reagent Kit ( Qiagen ) , according to the manufacturer's instructions . NIPBL—monoclonal rat anti-NIPBL , isoform A ( long isoform ) NP_597677 ( Absea , China , 010702F01 clone KT54 ) and isoform B ( short isoform ) NP_056199 ( Absea , China , 010516H10 clone KT55 ) SMC2 –rabbit polyclonal anti-SMC2 [78] Tubulin—mouse anti-tubulin ( Sigma ) There are no ethical issues . All cell lines used in this manuscript were previously published [20] . | The most frequent mutations in the human developmental disorder Cornelia de Lange Syndrome ( CdLS ) occur in the NIPBL gene . NIPBL is critical for chromatin-association of the cohesin complex and has a dual role as transcription factor . The regulation of the NIPBL gene is of great interest since organisms are very sensitive to NIPBL levels . For instance , severely affected patients showed only ~65% and mildly affected patients ~75% of NIPBL mRNA levels . One case reports a CdLS phenotype with as little as 15% reduction of the NIPBL transcript . Further , in a number of patients with CdLS phenotype no mutations in known CdLS genes are found , raising the question whether mutations could also occur in gene-regulatory elements . Here we investigate the role of a long non-coding RNA NIPBL-AS1 originating from a promoter shared with NIPBL and observe a co-regulation by a distal enhancer that we have identified and that seems to be conserved between tissues . Deletion of the enhancer by CRISPR leads to reduced expression of NIPBL and NIPBL-AS1 but also of NIPBL target genes that were found to be dysregulated in CdLS patient cells . The lncRNA NIPBL-AS1 itself has no role for NIPBL expression , but its transcription reduces the expression of the NIPBL gene and vice versa , thus revealing an interesting mechanism for the fine-tuning of expression levels . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"gene",
"regulation",
"messenger",
"rna",
"dna-binding",
"proteins",
"cloning",
"long",
"non-coding",
"rnas",
"dna",
"transcription",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"transcriptional",
"control",
"proteins",
"gene",
"expression",
"histones",
"molecular",
"biology",
"biochemistry",
"rna",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"non-coding",
"rna"
]
| 2017 | Regulation of the cohesin-loading factor NIPBL: Role of the lncRNA NIPBL-AS1 and identification of a distal enhancer element |
Scrapie is a transmissible spongiform encephalopathy ( TSE ) in sheep and goats . In recent years , atypical scrapie cases were identified that differed from classical scrapie in the molecular characteristics of the disease-associated pathological prion protein ( PrPsc ) . In this study , we analyze the molecular and neuropathological phenotype of nine Swiss TSE cases in sheep and goats . One sheep was identified as classical scrapie , whereas six sheep , as well as two goats , were classified as atypical scrapie . The latter revealed a uniform electrophoretic mobility pattern of the proteinase K–resistant core fragment of PrPsc distinct from classical scrapie regardless of the genotype , the species , and the neuroanatomical structure . Remarkably different types of neuroanatomical PrPsc distribution were observed in atypical scrapie cases by both western immunoblotting and immunohistochemistry . Our findings indicate that the biodiversity in atypical scrapie is larger than expected and thus impacts on current sampling and testing strategies in small ruminant TSE surveillance .
Transmissible spongiform encephalopathies ( TSEs ) are fatal neurodegenerative diseases that are caused by prions [1] and include , among others , Creutzfeldt-Jakob and Gerstmann-Sträussler-Scheinker disease ( GSS ) in humans , bovine spongiform encephalopathy ( BSE ) in cattle , and scrapie in small ruminants ( SRs ) , comprising sheep and goats . A hallmark of TSEs is the deposition of a misfolded , partially proteinase K–resistant isoform ( PrPsc ) of the host-encoded normal prion protein ( PrPc ) . Current molecular diagnostic procedures rely on the detection of the proteinase K–resistant core fragment , termed PrPres . Concerns were raised that BSE might have entered the SR population unnoticed . Moreover , in experimental SR BSE , infectivity was detected in various nonnervous tissues [2] and body fluids [3] and may present a risk for consumers and livestock . As a consequence , in 2002 , European Union member states enhanced TSE surveillance in SRs , and subsequently , the first natural BSE case in a French goat was identified [4] . Besides this , numerous so-called atypical scrapie cases that differed in their neuropathological and molecular characteristics from classical scrapie and experimental SR BSE were first reported in Norway [5] and later in other countries ( reviewed by [6] ) . Importantly , such cases also occurred in sheep with genotypes considered to confer resistance to scrapie and were missed by some screening test formats while others readily detected them [7] . In the meantime , specific SR TSE screening tests have been evaluated and approved by the authorities . Consequently , the number of atypical scrapie cases reported increased , but still very little information is available about the phenotype and epidemiology of this disease . To date , three forms of SR TSE can be discriminated by molecular PrPres typing [8]: classical scrapie , atypical scrapie , and BSE . In the present study , we aimed at analyzing the complete set of SR TSE cases detected by active and passive surveillance in Switzerland from 2004 to 2005 for the molecular and neuropathological phenotype , and at assessing the efficacy of current testing strategies . Our results indicate unexpected variations in the phenotype of atypical scrapie in sheep and goats .
In the framework of a 1-y active surveillance program , more than 30 , 000 regularly slaughtered ( RS ) and more than 3 , 000 fallen SR stock ( FS ) were analyzed for TSE in Switzerland . RS animals were screened with either of the two approved screening tests: the Prionics Check Western SR ( referred to as test A ) and the Bio-Rad TeSeE sheep and goat ELISA ( referred to as test B ) in caudal brainstem samples . By contrast , FS , as well as clinically suspicious ( CS ) animals were analyzed by immunohistochemistry ( IHC ) in different brain structures and both screening tests in caudal brainstem samples in parallel . Applying this testing scheme , nine SR TSE cases ( Table 1 ) were identified and confirmed by IHC . In three of these animals , clinical signs of a central nervous system disorder ( Table 1 ) were reported . Lymphoid tissues were available in four cases ( S5/FS , S6/FS , S7/CS , and G2/FS ) and tested negative with screening test B ( retropharyngeal lymph node ) and by IHC ( retropharyngeal lymph node , tonsils ) . In slaughtered SRs , ~47% of the animals were screened with test A and ~53% with test B in regional laboratories . Two slaughtered TSE cases were initially picked up by test A ( S3/RS and G1/RS ) and three by test B ( S1/RS , S2/RS , and S4/RS ) . The respective screening tests were repeated in the reference laboratory , and fresh brainstem samples from the same animals were cross-checked with the second screening test . All nine TSE cases scored positive in test B , but large variations in the optical density values were observed between replicates for some samples ( Table 1 ) . By contrast , only five TSE cases gave a clear positive signal in test A ( Figure 1 and Table 1 ) . For further classification of the TSE cases identified , a refined western immunoblot ( WB ) method was used , resulting in an improved resolution of individual PrPres bands . Analyses were undertaken repeatedly , and molecular masses and relative signal intensities of PrPres bands were measured . Case S1/RS showed a three-band pattern , with moieties migrating at ~33 kDa , ~30 kDa , and ~21 kDa , respectively ( Figure 2 ) . This electrophoretic mobility pattern ( EMP ) is indicative for either classical scrapie or BSE [8] . Consequently , the sample was analyzed using a discriminatory WB technique , thereby presenting features of classical scrapie and not of BSE ( Figure 3 ) . All other SR TSE cases revealed a different multiband pattern with three major bands migrating at a molecular mass of ~8 kDa , ~23 kDa , and ~31 kDa , respectively ( Figure 2 ) . In animals with high PrPres signal intensities , one to two additional minor bands in the range of 15–17 kDa were observed . Based on these observations and according to defined criteria [8] , these TSE cases were classified as atypical scrapie . All atypical scrapie cases revealed a consistent EMP ( Figure 4A ) and similar proportions of the intensity of bands ( Figure 4B ) , although some slight variations between individual samples were observed . With both criteria , atypical scrapie cases clearly differed from Swiss and British classical scrapie reference cases and S1/RS . In four of the atypical scrapie cases ( S5/FS , S6/FS , S7/CS , and G2/FS ) whole brains were available . In order to estimate the distribution profile of abnormal prion protein throughout the brains of these animals , samples of representative neuroanatomical structures were analyzed by the refined WB and compared to those of Swiss and British classical scrapie cases . Very similar EMPs were observed in all four atypical cases in these structures ( Figure 5A–5D ) . In G2/FS ( Figure 5C ) , PrPres was not detected in the caudal brainstem and the cerebellar cortex . In all other structures examined , especially in thalamus , the basal ganglia , and the cerebral cortex , it was present in high levels . Interestingly , a completely diverse picture was observed in sheep S6/FS where PrPres was detected in the cerebellar cortex , and barely in caudal brainstem , but not in the more rostral structures ( Figure 5B ) . In sheep S7/CS ( Figure 5D ) and S5/FS ( Figure 5A ) , PrPres deposits were most pronounced in the cerebellar cortex , hippocampus , basal ganglia , and the piriform lobe , but again the caudal brainstem sample reacted weakly . By contrast , in the classical scrapie reference cases from the United Kingdom ( unpublished data ) and Switzerland ( Figure 5E ) , PrPres accumulations were consistently detected at high amounts , notably in the caudal brainstem , but also in all other structures examined . Next , we attempted to specify the findings in WB by using an alternative prion protein ( PrP ) detection method . To this end , available formalin-fixed brain tissues of all SR TSE cases identified and a Swiss classical scrapie reference case , were analyzed by IHC , and the type and intensities of PrPsc deposits in defined neuroanatomical structures were assessed by light microscopy ( Table 2 ) . In the atypical scrapie cases , the PrPsc deposition type consisted of fine to coarse granules that were found in the gray and white matter ( Figure 6A and 6B ) . All the atypical cases , except S6/FS , revealed a faint positive immunolabeling in the caudal brainstem that involved mainly the spinal tract nucleus of the trigeminal nerve , and in two cases ( S7/CS and S5/FS ) additionally the reticular formation and the ambiguus nucleus . The dorsal motor nucleus of the vagus nerve was not affected . One remarkable finding was that cases S7/CS and S5/FS had equal amounts of PrPsc in the molecular and the granular layer of the cerebellar cortex , whereas cases S2/RS and S3/RS showed severe accumulations in the molecular layer , but to a much lesser extent in the granular layer . In line with the findings in WB , cases S7/CS , S5/FS , and G2/FS showed moderate to severe amounts of PrPsc in the more rostral structures , such as thalamus , basal ganglia , and cerebral cortex . Case S6/FS was essentially negative in all of these structures . It is noteworthy that in sheep S7/CS , dense , plaque-like deposits were identified in the ventral and the medial geniculate body of the thalamus ( Figure 6B , inset ) . In the classical scrapie case , S1/RS , severe amounts of intraneuronal , perineuronal , and coarse granular type PrPsc were detected in the caudal brainstem . In the Swiss classical scrapie reference case ( Figure 6C ) , PrPsc labeling of neurons and neuropil was detected throughout the brain in high amounts . Here the deposits were of the intraneuronal , perineuronal , and stellate , as well as coarse granular type , and involved also the dorsal motor nucleus of the vagus nerve of the caudal brainstem . Overall , the PrPsc distribution profile of SR TSE cases as determined by IHC correlated well with the results of the refined WB analyses in all animals , with some minor variations . The interpretation of the haematoxylin and eosin–stained tissue slides of cases S1/RS , S2/RS , S3/RS , S4/RS , and G1/RS was not possible due to autolysis of the tissue and/or wrong tissue treatment such as freezing prior to embedment . Nevertheless , in cases S5/FS , S6/FS , S7/CS , and G2/FS spongiform lesions were identified in the neuropil that were generally associated with PrPsc accumulation . In these four cases , spongiform lesions were absent in the brainstem at the level of the obex . In cases S7/CS ( Figure 6B ) and S5/FS , moderate to severe vacuolation was found in the molecular layer of the cerebellar cortex , the basal ganglia , and the cerebral cortex , whereas the midbrain and hippocampus were only mildly to moderately affected . Case S6/FS showed only mild vacuolation in the molecular layer of the cerebellar cortex ( Figure 6A ) . Goat G2/FS has already been described elsewhere [9] . No intraneuronal vacuoles were observed in any of the structures examined . A diffuse gliosis appeared in cases S7/CS and G2/FS . In the classical scrapie case , S1/RS , interpretation of vacuolation in the neuropil was complicated , but mild intraneuronal vacuolation was identified in the dorsal motor nucleus of the vagus nerve ( unpublished data ) .
In this study , we demonstrate that the vast majority of SR TSE cases identified in Switzerland by active surveillance were atypical scrapie . Their PrPres EMP was consistent , but prominent differences in the distribution of abnormal prion protein in the brains compared to each other and to those in classical scrapie–affected sheep were observed . Importantly , none of the cases revealed evidence for an infection with the BSE agent . According to the literature , atypical scrapie is characterized by a PrPres EMP with at least three to five bands , including a prominent lower band , estimated to migrate at 7–12 kDa [5 , 10–12] . In the present study , a detailed analysis of PrPres in Swiss atypical scrapie cases showed a uniform multiband EMP with three major bands migrating at ~8 kDa , ~22 kDa , and ~32 kDa . In a previous study , a side-by-side comparison by WB showed that the PrPres EMPs of sheep S7/CS , goat G2/FS , and a Nor98 case , the prototype of atypical scrapie , were essentially indistinguishable [9] . These results indicate that the Swiss atypical scrapie cases share a common molecular PrPres type similar to Nor98 regardless of the host species and genotype . The phenotype of a TSE is defined , among other criteria , by the neuroanatomical distribution of histopathological lesions and PrPsc deposits . Most atypical scrapie cases were identified by active TSE surveillance; complete brain was usually not available because sampling is performed through the foramen magnum accessing the brainstem ( and ideally also the cerebellum ) , but not the more rostral structures . Thus , until this date , very limited information is available about these criteria in brains of atypical scrapie–affected animals . In the literature for brainstem , no or only weak PrPsc deposits were detected [5 , 7] , whereas strong deposits were described for the cerebellum [5 , 13–15] . Positive WB signals have also been shown in samples from the cerebral cortex [5 , 15] , midbrain [14] , and thalamus [15] . In this study , we present for the first time a systematic comparative analysis of the PrPsc distribution in brains of atypical scrapie–affected SRs by both WB and IHC . Our results show , that unlike some human TSEs [16] , the PrPres EMP in WB and the PrPsc deposition type in IHC remains constant regardless of the brain structure examined . However , when we analyzed the neuroanatomical PrPsc distribution by WB and IHC prominent differences were observed . In one goat ( G2/FS ) , the cerebrum , thalamus , and midbrain were severely affected with a minimal involvement of the cerebellum and the caudal brainstem ( Figure 5C and [9] ) . By contrast , a completely different pattern was observed in one sheep ( S6/FS ) where PrPsc accumulated in the cerebellar cortex almost exclusively ( Figures 5B and 6A ) . In other cases ( S5/FS and S7/CS ) , severe deposits were detected in the cerebrum , thalamus , and midbrain and also involved the cerebellum to different extents , but again barely the caudal brainstem ( Figure 5A and 5D ) . Small differences between IHC and WB results were observed but can be biased by the sampling procedure of frozen tissues for WB . Unfortunately , a complete and detailed analysis of the RS cases was not possible , due to the lack of whole brains and severe tissue autolysis . However , some brain structures , in addition to caudal brainstem , were available for IHC of S2/RS and S3/RS , and both cases seem to fit into either of the schemes found in S5/FS and S7/CS or S6/FS , respectively . The localization and severity of spongiform lesions in S5–S6/FS , S7/CS , and G2/FS correlated well with the PrPsc deposits in IHC . Taken together , our results indicate a considerable diversity in the neuropathological representation of atypical scrapie in SRs . A possible explanation for these findings is that the animals died at different time points in the course of the disease . One could speculate that PrPsc deposits and histopathological lesions evolve from the cerebrum to the cerebellum , and to a lesser extent , the caudal brainstem during the pathogenesis . This concept may account for cases S5/FS and S7/CS , since both were clinically diseased with prominent signs of a central nervous system disorder . However , the PrPsc distribution in S6/FS clearly cannot be explained by this means . Moreover , in S2/RS and S3/RS the cerebellum was severely affected , but these animals were regularly slaughtered , which implicates that clinical signs were not obvious at that time . Thus , the heterogeneities observed do not seem to reflect different stages of the disease exclusively . Another factor that may contribute to the phenotype of atypical scrapie is the PrP genotype of the host . Epidemiological studies indicate associations between the AF141RQ and the AHQ PrP alleles with atypical scrapie in sheep [17 , 18] . Only one of the Swiss atypical scrapie cases ( S7/CS ) encoded the AF141RQ allele , but three were homozygote for ARR , a genotype that is considered to confer resistance to classical scrapie . An interesting observation was that , in contrast to the other sheep , in ARR/ARR genotypes , the cerebellar cortex in the molecular layer was consistently affected in a much more severe manner compared to the granular layer . Based on our data we can not establish a clear correlation between the neuropathological diversity and the PrP genotype in scrapie-affected sheep . More field cases and data from experimental transmission studies are needed . Hence , the cause of neuropathological diversities in atypical scrapie cases remains unclear and may be linked to other host factors , the infectious agent involved , or the interaction of both . The molecular phenotype in atypical scrapie is unique for animal TSEs , but reminiscent to the one described for GSS . In GSS , also unique for human TSEs , a prominent PrPres band similarly migrating at ~7 kDa in WB analysis has been described [19] . These lower fragments in atypical scrapie [10 , 11] and GSS [20] correspond to N- and C-terminal truncated PrP peptides . Since it is also detectable in nonproteinase K–treated tissue homogenates , it was suggested that these peptides are produced in vivo by a proteolytic pathway that partially degrades PrP in the extracellular compartment . PrPsc deposition profiles very similar to G2/FS , S5/FS , S6/FS , and S7/CS have been described for GSS: a cerebellar type [21] , a cerebral type , and combinations of both [22] . These distribution profiles are correlated to defined mutations in the host's PrP gene . To identify common genetic components in atypical scrapie and GSS , we compared the amino acid sequences of the prion protein in Swiss atypical scrapie cases with those found in GSS patients ( Figure 7 ) , but a common genetic background in the two diseases was not observed since the mutations described for GSS were absent in atypical or classical scrapie cases . Further , polymorphisms known to determine resistance or susceptibility to TSE in SRs were not found in GSS patients . Another striking similarity to GSS is the presence of plaque-like PrPsc aggregations in the thalamus of S7/CS . Such PrPsc deposits are a major feature of GSS in humans [22] and have not yet been described for atypical scrapie . None of the other cases under investigation showed this deposition type . It must be pointed out that S7/CS was in an advanced stage of the disease , and thus , at best comparable with samples derived from GSS patients . Whether these plaques are a general feature in late-stage atypical scrapie cases remains to be determined . For GSS the hereditary background is evident and an epidemiological link between these two diseases is not obvious until to date . However , since a direct comparison of GSS with atypical scrapie in the same experimental setup is still missing , one can only speculate that both diseases involve similar pathomolecular mechanisms . Atypical scrapie may therefore represent a valuable model for investigations in the pathogenesis of GSS and vice versa . From the diagnostic point of view , our results were unexpected , as none of the structures examined was consistently affected in all of the animals . This has a strong impact on sampling techniques in terms of active and passive SR TSE surveillance . European Union member states have implemented sampling of cerebellum in addition to the standard brainstem sample after increasing reports of atypical scrapie . Our results sustain this action , insofar as the cerebellum indeed consistently revealed higher amounts of PrPsc compared to caudal brainstem samples . For cases of a cerebellar type ( like S6/FS ) this procedure seems to be optimal . Beyond this , in most of the cases , especially in G2/FS , samples from the central parts of the brain such as thalamus or basal ganglia would be optimal to ensure a maximum in the sensitivity of PrPsc detection . However , such a strategy requires sampling of the whole brain and should be possible in passive surveillance but may be unrealistic due to practical and economical constraints for large active surveillance programs in most countries . This dilemma may partially be overcome by the application of highly sensitive screening tests on cerebellum samples . It would be wise to reconsider current sampling and testing strategies and to enhance disease awareness in order to increase the efficacy of passive TSE surveillance schemes . It has been demonstrated that atypical scrapie involves a transmissible agent [23] , but as it mainly occurs as an isolated case in a flock , it remains unclear whether the disease can spread naturally in the SR population . Atypical scrapie is possibly a sporadic TSE in SRs . Next , our work will be extended to comparative transmission studies in rodent models and SRs for a further characterization of the infectious agents involved and the biodiversity in the phenotype of atypical scrapie .
Eight TSE cases were detected via active surveillance in sheep ( S1/RS–S6/FS ) and goats ( G1/RS and G2/FS ) ( Table 1 ) . A clinically affected scrapie case ( S7/CS ) diagnosed in Switzerland in 2004 by passive TSE surveillance was included in this study . Classical scrapie reference material originated from two sheep inoculated with central nervous system material from a Swiss clinically scrapie–affected ram , and five British classical scrapie cases were provided from the Veterinary Laboratory Agency Archive , Weybridge , United Kingdom . BSE-positive control tissue was taken from a clinically affected Swiss cow . Negative control tissues were derived from sheep testing negative with both rapid and confirmatory tests . In Switzerland , the Check Western SR ( Prionics , http://www . prionics . com ) , referred to as screening test A , and the TeSeE Sheep/Goat ELISA ( Bio-Rad , http://www . bio-rad . com ) , referred to as screening test B , were approved for TSE screening in SRs . Both tests were performed according to the manufacturer's instructions . PrPres was purified according to the protocol of the TeSeE Sheep/Goat Western Blot ( Bio-Rad ) , separated by SDS PAGE , immunoblotted on nitrocellulose membranes , and detected with the monoclonal antibody P4 ( R-Biopharm , 400 ng/ml; http://www . biopharm . com ) and the secondary goat anti-mouse antibody IRDye 800 ( 200 ng/ml; Rockland , http://www . rockland-inc . com ) using an Odyssey infrared imager ( LI-COR Biosciences , http://www . licor . com ) . Protein bands were identified , quantified , and their molecular mass estimated with the Odyssey analysis software version 2 . 1 ( LI-COR Biosciences ) . For the BSE/scrapie discriminatory WB , samples were processed as described for the refined WB , but PrPres was detected separately with MAb P4 and MAb 6H4 ( 200 ng/ml; Prionics ) in parallel . For histopathology , 4-μm-thick sections were deparaffinized and stained with haematoxylin and eosin . Confirmatory IHC and the PrPsc distribution profiling in case G2/FS was performed as described previously [9] . For the PrPsc distribution profiling in sheep , the protocol was optimized as follows: 4-μm-thick tissue sections were deparaffinized and immersed in 98% formic acid ( 30 min ) . After citrated autoclaving ( 121 °C , 30 min ) , the endogenous peroxidase activity was blocked ( H2O2 in methanol , 10 min ) and slides placed in distilled water overnight at 4 °C . After incubation with 5% normal goat serum ( Dako , http://www . dako . com ) for 20 min , PrPsc was detected with the monoclonal antibody F99/97 . 6 . 1 ( 2 μg/ml; VMRD , http://www . vmrd . com ) . Subsequent steps were performed with the peroxidase/DAB Dako Envision detection kit ( Dako ) . PrPsc immunolabeling was scored semi-quantitatively on a scale from 0 ( absent ) to 3 ( severe ) . For additional information about Materials and Methods see Protocol S1 .
Sequences were translated from the GenBank database ( http://www . ncbi . nlm . nih . gov/Genbank ) and the accession numbers were caprine ( EF140716 ) , human ( NM_183079 ) , and ovine ( U67922 ) . | In the view of concerns that bovine spongiform encephalopathy has entered the small ruminant population , comprehensive active surveillance programs for transmissible spongiform encephalopathies ( TSEs ) in sheep and goats were implemented worldwide . In these , previously unrecognized atypical scrapie cases were identified that to date represent the majority of detected small ruminant TSE cases in some countries . The pathogenesis and epidemiology of atypical scrapie , as well as its relevance to both animal health and food safety , is still poorly understood . In the present study , we performed a systematic neuropathological analysis of recently diagnosed atypical scrapie cases in Switzerland . Our results show that the neuropathological presentation in atypical scrapie–affected small ruminants varies remarkably , and the results indicate a biodiversity of TSEs in sheep and goats larger than expected , with some similarities to known human TSEs . These findings will form the basis for future research on TSE phenotypes and help to design experimental studies necessary to generate data for risk assessments and the implementation of appropriate disease-control strategies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
]
| [
"infectious",
"diseases",
"sheep",
"none",
"vertebrates",
"animals",
"mammals"
]
| 2007 | Diversity in Neuroanatomical Distribution of Abnormal Prion Protein in Atypical Scrapie |
In nearly all picornaviruses the precursor of the smallest capsid protein VP4 undergoes co-translational N-terminal myristoylation by host cell N-myristoyltransferases ( NMTs ) . Curtailing this modification by mutation of the myristoylation signal in poliovirus has been shown to result in severe assembly defects and very little , if any , progeny virus production . Avoiding possible pleiotropic effects of such mutations , we here used pharmacological abrogation of myristoylation with the NMT inhibitor DDD85646 , a pyrazole sulfonamide originally developed against trypanosomal NMT . Infection of HeLa cells with coxsackievirus B3 in the presence of this drug decreased VP0 acylation at least 100-fold , resulting in a defect both early and late in virus morphogenesis , which diminishes the yield of viral progeny by about 90% . Virus particles still produced consisted mainly of provirions containing RNA and uncleaved VP0 and , to a substantially lesser extent , of mature virions with cleaved VP0 . This indicates an important role of myristoylation in the viral maturation cleavage . By electron microscopy , these RNA-filled particles were indistinguishable from virus produced under control conditions . Nevertheless , their specific infectivity decreased by about five hundred fold . Since host cell-attachment was not markedly impaired , their defect must lie in the inability to transfer their genomic RNA into the cytosol , likely at the level of endosomal pore formation . Strikingly , neither parechoviruses nor kobuviruses are affected by DDD85646 , which appears to correlate with their native capsid containing only unprocessed VP0 . Individual knockout of the genes encoding the two human NMT isozymes in haploid HAP1 cells further demonstrated the pivotal role for HsNMT1 , with little contribution by HsNMT2 , in the virus replication cycle . Our results also indicate that inhibition of NMT can possibly be exploited for controlling the infection by a wide spectrum of picornaviruses .
The family Picornaviridae encompasses 35 genera of small non-enveloped RNA viruses that include numerous animal and human pathogens ( www . picornaviridae . com ) . They are causative agents of diseases with huge impact on health care and economy ranging from the relatively harmless common cold typically caused by rhinoviruses to flaccid paralysis as live-threatening complication from poliovirus infection . Vaccines are available only against few members and currently no antiviral drug has been approved for clinical use [1–3] . The ~30 nm diameter icosahedral picornavirus capsid is—in most instances—built from 60 copies of viral proteins VP1 , VP2 , VP3 , and VP4 ( buried inside ) . It contains a 6 . 7 to 9 . 7 kb single-stranded positive-sense RNA genome ( ss ( + ) RNA ) , whose single open reading frame is translated into a large polyprotein comprising a structural ( P1 ) and two nonstructural domains ( P2 and P3 ) . These are separated from each other and further cleaved by virus-encoded proteases into about a dozen smaller proteins required for replication of the viral RNA and capsid assembly . Processing of the P1 precursor by the viral proteases 3Cpro/3CDpro yields the capsid proteins VP0 , VP1 and VP3 , which remain associated as a 5S protomer . Five protomers assemble into a 14S ( immature ) pentamer and twelve pentamers construct the 75S to 80S empty procapsid . Progeny ss ( + ) RNA produced in replication complexes located on the surface of virus-induced vesicles is either threaded into the procapsid , or more likely , pentamers recruited by the nonstructural protein 2C assemble and condense around the nascent viral RNA . In the latter instance , accumulating procapsids are either regarded as a storage form of pentamers or a dead-end by-product of virion morphogenesis . The resulting 150S provirions undergo autocatalytic cleavage of VP0 into VP2 and VP4 giving rise to infectious virions . Capsid maturation is rapid in enteroviruses ( e . g . completed after 1 h in rhinovirus 16 [4] ) but quite protracted for hepatoviruses ( at neutral pH ) [5] . Parecho- , kobu- and saliviruses lack a maturation cleavage and maintain a shell composed of VP1 , VP3 , and VP0 ( reviewed in [6] ) . The polyprotein of almost all picornaviruses and derived processing products sharing the same N-terminus ( i . e . P1 , VP0 , VP4 ) are N-myristoylated [7] . This predominantly co-translational modification involves the transfer of myristate , a 14-carbon saturated fatty acid ( C14:0 ) , from myristoyl-CoA to an N-terminal glycine in the context of a myristoylation signal ( GXXX ( S/T ) ) following removal of the initiator methionine from the nascent polypeptide chain by methionine aminopeptidase . In apoptotic cells , certain proteins undergo post-translational myristoylation upon exposure of a cryptic glycine by caspase-mediated cleavage [8] . A substantial number of viral and cellular proteins are N-terminally myristoylated , which contributes to their subcellular location and function by modulating protein-membrane and protein-protein interactions [9–11] . In higher eukaryotes , the reaction is catalyzed by the two ubiquitously expressed N-myristoyltransferase isozymes NMT1 and NMT2 ( EC 2 . 3 . 1 . 97 ) . Despite overlapping substrate specificity [12] , they exert unique roles in a number of biological processes [13 , 14] . The atomic structures of poliovirus 1 ( PV1 ) [7] , coxsackievirus B3 ( CVB3 ) [15] and several other enteroviruses ( [16] and references therein ) reveal a spatially very similar organization of the myristic acid chain covalently attached to the N-terminus of VP4 , though C14 up to C11 of the fatty acid may be disordered . The myristoyl groups initially cluster together around the icosahedral 5-fold axis and then splay apart to cradle a twisted parallel β-tube composed of the five VP3 N-termini of a pentamer . Each myristoyl chain interacts with its 5-fold related siblings and additionally engages the VP4 and VP3 N-terminal extensions of adjacent 5-fold related protomers . This network of contacts appears to stabilize the mature ( VP0 cleaved into VP2 and VP4 ) pentamer subunits of the virion . Indeed , mutation of Thr 28 in VP4 of PV1 , which forms a conserved hydrogen bond with the carbonyl oxygen of a myristoyl moiety from an adjacent five-fold symmetry-related VP4 , resulted in unstable virions in concert with assembly anomalies . Notably , the manifestation of the assembly defect was strongly dependent on the replaced amino acid , which complicated interpretation of the results [17] . Only nonviable mutants were obtained , when N-myristoylation was ablated by site-directed mutagenesis of the glycine at the N-terminus of the PV1 polyprotein . However , no consistent picture has emerged on the underlying cause , which ranged from defects in polyprotein processing and RNA replication to impaired viral morphogenesis at different stages [18–24] . Studies on the in vitro and in vivo formation of myristoylation-deficient assembly intermediates of foot-and-mouth disease virus ( an Aphthovirus ) showed profoundly different outcomes regarding the lipid´s contribution in this process [25–29] . To overcome the technical difficulties associated with these systems [30 , 31] , a viable PV1 mutant with about 50% reduced myristoylation was generated [32] . Its analysis demonstrated a kinetic preference for modified protomers in the assembly of immature ( VP0 containing ) pentamers , suggesting that the myristoyl moiety on VP0 might drive the protomer to pentamer transition . Remarkably , unmodified pentamers were excluded from mature virions but not from procapsids . At variance with these results , mutant PV1 with a substitution of proline either at position 2 or 5 in the consensus GXXXS/T myristoylation signal sequence , while featuring a drastically reduced myristoylation , still yielded an appreciable number of progeny virions [23] . They lacked infectivity , which was interpreted as ( additional ) requirement for VP4 myristoylation in the early steps of cell entry , possibly uncoating , as previously proposed from a structural analysis of PV1 [7] . However , an effect of the proline by itself has not been excluded by these researchers . In cardioviruses [33 , 34] , aphthoviruses [35] , and even some rhinoviruses of the Enterovirus genus [36] the N-terminal part of VP4 including the attached myristate are disordered . At physiological temperatures , capsids of picornaviruses tend to “breath” , allowing partial exit of the normally internal VP4 , which includes the N-terminal myristate [37 , 38] . This indicates a considerable degree of plasticity and rather weak interactions of this lipid with internal structures [39] , casting some doubt on a prominent capsid stabilizing function . Apart from targeting mostly PV1 , a caveat of these virus mutant studies is that replacement of amino acid residues may per se interfere with assembly and thus compound any effect due to the lack of the myristoyl moiety or of certain lipid-protein interactions . The potential attractiveness of this modification as a drug target with a likely high barrier to the emergence of resistance prompted us to revisit its functional relevance for picornaviruses by orthologous non-mutational approaches . Until now , there is only one preliminary study in this direction , which , interestingly , showed a just moderately decreased infectivity of enterovirus 71 following inhibition of cellular myristoylation by 2-hydroxy myristate ( 2-HMA ) , in parallel with a diminished VP0 scission [40] . Here , by gene knock-out in haploid HAP1 cells , we demonstrate a greater importance of NMT1 as to the NMT2 isozyme . By employing the potent pan-NMT inhibitor DDD85646 we found that viral assembly is only partially compromised despite a severely reduced VP0 myristoylation , though CVB3 progeny particles were markedly enriched in unstable noninfectious provirions; this indicates an important role for N-terminal myristoylation in efficient VP0 scission for virus maturation . In contrast , parecho- and Aichi virus , which do not process their VP0 , are drug-insensitive . The low specific infectivity of CVB3 still made in the presence of DDD85646 is most likely due to defective viral RNA transfer through the endosomal membrane . Finally , our study identifies NMTs as suitable drug target for combating a broad spectrum of picornaviruses .
Revisiting the importance of myristoylation in the picornaviral life cycle , we used coxsackievirus B3 ( CVB3 ) as a model and investigated the role of the two N-myristoyltransferase isozymes NMT1 and NMT2 , which are ubiquitously expressed in mammalian cells . To this end we assessed the impact of the gene knockout of each of the two human NMT isozymes on viral infection in the near-haploid fibroblastoid human cell line HAP1 [41] . HsNMT1 and HsNMT2 ( Hs , homo sapiens ) are encoded by single multi-exon genes on human chromosome 17 ( NMT1; Gene ID: 4836 ) and chromosome 10 ( NMT2; Gene ID: 9397 ) , respectively . Three enzymatically active HsNMT1 isoforms of different molecular weight have been described . Two of them may result from alternative splicing or initiation at internal start codons as a consequence of leaky ribosomal scanning ( [11] and references therein ) . By contrast , only a single HsNMT2 protein has been found in human tissue [12] . To abolish expression of all reported HsNMT1 isoforms we created a knockout ( KO ) HAP1 cell line with a CRISPR/Cas9-induced frame shift mutation in exon 3 of NMT1 . To eliminate HsNMT2 expression , we similarly introduced a frame shift in the NMT2 exon 1 . Repeated attempts to obtain an NMT1 exon 3 / NMT2 exon 1 double-KO cell line failed attesting to the essential nature of myristoylation for long-term cell growth and survival . The morphology of the two KO clones was indistinguishable from wt HAP1 cells ( S1A Fig ) , although the NMT1 KO cells grew somewhat slower mirroring the effect of siRNA-mediated HsNMT1 knockdown in ovarian cancer cells [13] . Western blotting confirmed the expected lack of the individual isozymes in the respective HAP1 mutants ( S1B Fig ) . As previously reported by others , siRNA knockdown of HsNMT1 in SK-OV-3 human ovarian cancer cells resulted in a 60% increase in HsNMT2 protein [13] . In our HAP1 cells , however , we did not see any compensatory up-regulation of the non-targeted NMT isozyme . Lack of HsNMT1 reduced CVB3 production in the mutant HAP1 cells by 2 . 1 log10 in a single-cycle replication as compared to the parental cells . On the contrary , in HAP1 cells devoid of HsNMT2 , the virus titer attained similar values as in wt cells ( Fig 1 ) . Reconstitution of HsNMT1 expression in about 80% of HAP1 NMT1 KO cells ( S1C Fig ) by transient transfection practically recovered CVB3 production ( S1D Fig ) , indicative of a lack of CRISPR/Cas9 induced off-target effects possibly impacting on virus growth . In summary , the experiments suggested a dominant role of the HsNMT1 isozyme for CVB3 production , at least in HAP1 cells . The NMT1 KO resulted in a substantially decreased production of infectious CVB3 , presumably attributable to diminished myristoylation of the viral proteins . However , other effects indirectly perturbing virus replication related to disruption of non-catalytic protein interactions of NMT1 [42–46] and/or ensuing from overall changes of the cellular myristoylome comprising >100 N-myristoylated proteins [47] cannot be excluded . We thus tested DDD85646 for its impact on virus yield . This pyrazole sulfonamide derivative ( Fig 2A ) inhibits N-myristoylation of trypanosomal proteins [48] and was subsequently validated as a potent HsNMT1/2 inhibitor [47] . The drug competes with peptide substrates with a 50% inhibitory concentration ( IC50 ) of ~17 nM ( HsNMT1 ) and ~22 nM ( HsNMT2 ) as determined in vitro [49] . For comparison , we used 2-hydroxy myristoic acid ( 2-HMA; Fig 2B ) . After cellular uptake , this myristic acid analogue is metabolically converted to the CoA-thioester , which is a potent competitive inhibitor of protein myristoylation ( in vitro IC50 ~50 nM ) [50] . 2-HMA was recently found to decrease the titer of enterovirus-71 ( EV-71 ) by about 10-fold at 250–500 μM concentration [40] . In order to rule out that the reduction in virus titer results from a generalized cytotoxic effect of the NMT inhibitors on the cells , we determined HeLa cell viability with the XTT assay over 1 to 3 days of drug treatment ( Fig 2C ) . After 24 h , viability slightly decreased to ~80% at 5 to 10 μM DDD85646 . The extrapolated 50% cell cytotoxic concentration ( CC50 ) was about 230 μM . Later , cell death increased , thus lowering the CC50 to 0 . 2 μM after 3 days of exposure , essentially as found in a myristoylome profiling study , using this NMT inhibitor [47] . The increase in drug-induced cytotoxicity from day 1 onward likely reflects the turnover of the endogenous pool of myristoylated proteins ( e . g . c-Src tyrosine kinase ) that become progressively replaced by non-myristoylated counterparts with compromised function [51] . As shown in Fig 2D , 2-HMA was not markedly cytotoxic even after 2 days of incubation , likely due to incomplete inhibition of cellular myristoylation by this compound even at highest achievable doses [52] . HeLa cells were then infected with CVB3 in the absence or presence of increasing concentrations of DDD85646 ( up to 5 μM where > 80% of uninfected HeLa cells were still viable ) or 2-HMA ( up to 100 μM ) added 1 h post infection ( p . i . ) and virus production was measured 7 h p . i . As shown in Fig 2E , DDD85646 reduced the yield of infectious CVB3 progeny in a dose-dependent manner by 4 . 5 log10 at 5 μM relative to the DMSO-treated solvent control . The 50% effective concentration ( EC50 ) was clearly below 0 . 5 μM , the lowest concentration tested , which still reduced the titer by 2 . 7 logs . 2-HMA was also effective ( EC50 = 7 . 5 μM ) , but the highest achievable reduction of virus titer was only 2 . 5 log10 at 20 μM; above this concentration its inhibitory activity quickly plateaued ( Fig 2F ) . The relatively low solubility of 2-HMA paired with a requirement for intracellular conversion to the coenzyme A derivative presumably limits its antiviral effect , as also observed by others [40 , 53] . Addition of MA fully relieved the effect of 2-HMA ( S2B Fig ) but not of DDD85646 in CVB3-infected HeLa cells ( S2A Fig ) as expected from the NMT´s sequentially ordered bi-bi mechanism [54] . The observed inhibitory properties of these drugs thus agree well with previous in vitro results for trypanosomal NMT [48] . At 5 μM , DDD85646 was also active against CVB3 in various other cell lines with two of them frequently employed as models for organs naturally infected by this virus ( Caco-2 for intestinal tract enterocytes , A549 for lung epithelial cells; Table 1 ) . In each instance , cell viability was greater than 75% as determined by the XTT assay ( S3 Fig ) . Altogether , this confirms the essential role of HsNMT in CVB3 infectious progeny production . At 5 μM , DDD85646 was also active against four other enteroviruses tested ( Table 1 ) : coxsackievirus B1 ( CVB1 ) , representatives of the 3 different rhinovirus species ( RV-A2 , RV-B14 , and RV-C15 ) , as well as against mengovirus , a member of the Cardiovirus genus . The drug decreased virus titers by 5 . 5 log10 ( RV-B14 ) to 2 . 0 log10 ( CVB1 ) compared to the DMSO control . Strikingly , human parechovirus 1 ( HPeV-1 , formerly echovirus 22 , genus Parechovirus ) and Aichi virus 1 ( AiV-1 , genus Kobuvirus ) were completely insensitive to drug treatment . In both viruses , VP0 remains uncleaved , resulting in their native capsid being built from 60 copies of VP0 , VP3 and VP1 [6] . The above results thus suggest that myristoylation is dispensable for the infection cycle of those picornaviruses that lack maturation cleavage of VP0 . We next verified whether DDD85646 treatment curtailed myristoylation of the structural proteins VP0/VP4 of CVB3 . We utilized metabolic tagging with the Alk-12 myristic acid analogue followed by bioorthogonal ligation of the terminal alkyne moiety to the azido group of the fluorescent reporter 5-TAMRA-azide via CuI-catalyzed azide-alkyne cycloaddition ( CuAAC , the so-called “click” reaction ) for subsequent in-gel visualization ( outlined in Fig 3A ) . A similar strategy was recently employed in a proteome-wide analysis of myristoylated proteins in herpes simplex virus-infected cells [55] . First , we confirmed that Alk-12 became attached to a variety of HeLa proteins in a NMT-controlled manner by the DDD85646 dose-dependent reduction of their in-gel fluorescence over a 24 h treatment period ( S4A Fig ) . Incorporation of Alk-12 was practically abolished by the drug at 5 μM . Importantly , incorporation of the clickable methionine analogue L-azido-homoalanine demonstrated that endogenous protein synthesis proceeded unperturbed up to the highest drug concentration tested ( i . e . 10 μM; S4B Fig ) . This fits well with the only slightly reduced metabolic activity determined in the XTT assay ( see Fig 2C , 24 h exposure to DDD85646 ) . We then infected HeLa cells with CVB3 at a multiplicity of infection ( MOI ) of 10 in absence or presence of 5 μM DDD85646 ( the concentration , which resulted in practically complete inhibition of infectious virus production ) and pulse-labeled with Alk-12 at 4 h p . i . At that time host cell shut-off is already well developed , which should result in preferential incorporation of the alkynyl-myristic acid analogue into viral proteins . Cells were harvested at 6 h p . i . and metabolically tagged proteins visualized by in-gel fluorescence ( Fig 3B ) . In extracts from CVB3-infected cells , a strongly fluorescent band was detected at about 36 kDa that most likely corresponded to VP0 with covalently linked Alk-12 . This band almost disappeared when the NMT inhibitor was present during viral infection ( compare the two rightmost lanes in Fig 3B ) , and , accordingly , it was not detected when the cells were sham-infected for control either with or without inhibitor ( first two lanes in Fig 3B ) ; in the latter case incorporation of Alk-12 into a cellular protein migrating at about 50 kDa was strongly reduced by DDD85646 . By densitometric analysis , the fluorescent VP0 signal decreased ~100-fold at 5 μM of the drug . The small Alk-12 modified VP4 ( 7 . 4 kDa ) was not visible due to a substantial fluorescent background obscuring bands in the low molecular weight range ( S4C Fig ) . Regardless of this technical problem , our results clearly demonstrate that DDD85646 effectively inhibits incorporation of Alk-12 into VP0 at the same concentrations that practically eliminate CVB3 infectivity , suggesting a cause-effect relationship . Consequently , we used a 5 μM drug concentration in all subsequent experiments to efficiently suppress myristoylation with minimal impact on cell viability over the testing period . HsNMTs may exert a catalytic function in CVB3 multiplication beyond modification of the viral polyprotein . After exclusion of a direct virucidal effect of DDD85646 ( S5 Fig ) , we performed a time-of-drug addition experiment to determine the stage of the virus life cycle at which NMT activity was of importance ( Fig 4A ) . Addition of DDD85646 2 h prior to challenging HeLa cells with CVB3 resulted in a greatly decreased virus titer . Virus production was equally inhibited when the drug was added 2 h p . i . , indicating that a severely diminished HsNMT activity had minimal influence on attachment , entry , and uncoating . The inhibitory effect was slightly reduced when DDD85646 was provided 3 h p . i . , matching that of guanidine hydrochloride ( GuHCl ) , a potent inhibitor of RNA synthesis of many picornaviruses [56] , added at the beginning of infection . A modest antiviral activity was still observed when virus multiplication was allowed to proceed for 5 h before drug addition; DDD85646 became ineffective when supplied at 6 h p . i . , i . e . shortly before one replication cycle has been completed . For B type coxsackieviruses , the switch from cellular to viral protein translation usually occurs 2 h p . i and viral protein synthesis levels off at about 5 h p . i . [57] . This agrees well with the period where HsNMT activity is important for virus multiplication as revealed by the time-of-drug addition experiment . We conclude that the role of HsNMT1/2 in the CVB3 infection cycle is confined to the synthetic phase , likely to maintain co-translational myristoylation of the picornaviral polyprotein . To determine whether DDD85646 interferes with the replication of the viral genome , we used the recombinant coxsackievirus RLuc-CVB3 , which expresses Renilla luciferase as sensitive indicator for viral RNA replication [58] . In contrast to GuHCl treated control cells , a strong luminescence was recorded at 7 h p . i . , regardless of addition of DDD85646 at the time of virus challenge ( Fig 4B ) . This indicated unperturbed viral RNA production and hence a correct processing of the non-structural precursor proteins P2 and P3 . To independently corroborate the data obtained with RLuc-CVB3 , we assessed replication of wt CVB3 in DDD85646-treated and control HeLa cells by RT-qPCR of newly synthesized plus-strand RNA . No significant difference in RNA replication kinetics was found in a single-cycle experiment ( Fig 4C ) . From the normal viral RNA synthesis in presence of DDD85646 , we infer that reduction of VP0/VP4 myristoylation blocks step ( s ) required for initiation of a second cycle of infection . To substantiate this reasoning , we employed recombinant CVB3-eGFP that expresses a fluorescent reporter for easy detection of infected host cells . At an MOI of 0 . 1 , a similar small number of cells became infected at 5 h p . i . regardless of the presence of DDD85646 and , as judged from the fluorescence intensities , the viral replication rates were also similar . However , virus spread monitored at 24 h p . i . was strongly reduced in drug-treated cells when compared to the DMSO solvent control ( Fig 5A ) . While most control HeLa cells showed a cytopathic effect ( CPE ) , i . e . rounding and detachment from the support following the 24 h multi-cycle infection , a large number of cells were protected from CVB3-eGFP infection when DDD85646 was present . To verify that these surviving cells did not become intrinsically less supportive for virus replication as a result of an up-regulated innate immune response [59] or drug-induced changes , such as arrest in cell cycle [51 , 60] , we re-infected them with wt CVB3 at an MOI of 1 , but this time in complete absence of the NMT inhibitor . Twenty four hours later , these cells showed massive CPE , ruling out that cell-intrinsic alterations prevented cell-to-cell transmission ( Fig 5B ) . From the above experiment , we deduce that cell-to-cell transmission of CVB3 grown in the presence of DDD85646 is impaired . This might be due to aberrant processing of the P1 precursor into capsid proteins and/or their perturbed assembly , as previously reported for myristoylation-deficient poliovirus mutants [18–24] . Western blot analysis of cell extracts from CVB3-infected HeLa cells prepared 6 h p . i . showed similar expression levels of VP1 and VP3 in absence and presence of DDD85464 ( Fig 6A , two uppermost panels ) . Therefore , the drastic inhibition of HsNMT1/2 did not affect 3CDpro-mediated cleavage of the structural precursor P1 . By contrast , maturation cleavage of VP0 was inefficient , as indicated by the depletion of VP2 ( Fig 6A , third panel ) . A strong VP0 signal was also present in the lysate from CVB3-infected untreated HeLa cells , which stems largely from the substantial amounts of immature 5S protomers normally produced in infected cells . The result suggested that virus morphogenesis was hampered since in enteroviruses cleavage of VP0 is intimately linked to RNA encapsidation [6] . It was also consistent with the about tenfold lower yield ( i . e . 1 log10 ) of encapsidated ( SuperNuclease-protected ) viral RNA copies recovered at 7 h p . i . , measured by RT-qPCR ( Fig 6B ) . Strikingly , the titer of the virus produced in presence of DDD85646 deceased about 5 , 000 fold ( i . e . by 3 . 7 log10 , Fig 6C , despite unaltered infectivity ( in PFU/μg RNA ) of the capsid-extracted viral genome following transfection into HeLa cells to by-pass receptor-mediated entry/uncoating steps , a method frequently employed to assess its integrity ( e . g . [61 , 62] ) ( S6 Fig ) . Taken together , these data document that restricting myristoylation of CVB3 resulted in about 90% less viral progeny , which had an approx . 500 fold lower specific infectivity ( i . e . the ratio of infectious particles measured as PFU/ml and physical particles measured as genomes/ml ) when compared to virus grown under control conditions ( Fig 6D ) . Binding of viral progeny collected from drug-treated and DMSO-treated cells to HeLa cell was similar ( S7A Fig ) . Monoclonal antibody-mediated blockage furthermore demonstrated virus attachment to both , coxsackie-adenovirus receptor ( CAR ) and decay-accelerating factor ( DAF ) displayed on these cells , in accordance with previous results on CVB3 variants [63–66] , and no gross difference in the inhibitory pattern was observed ( S7B and S7C Fig ) . These results attest an overall unchanged 3D structure of CVB3 , when myristoylation of viral structural proteins VP0/VP4 is severely diminished , though subtle alterations of the particle inconsequential for cell attachment cannot be excluded . As inhibition of HsNMTs with DDD85646 led to an about tenfold lower yield of full ( i . e . CVB3 RNA-containing ) virus particles , we next examined the step ( s ) at which virus assembly was critically blocked . Lysates of the same number of HeLa cells infected with CVB3 in presence or absence of the drug were prepared after one infection cycle ( 7 h p . i . ) and 5S protomers and 14S pentamers fractionated on a 5–25% ( w/v ) sucrose density gradient . Every second fraction of the gradient was examined in a Western blot developed with VP1 and VP0/VP2 specific antibodies to reveal a possible variation in the steady-state levels of these early assembly intermediates . Control lysates exhibited a rather narrow peak of 5S protomers close to the top of the gradient and only trace amounts of 14S pentamers at steady-state . Larger virus structures were pelleted under these conditions . Based on the high proportion of VP2 over VP0 the pellet seemed to be largely composed of mature 150S virions ( Fig 7A , upper two panels ) . Lysates from NMT inhibitor treated cells showed a considerably broader peak for the viral protomers , extending somewhat beyond the 5S value , while the level of 14S pentamers appeared just slightly elevated compared to the control lysate ( Fig 7A , lower two panels ) . The immunoblot of the pellet comprising the large viral particles still made in presence of the NMT inhibitor showed some VP0 , while VP2 was practically absent . This suggested that mostly immature viral particles such as 150S provirions and 75S procapsids had been pelleted . The amount of VP1 reporting on the total number of virus particles was furthermore reduced in the drug-exposed sample by about 10-fold in comparison to the control lysate as determined by densitometry , paralleling the decrease in genomes/ml following DDD85646-treatment ( Fig 6B ) . Further resolution and immunoprobing of these large structures by re-centrifugation on a steeper sucrose gradient was unsuccessful due to the limiting amount of material . Accumulation of largely immature and hence non-infectious , viral RNA-containing particles such as provirions would most probably explain the observed sharp drop in specific infectivity ( Fig 6D ) . To substantiate the above findings , in a separate experiment we further characterized the larger CVB3 assemblies formed in presence of the NMT-inhibitor . Cell lysates were prepared 7 h p . i . from DDD85646 or DMSO-treated CVB3-infected HeLa cells . The lower amount of RNA-containing viral particles ( in genomes/ml ) obtained in presence of the drug was compensated for by using ten times more sample than for control conditions . Viral assemblies sedimenting at >14S were initially separated from smaller material by pelleting through a 30% ( w/v ) sucrose cushion . The pellet was resuspended and viral particles further resolved by ultracentrifugation in a 10–30% ( w/v ) sucrose density gradient . Purified RV-A2 virions ( 150S ) and subviral particles produced by heating ( 80S ) were used in parallel as sedimentation markers ( S8 Fig ) . Fractions were collected from the top and probed for VP1 and VP0/VP2 by Western blotting with specific antibodies as above . As expected , the peak at 150S seen for the control CVB3 sample consisted of mostly mature virions as judged from the presence of VP1 and VP2 together with minute amounts of VP0 ( Fig 7B , upper two panels ) . Strikingly , the 150S peak of particles made in presence of DDD85646 showed , besides VP1 , a massive increase of VP0 compared to the solvent control , resulting in a VP0 to VP2 ratio of roughly 5 to 1 ( Fig 7B , lower two panels ) . This fraction most likely comprised either a mixture of provirions with a small amount of mature virus particles or just provirions with about 20% of their VP0 cleaved into VP4 and VP2 . We wish to note that others also observed co-sedimentation of enterovirus provirions and virions ( at 150S ) , in contrast to earlier reports [67–72] and conclude that impaired myristoylation profoundly affects the cleavage of VP0 that normally converts provirions into infectious virions concomitant with RNA encapsidation . The lack of detectable viral material in the light gradient fractions obtained from the control sample suggests that the prior centrifugation through the sucrose cushion efficiently removed 5S and 14S early assemblies ( Fig 7B , upper panels ) . However , despite this pre-fractionation step , we still detected substantial amounts of VP0 and VP1 together with little VP2 in the top fractions of the gradient with the sample of the DDD85646-treated cells . It might indicate that these slowly sedimenting 5S/14S intermediates originate from 150S particles that disintegrated during sucrose gradient centrifugation and/or upon the initial pelleting step . Some faster moving material extending up to the 75S peak conceivably represents partially broken virus . This assumption is also in agreement with the reduced level of VP1 in the 150S peak , as compared to VP1 in the analogous peak of the DMSO control , taking into account that ten times more of the sample obtained from the infection in the presence of the drug was applied onto the gradient . Furthermore , the ratio of VP0/VP2 at the top of the gradient was distinctly higher than in the 150S peak . This led us to conclude that virus collected from drug-treated cells rather represents a mixed population mainly composed of fragile provirions [ ( + ) ssRNA- ( VP0VP3VP1 ) 60] and some , more robust , virions [ ( + ) ssRNA- ( VP4VP2VP3VP1 ) 60] as opposed to a single class of uniformly stable RNA-containing 150S particles with a substantial fraction of VP0 in their capsid remaining uncleaved . Particles corresponding to the 75S and 150S peaks ( Fig 7B , boxed fractions next to the respective cartoons ) were further examined by negative stain transmission electron microscopy . The 75S particles of the control CVB3 ( CVB3DMSO ) as well as of the virus grown in the presence of the drug ( CVB3DDD ) appeared as hollow spheres with a dense core indicative for lack of genomic RNA and accumulation of stain , while the respective 150S nucleoprotein particles appeared as bright spheres ( Fig 7C , left-hand and right-hand combined panels ) . Both had an average diameter of approx . 30 nm in accord with published values [15 , 73] . Reflecting their lower amounts ( Fig 7B ) , only few 75S and 150S particles were evident in electron micrographs of the sample from DDD85646-treated cells , however they were indistinguishable from their counterparts produced under control conditions . An N-linked myristoyl group , usually together with an additional targeting signal , is frequently involved in membrane attachment ( as portrayed by the two-signal model; [74] ) . Picornavirus replication complexes ( RCs ) , virion precursors , and perhaps newly formed full particles , localize to the surface of membranous vesicles , which are produced de novo [75–77] . This serves the coordination of ss ( + ) RNA synthesis with encapsidation [78] . According to recent findings , 14S particles , through their VP1 and/or VP3 proteins , interact with the non-structural protein 2C present in RCs . This is thought to facilitate binding of pentamers to nascent genomic RNA produced in close proximity [79] . In the context of the two-signal model , we hypothesized that the weakly hydrophobic myristate attached to VP0 , by integrating into the vesicle´s lipid bilayer , might promote the interaction of intrapentameric VP1/3 with 2C . A perturbed VP0 myristoylation may lead to a decreased 2C-mediated recruitment of 14S precursors to membrane-associated RCs , resulting in inefficient virus mophogenesis . Previous studies examining the role of myristoylation in membrane targeting of polioviral precursors were rather inconclusive [80 , 81] . We find by confocal immunofluorescence ( IF ) microscopy , that abrogation of myristoylation by DDD85646 had no significant effect on the colocalization of 2C and VP1 in perinuclear clusters of fluorescence at 5 h p . i . ( Fig 8 ) . A drawback of the immunofluorescence analysis is that the used enterovirus VP1-specific monoclonal antibody is unable to discriminate between intermediates and final stages of virus assembly . We hence further investigated the role of myristoylation in the morphogenesis of CVB3 at the ultrastructural level . HeLa cells were challenged with the virus in absence ( DMSO control ) or presence of DDD85646 , incubated at 37°C for 7 h and ultrathin sections were prepared for transmission electron microscopy ( TEM ) . One hundred micrographs for each condition were inspected for subcellular changes and formation of morphogenetic structures following infection by CVB3 as previously reported for various enteroviruses [82–85] . Non-infected HeLa cells in the absence and presence of DDD85646 displayed an organellar shape and localization typical of normal cells ( Fig 9A and 9B ) . By contrast , CVB3-infected cells became rounded and revealed ultrastructural alterations associated with apoptosis , known to be triggered by this virus in HeLa cells [86] . The cell nucleus ( N ) was shrunken , the nuclear membrane ( Nm ) fenestrated at multiple sites , and the chromatin ( Ch ) condensed; most mitochondria ( M ) showed normal morphology in contrast to the endoplasmic reticulum ( ER ) , which was dilated and marginalized . The cytosol was filled with numerous double-membrane vesicles ( DMVs ) , on which RCs are anchored [87] ( Fig 9C and 9D ) . Several multilamellar vesicles ( MMV ) were also apparent ( Fig 9 , insets C2 and D2 ) . These virus-induced membrane structures , possibly of autophagic origin [88] , closely resemble those described in a recent electron tomography study of CVB3-infected Vero E6 cells [89]; they were not markedly affected by the drug . Essentially the same number ( 76% and 78% ) of control and drug treated cells ( S9A Fig ) showed about 1 to 2 irregular virus-related aggregates ( virus blebs , Vb ) per cell ( Fig 9 , insets C3 and D1 ) . These were usually in close proximity to DMVs , contained mostly amorphous material lacking a clear border , and occasionally included round uniformly dense objects with the size of virus particles . Such aggregates likely correspond to the electron-dense material believed to represent assembly precursors and newly formed ( pro ) virions [84 , 90]; they were unaltered in the presence of DDD85646 . Remarkably , paracrystalline arrays ( Ar ) , which often appear at this rather late stage of picornavirus infection , were detected almost exclusively in infected control cells ( Fig 9C ) . Usually , 2–3 arrays consisting of several rows of uniformly dense particles of about 28 nm diameter were present per cell ( Fig 9 , inset C1 ) . Cytoplasmic protrusions of apoptotic cells filled with elaborate particle clusters were also evident only in untreated infected cells ( S9B Fig ) . Similar paracrystalline virus aggregates were previously reported for CVB5-infected HeLa cells [85] . They represent ordered , rather closely packed progeny virions and form when the virus attains sufficiently high intracellular concentrations . The particles in these lattice-like structures are somewhat smaller than determined by X-ray crystallography for purified virus ( 28 nm vs . 30 nm ) due to cell shrinkage during preparation . The above data , combined with the VP1/2C immunofluorescence colocalization study led us to conclude , that the massively reduced VP0 myristoylation does not visibly alter early precursor recruitment to still intact replication complexes . The lack of virus arrays must thus be related to a substantially diminished number of CVB3 progeny produced at these sites in presence of DDD85646 , too low for their ordered intracellular aggregation , in keeping with the several-fold decrease in CVB3 particles recovered from cells lysates as determined by RT-qPCR .
To overcome the limitations of previous myr-deficient PV1 mutant studies , we here used gene knockout and pharmacological inhibition of cellular N-myristoyltransferases ( NMTs ) to explore the effect of radically diminished N-terminal myristoylation on the replication of coxsackievirus B3 ( CVB3 ) and other representatives of the genus Enteroviruses and to evaluate the usefulness of NMT as a potential drug target . For the purpose of comparison we focused on CVB3 , as its myristoyl group linked to VP4 displays interactions at the inner side of the capsid almost superimposable to PV1 [15] . A recent study exploring the translation initiation in CVB3 incidentally discovered an essential role for the myristoylated Gly 2 in infectivity , where its mutation to Arg resulted in abnormal polyprotein processing in vitro , but other possible defects have not been analyzed [91] . In mammalian cells , myristoylation is carried out by the two almost ubiquitously expressed NMT1 and NMT2 isozymes with overlapping substrate specificity , though their individual pathophysiological role is still quite elusive [11] . Using HAP1 NMT1 and NMT2 gene knockout clones we found that the HsNMT1 isozyme was of prime importance for CVB3 production in these cells , while gene ablation of HsNMT2 had little impact ( Fig 1 ) . A similar conclusion was very recently drawn for enterovirus 71 upon siRNA-mediated knockdown of the individual HsNMT isozymes in RD cells [40] and large-scale siRNA screens for host factors involved in HIV-1 and HBV replication also showed their critical dependence on HsNMT1 [92 , 93] . A potentially differential involvement of the three HsNMT1 isoforms ( NMT1L , NMT1M and NMT1S [94] ) as described for HIV-1 [95] warrants further investigation . To exclude the accidental elimination of nonenzymatic ( e . g . scaffolding function ( s ) ) by the NMT gene knockout , we employed DDD85646 [48] ( Fig 2A ) as a potent non-selective inhibitor of the human HsNMT1/2 isoforms with a high on-target mode of action in HeLa cells [47] . Orthogonal chemistry paired with in-gel fluorescence analysis demonstrated that the drug effectively decreased metabolic incorporation of the myristic acid analogue Alk-12 into VP0 in CVB3-infected cells down to at least 1% of the control ( Fig 3B ) , being accompanied by a drastic reduction of infectious virus yield by more than 4 log10 ( Fig 2E ) . Neither CVB3 entry , nor replication or 3C ( D ) pro-mediated processing of capsid precursor protein P1 were affected ( Figs 4A–4C and 6A ) . Previous in vitro translation experiments by others reported impaired polyprotein proteolysis by 3C ( D ) pro for PV1 and CVB3 myr-null mutants [20 , 21 , 91] . However , a subsequent in vivo analysis of myr-deficient mutant PV1 did not confirm these findings [22 , 32] , in line with our data on CVB3 ( Fig 6A ) . We found instead that CVB3 morphogenesis was clearly perturbed , as the number of RNA-filled particles in lysates of DDD85646-treated infected cells dropped by 90% relative to control values ( Fig 6B ) . Fractionation of progeny virus and assembly intermediates on sucrose density gradients in combination with TEM confirmed the formation of a lower number of 150S particles as well as 75S empty procapsids , ultrastructurally indistinguishable from the virus control ( Fig 7B and 7C ) . The pronounced accumulation of CVB3 5S protomers in presence of DDD85646 ( Fig 7A ) indicated , that the myristoyl group attached to VP0 is required for their efficient assembly into 14S precursors . However , the five-stranded β-tube built by the N-termini of VP3 packing around the icosahedral 5-fold axes , which can form only on pentamer assembly [96] apparently suffices for nucleation and/or moderate stabilization of this essential early intermediate as previously predicted [97] , evidently allowing its further transition into higher order structures . Our finding agrees well with a past study by Moscufo et al . , demonstrating a kinetic advantage of pentamer formation from myr-modified protomers over unmodified protomers made by a viable PV1 mutant with 50% reduced VP0 myristoylation [32] . Interestingly , for FMDV , either defined , but aberrantly sedimenting 17S structures or properly assembled 75S virus-like particles ( VLPs , equivalent to empty procapsids ) were obtained in vitro with recombinant protomers featuring an unmodified Gly 2 [26–28] . The reason for these disparate results has not yet been addressed . Notably , the myristoyl group including the linked N-terminus of VP0 are disordered in the empty PV1 procapsid [96] and presumably also in the 14S precursors , with which it is in a dynamic equilibrium [6] . To what extent the stabilizing interactions of the myristate chains established in the mature virus capsid are also maintained in immature pentamers is thus elusive due to the lack of clear structural evidence . Substitution of Thr 28 in VP4 by Val to abolish its hydrogen bond with the myristoyl group from an adjacent VP4 shown in the X-ray structure of PV1 did not prevent formation of 14S particles , though assembly defects at later stages were observed [17] . Future studies are required to comprehensively determine the myristoylation-dependent interactions , which contribute to efficient protomer association and/or stability of the pentameric intermediate . Contrary to CVB3 grown under conditions of severely inhibited VP0 myristoylation , PV1 mutants with a Gly 2 to Ala or Arg substitution shown to fully abolish this VP0 modification did not yield any 150S particles . Curiously , these mutants resulted in distinct assembly phenotypes in different labs , giving rise to just 5S particles [24] , 5S and some 14S particles [22] , or even 80S empty procapsids [19] , possibly due to the different experimental approaches [30 , 31] . However , it cannot be completely excluded , that assembly and RNA encapsidation may have also been adversely affected from substitution of the N-terminal glycine in VP0 by a side chain-bearing , less flexible amino acid residue , in addition to elimination of the myristoyl group . Our results with the myr-depleted CVB3 are remarkably similar to those previously obtained with a PV1 mutant featuring a replacement of Ala 3 or Ser 6 by Pro in the N-terminal myristoylation motif . Despite an almost undetectable modification of VP0 [21] , the level of 14S pentamers was only marginally reduced . Also , practically noninfectious virus particles sedimenting at 150S were still detected at 5- to 10-fold reduced amounts relative to wild type PV1 [22 , 23] . Whether the decreased virus formation observed in these two instances in some way still depends on the very slight VP0 myristoylation is a subject for further investigations . Recently it has been demonstrated that the non-structural protein 2C in the membrane-bound replication complex ( RC ) interacts with VP1 and/or VP3 possibly at the level of pentamers , thought to be crucial for progeny RNA encapsidation [79 , 98] . Our confocal immunofluorescence ( IF ) analysis showed that the co-localization of 2C with VP1 of CVB3 in perinuclear membranes was not significantly altered in presence of DDD85646 ( Fig 8 ) , This finding rules out a prominent role of VP0 myristoylation in targeting of 14S precursors to the membrane-associated replication complex ( RC ) , in agreement with a previous cell fractionation study examining the membrane association of wt and myr-deficient PV1 assembly intermediates [80] and is further corroborated by electron microscopy discussed below . Of note , in a very similar IF analysis , Thibaut et al . recently reported a disrupted co-localization of 2C with VP1 of CVB3 ( Nancy ) grown in glutathione ( GSH ) depleted cells . The protomers made under these conditions were conformationally altered and unable to assemble into pentamers [58] . Given the disparate ( co ) localization results , an ( indirect ) interaction between the myristoyl group of VP0 and GSH in stabilizing these precursors seems unlikely . Ultrastructure analysis clearly showed that dense aggregates ( virus blebs ) in close proximity to mostly DMVs , proposed to be the sites of progeny particle formation [75 , 82 , 84 , 85 , 99] , remained unaltered compared to CVB3-infected control cells ( Fig 9 ) . Strikingly , virus paracrystalline arrays , which only form at high intracellular concentrations of progeny virus , were practically absent in infected , DDD85646-treated HeLa cells , consistent with the substantially lower number of full particles as determined by RT-qPCR . Since this became already evident in situ by the virtual absence of ordered viral aggregates , it strongly suggests that mechanical disruption of ( pro ) virions during freeze-thawing for progeny release contributed only little if at all to the drastically diminished yield . Despite the less efficient transition of protomers made in presence of DDD85646 into 14S particles ( Fig 7A ) , the above observations indicate that these early intermediates are not markedly depleted from the spatially still properly organized viral replication sites . This implicates a reduced flux also through the late morphogenetic pathway , most likely resulting from impaired encapsidation of the nascent ( + ) ssRNA by the myr-deficient 14S pentamers and/or a heightened tendency of partially formed provirions to spontaneously disassemble . In this context , the exclusive formation of completely myristoylated virions in a viable PV1 Ala 3 to Asp mutant with 50% reduced VP0 myristoylation [32] could arguably result from a profoundly higher efficacy of properly modified pentamers in condensation of the viral RNA genome . Also in accord with a defect in RNA encapsidation , unmodified protomers derived from a myr-deficient P1 precursor-expressing recombinant vaccinia virus were only inefficiently incorporated into virions of a coinfecting wild type PV1 [18] . A considerable fraction of CVB3 particles still made in presence of DDD85646 and sedimenting at 150S was in fact quite fragile and ( partially ) disintegrated during ultracentrifugation ( Fig 7B , CVB3DDD ) . As inferred from the VP0/VP2 stoichiometry the 150S particles furthermore seemed to consist mostly of provirions . From their delicate nature they intriguingly resembled the highly unstable PV1 provirions obtained by mutation of VP2 His 195 , required for the maturation cleavage of VP0 [100] . Thus , a key finding is , that the autocatalytic VP0 scission , which depends on viral RNA encapsidation and is crucial for virion stability and infectivity [71 , 100 , 101] , appears to be critically regulated by myristoylation of VP0 . We note that such hypothesis was already previously proposed by Marc et al . [23] based on compositional analysis of their myr-deficient PV1 mutants , and is now strongly supported by our non-mutational approach . Though quite speculative , the absence of the N-terminal C14:0 group may trigger a conformational change , which propagates along the backbone of VP0 , resulting in distant structural alterations affecting the proper disposition of the maturation cleavage site ( encompassing the C-terminus of VP4 and N-terminus of VP2 ) . This would reduce but not fully prevent its scission by the nearby catalytic center comprising His 195 , nucleophilic water and likely a base of the encapsidated RNA [100 , 102] to account for the mature virions still produced in presence of DDD85646 . A similar scenario was recently put forward to explain impaired cleavage of VP0 by mutations of Ala 107 on the outer surface of EV71 VP1 [103] . The mature CVB3 particles made in DDD85646-treated cells , if fully infectious , should theoretically suffer a just 6-fold loss in specific infectivity ( the ratio of PFU to encapsidated genomes ) due to their dilution by a 5-fold excess of viral RNA containing provirions , considered noninfectious [71 , 101] . However , their specific infectivity instead strikingly decreased by several hundred-fold ( Fig 6D ) and cell-to-cell spreading was dramatically reduced compared to normal virus ( Fig 5A ) . Thus only a very small proportion of these particles were able to productively release their genome into the cells . Their encapsidated RNA was still intact ( S6 Fig ) and neither attachment to the co-receptor DAF nor to CAR used for internalization [104] was substantially affected ( S7A–S7C Fig ) . Moderate heating as an accepted surrogate for receptor-mediated uncoating of enteroviruses [105] furthermore led to complete genome release from these CVB3 particles as determined by RT-qPCR ( S10 Fig ) . The observed block in infectivity must therefore reside in an early post-attachment step similarly as proposed by Marc et al [23] for their very weakly myristoylated PV1 mutants of greatly reduced infectivity . These in vivo results , now furnished by a virus comprising a native ( non-mutated ) myr-deficient VP4 can be best explained by the current model of enterovirus uncoating . It involves binding of the virion to the host cell , internalization , followed by viral receptor and/or a low endosomal pH-triggered conformational changes of the capsid proteins resulting in an expanded 135S A-particle , which possibly asymmetrically [106] externalizes the hidden VP1 N-termini and expels VP4 to form a pore for translocation of the genomic RNA into the cytosol ( [107] and references therein ) . Formation of such an intermediate or infectosome is considered impossible when VP4 is still part of VP0 in provirions [71] . Myristoylation of recombinant VP4 furthermore noticeably facilitated pore formation in model membranes [108] . The drastically diminished specific infectivity of CVB3 ( made in presence of DDD85646 ) is hence most likely attributable to the on average very low number of myristyolated VP4 molecules , largely insufficient for breaching the endosomal bilayer . In the enteroviruses , the N-terminal glycine becomes accessible by removal of the initiating methionine through the action of the methionine aminopeptidase , whereas in the cardio- and aphtoviruses it is exposed by proteolytic removal of a leader peptide [6] . Irrespective of these differences , DDD85646 not only potently inhibited various members of the genus Enterovirus but also mengovirus , a cardiovirus . In stark contrast , HPeV-1 ( a parechovirus ) and AiV-1 ( a kobuvirus ) were practically resistant to DDD85646 treatment compared to CVB3 grown in the same cells ( Table 1 ) . At variance with most other representatives of the large Picornaviridae family , the VP0 proteins of parecho- and kobuviruses remain intact in the mature virions . Parechoviruses lack a consensus myristoylation signal in VP0 and are thus most likely not myristoylated [109] explaining their unresponsiveness to NMT inhibition . For HPeV-1 and 3 and Ljungan virus , numerous interactions between the RNA genome and capsid proteins VP1 and VP3 were identified in high-resolution 3D structures , suggesting a role in virion formation and stability [110–113] . These interactions not found in other picornaviruses conceivably rendered lipidation of VP0 evolutionary dispensable for assembly of parechoviruses . In kobuviruses , a large leader peptide initially attached to the P1 precursor is rapidly cleaved off by the viral protease 3Cpro [6] , thereby exposing the N-terminal glycine of AiV-1 VP0 . This glycine is part of a conserved consensus myristoylation motif ( GXXX ( S/T ) ) and was previously described to be N-terminally blocked [114] , most probably by a myristate as indicated by the metabolic incorporation of the myr-analogue Az-12 into VP0 of AiV-1 infected Vero cells ( S11 Fig ) . A 5´ terminal RNA stem-loop and the leader peptide L were found essential for AiV-1 assembly [115 , 116] , again indicating a possible redundancy for VP0 myristoylation in this process . Kobuviruses together with the closely related saliviruses ( genus Klassevirus ) are furthermore singled out from the other picornaviruses by having their nonstructural 3A protein N-terminally myristoylated , despite lacking a consensus myristoylation signal . Interestingly , 3A myristoylation was not essential for AiV-1 replication as was shown by site-directed mutagenesis [117] . The reason for maintenance of a myristoylated VP0 and 3A despite lacking an obvious function in infectious AiV-1 production , as inferred from its insensitivity to DDD85646 , remains to be determined . Though not tested here , we would further predict that hepatitis A virus ( HAV ) of the genus Hepatovirus is also resistant to NMT inhibition , since its VP4 is non-myristoylated [118] . The strong hydrophobicity of this unusually small polypeptide compared to other picornaviruses is sufficient for endosomal pore formation [119] , and 2A as ( temporary ) VP1 appendix appears to take over the myr-moiety´s role in virus capsid assembly [120] . Despite lacking activity against members of a few genera , our data suggest that DDD85646 has promising potential to be developed as a broad-spectrum anti-picornaviral drug . Furthermore , until now we failed to isolate stable drug-resistant CVB3 mutants in dose-escalation experiments . Drugs targeting host-factors such as NMT represent an attractive alternative to direct-acting antivirals , as the latter suffer from the rapid emergence of drug escape variants in RNA viruses , due to the high error rate of the RNA–dependent RNA polymerases [2 , 121] . Taken together we have discovered that acute inhibition of NMT activity by the pan-HsNMT1/2 isozyme inhibitor DDD85646 has a potent anti-picornavirus activity with minimal cytopathic effect on infected cells over several replication cycles . With respect to CVB3 , the almost complete suppression of myristoylation led to a ~10-fold reduction in virus progeny , drastic accumulation of provirons relative to virions , and a profound defect of these particles in RNA release . These seemingly disparate deficits appear to be all connected to insufficient acylation of the viral structural proteins VP0/VP4 , resulting in poor viral RNA encapsidation , attenuated maturation cleavage and likely lack of functional pore formation for cytosolic transfer of the viral genome through the endosome membrane . After finishing revision of our manuscript we realized that a paper has been very recently published by Mousnier et al . , examining the inhibition of RV , PV and FMDV replication by a newly developed NMT inhibitor ( IMP-1088 ) [122] . While there is a certain overlap with our data independently strengthening our conclusions , our analysis substantially extends their findings by demonstrating the pivotal role of NMT1 isozyme in this lipid modification , by showing the importance of myristoylation for the VP0 maturation cleavage and also for RNA uncoating . We furthermore report for the first time the unexpected insensitivity of two genera of picornaviruses ( Parecho- and Kobuviruses ) towards profound NMT inhibition .
All cell culture reagents were from Gibco , except fetal calf serum ( FCS ) , which was from Sigma-Aldrich . NP-40 and Tris were purchased from AppliChem , SDS was from Serva , D-sucrose from Roth; other chemicals were obtained from Sigma-Aldrich unless otherwise stated . DNase I , RNase A and EDTA-free complete protease inhibitor cocktail were from Roche , SuperNuclease was obtained from SinoBiological and InstantBlue from Expedon . RNase A from bovine pancreas was prepared as 100 mg/ml stock solution in 10 mM Tris-HCl , pH 7 . 5 , 15 mM NaCl . DDD85646 ( 2 , 6-Dichloro-4-[2- ( 1-piperazinyl ) -4-pyridinyl]-N- ( 1 , 3 , 5-trimethyl-1H-pyrazol-4-yl ) -benzenesulfonamide , CAS No . 1215010-55-1 ) was synthetized as previously described ( [123] and S12 Fig ) ; it was dissolved in DMSO at 50 mM and kept in aliquots at –20°C . 2-hydroxy-myristate ( 2-HMA ) was obtained from Enzo Life Sciences; a 10 mM stock solution was prepared in DMSO and kept at –20°C . A 1 M stock solution of guanidine hydrochloride ( GuHCl ) was prepared in H2O and kept at 4°C . Alk-12 ( myristic acid alkyne , tetradec-13-ynoic acid ) was from Cayman Chemical , L-azdiohomoalanine ( AHA ) from Invitrogen ( Thermo Fisher Scientific ) , fatty acid free BSA was from Calbiochem , and LDS sample buffer was from Novex ( Thermo Fisher Scientific ) . RIPA buffer ( 50 mM Tris-HCl pH 8 . 0 , 500 mM NaCl , 0 . 1% SDS , 0 . 5% sodium deoxycholate , 1% NP-40 , 1 mM EDTA , protease inhibitors ) , sodium phosphate buffer ( 100 mM sodium phosphate pH 7 . 4 , 0 . 1% Triton-X 100 , protease inhibitors ) , virus buffer ( 20 mM Tris-HCl , pH 7 . 5 , 2 mM MgCl2 ) were made in-house and filter sterilized . The polyclonal rabbit anti-CVB3 VP3 antiserum [124] and rabbit anti-CVB3 2C antiserum [125] were provided by Karin Klingel . The mouse monoclonal anti-enterovirus VP1 antibody was from Dako ( Clone 5-D8/1 ) . The rabbit anti-CVB3 VP2 polyclonal antiserum [126] was kindly donated by Andreas Henke from the Institute of Virology and Antiviral Therapy , Medical Center at the Friedrich-Schiller-University Jena , Germany . The rabbit anti-NMT1 polyclonal antibody was purchased from Gentex and the mouse anti-NMT2 monoclonal antibody was obtained from BD Biosciences . The monoclonal mouse anti-γ-tubulin antibody was from Sigma . HRP-conjugated secondary antibodies were purchased from Jackson ImmunoResearch , and Alexa-Fluor conjugated antibodies were obtained from Thermo Fisher Scientific . For Western blot analysis all antibodies were diluted in PBSTB ( 0 . 1% Tween 20 supplemented with 2% BSA ) in following dilutions: the anti-enterovirus VP1 antibody 1:1 , 000 , anti-CVB3 VP2 1:1 , 000 , anti-CVB3 VP3 1:500 , anti-NMT1 1:5 , 000 , anti-NMT2 1:1 , 000 , anti-tubulin 1:5 , 000 , HRP-secondary antibodies 1:10 , 000 . All cell lines except HAP1 were originally obtained from ATCC . HeLa Ohio were maintained in minimal essential medium ( MEM ) , HEK293 and Vero cells in DMEM , and A549 in RPMI , supplemented with 10% fetal calf serum ( FCS ) , 1% penicillin and streptomycin and 1% L-glutamine ( Pen-Strep-Glu ) . Caco-2 cells were grown in DMEM , 20% FCS with Glu and antibiotics . HAP1 cells were obtained from Horizon Discovery and grown in IMDM supplemented with 10% FCS and 1% Pen-Strep . HeLa-CDHR3/Y529 cells overexpressing the RV-C receptor ( the cadherin-related family member 3 ( CDHR3 ) Y529 variant ) were produced and maintained as described in [127] . All cells were kept in a humidified 5% CO2-containing atmosphere at 37°C . In infection assays , serum concentration was reduced to 2% FCS . For infection with rhinoviruses , the infection medium was additionally supplemented with 30 mM MgCl2 and cells incubated at 34°C . Rhinovirus-2 ( RV-A2 ) , rhinovirus-14 ( RV-B14 ) , coxsackievirus-B1 ( CVB1 ) , mengovirus ( MV ) , and human parechovirus 1 ( HPeV-1 ) were purchased from ATCC . Plasmid p53CB3/T7 , containing the full-length cDNA of coxsackievirus B3 ( CVB3 , strain Nancy ) [128] and plasmid pRLuc-53CB3/T7 , which contains the Renilla luciferase gene upstream of the capsid coding region for production of a Renilla luciferase ( RLuc ) expressing CVB3 ( RLuc-CVB3 ) [129] were a generous gift from Frank Kuppeveld ( Institute of Veterinary Research , Faculty of Veterinary Medicine , University of Utrecht , The Netherlands ) . Plasmid pMKS-1-eGFP encodes an enhanced green fluorescent protein ( eGFP ) -expressing CVB3 ( CVB3-eGFP H3 strain ) [130] . The plasmid pAV-UCSF [117] , carrying a full-length cDNA of Aichi virus 1 ( genus Kobuvirus ) was kindly donated by Joseph L . DeRisi ( University of California San Francisco , USA ) . The plasmid pC15 [131] , a gift from Ann Palmenberg ( Institute for Molecular Virology , University of Wisconsin , USA ) , contains a full-length cDNA copy of the RV-C15 genome . Infectious plasmids were linearized at the 3´-end of the cDNA insert with an appropriate restriction enzyme and viral RNA ( vRNA ) transcribed in vitro using the T7 Ribomax Large Scale RNA Production System ( Promega ) . The RNA was purified by phenol-chloroform-isoamyl alcohol extraction followed by ethanol precipitation , dissolved in RNase-free water ( Quiagen ) supplemented with RNasin ( Promega ) and stored at –80°C until further use . HEK293 cells were transfected with in vitro transcribed vRNA using Lipofectamine 2000 ( Invitrogen ) according to the supplier´s instructions . Two days later , cells were disrupted by three freeze-thaw cycles to release the ( recombinant ) virus and the clarified supernatant was used for preparation of respective virus stocks in HeLa cell monolayers ( HeLa-CDHR3/Y529 cells for RV-C15 and Vero cells for Aichi virus ) . Infectious virus titers were determined by endpoint titration expressed as 50% tissue culture infectious doses ( TCID50 ) per ml [132] , routinely yielding 107–108 TCID50/ml . Virus stocks were aliquoted and kept at –80°C . Cell viability assay was essentially done as described before ( Thinon et al . [47] ) . After 24 , 48 , and 72 h , XTT reagent ( AppliChem ) instead of MTT was added to the cells in the presence of phenazine methosulfate ( PMS ) and A490 and A630 of the colored formazan derivative reductively formed by metabolically active cells were determined 2 h later . A490–A630 was calculated for each measurement and the average value of the puromycin control subtracted from the average value of the other samples . The metabolic activity was calculated as percentage relative to the solvent control ( referred as 100% viability ) . CC50 values were calculated by fitting the data to the IC50 function using GraphPad Prism 6 . 0 ( GraphPad Software ) . The effect of the inhibitor ( DDD85646 , 2-HMA , GuHCl ) in a single cycle infection was determined as reduction of TCID50 with respect to the solvent control . Virus ( CVB3 , CVB1 , RV-A2 , RV-B14 , RV-C15 , Mengovirus , HPeV-1 , AiV-1 ) was added to subconfluent cells ( HeLa , A549 , Caco-2 , Vero ) as specified in Figs 2 , 4 , and 6 and Table 1 , grown in 12-well plates at an MOI of 1 and incubated in appropriate serum-reduced infection medium for 1 h at 37°C ( 34°C for rhinoviruses ) . Non-internalized virus was removed by washing the cells three times with PBS and fresh infection medium with inhibitor was added at concentrations specified in the corresponding figures . Infected cells were subjected to three cycles of freezing-thawing at a time p . i . appropriate for the respective virus ( listed in Table 1 ) , cellular debris was removed by low speed centrifugation and virus titers ( TCID50 ) were determined as above . Cells grown in 12-well tissue culture plates were infected with CVB3 at an MOI of 1 for 1 h at 37°C in infection medium and washed with PBS . DDD85646 was added to a final concentration of 5 μM at the times specified in Fig 4 starting from 2 h prior infection ( –2 ) up to 6 h p . i . ( +6 ) and maintained until the end of the experiment . GuHCl ( 2 mM ) , used as a control for inhibition of virus replication , was added at 1 h p . i . ( +1 h ) . At 7 h p . i . , cells were subjected to 3 freeze-thaw cycles for virus release . TCID50 was then determined as above . HeLa cells grown in 12-well plates were infected with RLuc-CVB3 at an MOI of 1 . Immediately after infection , DDD85646 ( 5 μM ) or GuHCl ( 2 mM ) was added . Addition of DMSO ( 0 . 1% final concentration ) served as solvent control for the NMT inhibitor . At 7 h p . i . , cells were lysed for determination of Renilla luciferase activity using the Renilla Luciferase Assay System ( Promega ) according to the manufacturer´s instructions . Luminescence was detected with the Synergy H1 hybrid multi-mode microplate reader ( BioTek ) . The increase of relative light units ( RLU ) in the absence of GuHCl was taken as measure of RNA replication . HeLa cells were seeded into 12-well plates to reach 80–90% confluency at the day of infection . Cells were inoculated with CVB3 at an MOI of 5 for 30 min at 4°C followed by removal of unbound virus by washing three times with PBS . Cells in one well were immediately lysed in 350 μl RLT buffer for total RNA isolation using the RNeasy Mini kit . The temperature was subsequently raised to 37°C ( at 5% CO2 ) for 30 min to trigger virus internalization into the cells of the other wells . New infection medium containing either 0 . 1% DMSO or 5 μM DDD85646 was then added to the cells and incubation continued up to 6 h . In 1 h intervals , the medium was aspirated from a well of drug- and control-treated cells followed by a gentle wash with PBS . The cells were then directly lysed in RTL buffer as before . Lysates were frozen at –80°C until further use . Total RNA was isolated with the RNeasy Mini kit and viral genomic RNA in each sample quantified by RT-qPCR as outlined below and normalized to the mRNA level of GAPDH determined analogously ( FW: 5’ gaaggtgaaggtcggagt 3’; RW: 5’ gaagatggtgatgggatttc 3’ ) by the ΔΔCT method [133] . HeLa cells were infected with CVB3 at an MOI of 5 for 1 h at 37°C . Unbound virus was removed by washing two times with PBS and fresh infection medium containing DDD85646 ( 5 μM ) or DMSO ( 0 . 1% ) was added . Seven h p . i . cells were subjected to 3 freeze-thaw cycles . Cellular debris was removed by low speed centrifugation and an aliquote of the supernatant was saved for virus titration . To eliminate any unpackaged viral RNA , the remaining virus-containing supernatant was treated with SuperNuclease at 25 U/ml final concentration for 30 min at 37°C . Encapsidated ( SuperNuclease-protected ) genomes were then isolated with Trizol LS ( Invitrogen ) and finally resuspended in RNase-free water . Reverse transcription ( RT ) of 5 μl of this solution was carried out by using Moloney Murine Leukaemia Virus Reverse Transcriptase ( MMLV , Promega ) according to the manufacturer’s instructions using an oligo dT primer . Quantitative real time qPCR was performed on an Eppendorf Realplex 2 Mastercycler with the Kappa Sybr Fast qPCR Master Mix ( Peqlab ) according to the manufacturer’s instructions . Pan entero specific oligonucleotides ( FW: 5’ tcctccggcccctgaatg 3’; RW: 5’ gaaacacggacacccaaagta 3’ ) were used for amplification of the cDNA . The number of genomes ( corresponding to the original quantity of intact , viral RNA-containing particles ) was calculated from a standard curve obtained by qPCR of known amounts of plasmid p53CB3/T7 . PFU/ml were calculated from the TCID50/ml value using the formula 1 PFU/ml = 0 . 7 x TCID50/ml; by normalization to the measured genome copy number/ml the specific infectivity is derived , expressed as the number of PFU per 1010 viral genomes . HeLa cells grown in 6 cm dishes until 90% confluent were incubated with medium alone or with increasing concentrations of DDD85646 ( at 0 . 25 , 0 . 5 , 1 , 5 , and 10 μM ) for 30 min at 37°C and 5% CO2 . For labelling with Alk-12 , the medium was replaced with fresh medium containing 0 to 10 μM of the drug and 22 μM Alk-12 . For labelling with L-AHA , the medium was replaced with Met/Cys-free DMEM supplemented with 2% dialyzed FCS , 1% Pen-Strep with DDD85646 and 20 μM L-AHA . Twenty-four h later , cells were scraped off the plate , pelleted at 500 g in a benchtop centrifuge and the pellet was washed twice with ice-cold PBS . Cells were lysed in 100 μl of 100 mM sodium phosphate buffer and 25 U/ml SuperNuclease for 20 min on ice . To incorporate Alk-12 into viral proteins , HeLa cells were challenged with CVB3 at an MOI of 10 in infection medium for 1 h at 37°C . The inoculum was removed , cells washed with PBS , and incubation continued following addition of infection medium with 5 μM DDD85646 or 0 . 1% DMSO ( control ) . Cells were washed with PBS at 3 . 5 h p . i . and labelling medium ( MEM supplemented with 2% fatty acid free BSA , 1% Pen-Strep-Glu ) containing DMSO or 5 μM DDD85646 was added , and incubation continued for 30 min . Alk-12 was then added to a final concentration of 22 μM . At 6 h p . i . cells were lysed as described above . Cellular debris was removed by centrifugation at 4°C and 20 , 000 g for 25 min in a benchtop centrifuge . Supernatants were snap frozen and kept at –80° until further use . Click chemistry grade Na-ascorbate , Tris ( 3-hydroxypropyltriazolylmethyl ) amine ( THPTA ) , Cu2SO4 , and the fluorescent capturing reagents Cy5 . 5-Alkyne and 5-TAMRA-Azide were from Jena Biosciences . Na-ascorbate and Cu2SO4 were freshly prepared before each experiment as 50 mM solutions in H2O . Stock solutions of THPTA ( 10 mM in water ) and capturing reagents ( 10 mM in DMSO ) were prepared and kept in aliquots at –20°C . Cell lysates were thawed on ice . The protein concentration was determined with the BCA Protein Assay Kit ( Pierce , Thermo Fisher Scientific ) and adjusted to 1 mg/ml with 100 mM sodium phosphate buffer . A click mixture was prepared by adding the components from stocks in the following order with vigorous mixing after each addition; capture reagent Cy5 . 5-Alkyne ( for L-AHA ) or 5-TAMRA-Azide ( for Alk-12 ) ( 1 μl , final concentration 0 . 1 mM ) , Cu2SO4 ( 2 μl , final concentration 1 mM ) , sodium ascorbate ( 2 μl , final concentration 1 mM ) , THPTA ( 1 μl , final concentration 0 . 1 mM ) . Six μl of a master mix was then added to 94 μl of the sample in a non-transparent microcentrifuge tube . Incubation was at 25°C with shaking at 600 rpm for 1 h . Next , 400 μl ice-cold acetone was added , the samples were quickly vortexed , kept at –20°C for 1 h and centrifuged at 14 , 000 g for 30 min to pellet precipitated proteins . The pellets were air-dried , 75 μl of 2% SDS in PBS , 10 mM EDTA was added and samples were vortexed . Once the proteins were completely dissolved , 25 μl of 4x LDS was added ( the final concentration of proteins was ~ 1 mg/ml ) . The samples were incubated at 95°C for 5 min and 10 μg protein each were loaded on a 10% or 15% Tris-Tricine SDS polyacrylamide gel . After electrophoretic separation , the gel was washed three times with MilliQ water and fluorescence was recorded with a Typhoon FLA 9500 laser scanner ( GE Healthcare ) ; equal protein loading was verified by staining with InstantBlue . The same protein samples were resolved by separate SDS-PAGE , transferred to an Immobilon-P PVDF membrane ( EDM Millipore ) and probed with an anti-enterovirus VP1 , anti-CVB3 VP2 , anti-CVB3 VP3 , and anti-γ-tubulin antibody ( as loading control ) as specified further below . HeLa cells seeded in one 15 cm culture plate at ~80% confluency were infected with CVB3 at an MOI of 10 . After 1 h adsorption , unbound virus was washed away and infection medium containing DMSO or 5 μM DDD85646 added . Seven h p . i . cells were collected with trypsin , washed 2x with PBS and lysed in 500 μl of TNE buffer ( 10 mM Tris-HCl pH 7 . 4 , 100 mM NaCl , 1 mM EDTA , 0 , 5% NP-40 , 1x protease inhibitor stock ) for 20 min on ice . Cellular debris was removed by low speed centrifugation , and the resulting supernatant fractionated through a 5–25% ( w/v ) sucrose density gradient ( prepared in TNE buffer ) by centrifugation for 16 . 5 h at 202 , 048 g in a SW40 Ti rotor ( Beckman ) at 4°C ( zero break ) . Fractions ( ~400 μl ) were collected from the top of the gradient and the pellets obtained from this centrifugation were each resuspended in 400 μl TNE buffer . 10 μl of each odd fraction inclusively the dissolved pellet were subject to SDS-PAGE , the proteins transferred to an Immobilon-P PVDF membrane and analyzed as described below . For separation of larger viral assemblies , a 10–30% ( w/v ) sucrose density gradient was used . HeLa cells grown in twenty 15 cm culture plates were infected with CVB3 at MOI of 5 . Following 1 h adsorption , DMSO or DDD85646 ( 5 μM ) was added to the infection medium . Seven h p . i . cells were disrupted by 3 cycles of freezing and thawing , clarified supernatants combined and viral material was pelleted in a SW32 Ti rotor ( Beckman ) through a 30% ( w/v ) sucrose cushion prepared in virus buffer for 3 . 5 h at 118 , 018 g and 10°C . Pellets were resuspended in virus buffer containing DNase I ( Roche ) and RNase A ( Roche ) at a final concentration of 5 μg/ml each , and incubated for 15 min at RT . N-laurlysarcosine was added to a final concentration of 1% ( w/v ) . The samples were kept overnight at 4°C and insoluble material was removed by a 15 min centrifugation at 20 , 000 g , 4°C . Virus from DDD85646-treated cells ( CVB3DDD ) and virus from DMSO treated cells ( CVB3DMSO; 1/10 to compensate for the different yield ) diluted to equal volumes of 0 . 5 ml with virus buffer was each overlaid onto 5 ml of a 10–30% ( w/v ) sucrose gradient made in virus buffer and centrifuged for 30 min in a SW55 Ti rotor ( Beckman ) at 286 , 794 g and 4°C . Two hundred μl aliquots were collected from top to bottom and frozen at –80°C until further use . Ten μl of each fraction were run on a 15% Tris-Tricine SDS polyacrylamide gel and proteins transferred to an Immobilion-P membrane for Western blot analysis . Viral proteins were detected using anti-CVB3 VP2 or anti-enterovirus VP1 specific antibodies as described below . Fractions containing empty or full particles were pooled separately and viral material was pelleted in a TLA100 . 3 rotor ( Beckman ) at 207 , 995 g for 1 h at 4°C . The pellets were resuspended in 30 μl of virus buffer and stored at –80°C until further use . Whole cell extracts were prepared by lysis of cells in RIPA buffer for 30 min on ice and clarified at 20 , 000 g for 5 min at 4°C . Equal amounts of total protein ( 10 to 30 μg ) determined with the BCA Assay Kit were separated on a 10% or 15% Tris-Tricine SDS polyacrylamide gel and transferred to an Immobilion-P PVDF membrane using the Trans-Blot Turbo system ( Bio-Rad ) . The membrane was blocked for 1 h at RT in PBSTB and incubated with primary antibody overnight at 4°C . After washing , the membrane was incubated with HRP-conjugated secondary antibody for 1 h at RT . Bands were revealed by chemiluminescence ( SuperSignal West Pico from Pierce , Thermo Fisher Scientific ) and images were captured with a Chemidoc system ( Bio-Rad ) . The band intensity was quantified using the ImageLab Software ( Biorad ) . Expression of γ-tubulin as loading control was determined by stripping the membrane in glycine buffer ( 2 . 5 mM glycine , 1% SDS , pH 2 . 0 ) for 1 h at RT , followed by rinsing with PBST , re-blocking in PBSTB and probing with a mouse monoclonal anti-γ-tubulin antibody according to the procedure described above . For VP1 and 2C co-localization studies , HeLa cells were grown until subconfluent on a 12 mm coverslip ( Marienfeld ) placed in the well of a 24-well plate . The medium was replaced by infection medium and CVB3 was added at an MOI of 5 . After 1 h at 37°C , virus inoculum was removed and replaced with fresh medium containing 5 μM DDD85646 . DMSO ( 0 . 1% ) was used as solvent control . Five h p . i . cells were washed with PBS , fixed with 4% PFA , and permeabilized with PBS containing 0 . 5% saponin . After blocking in 0 . 5% fish gelatin in PBS for 15 min , cells were incubated with a combination of anti-enterovirus VP1 antibody and anti-CVB3 2C—both diluted 1:200 in blocking solution—for 1 h . Excess antibody was removed by washing with PBS , followed by a 1 h incubation of the cells with appropriate Alexa-Fluor conjugated secondary antibody diluted 1:1 , 000 in blocking buffer . Cell nuclei were counterstained with DAPI . All steps were carried out at room temperature . The samples were examined in a LSM 700 confocal laser scanning microscope ( Zeiss ) , and images were processed with ImageJ . For analysis of virus spread , HeLa cells were grown in 12-well plates and infected with CVB3-eGFP at an MOI of 0 . 1 in infection medium for 1 h at 37°C and 5% CO2 . Unbound virus was removed by washing several times with PBS , fresh infection medium containing 5 μM DDD85646 or 0 . 1% DMSO ( solvent control ) was added and incubation continued . Five and 24 h p . i . cells were examined using an Axio Observer . Z1 epifluorescent microscope . The medium of the surviving cells was replaced with fresh medium without drug , followed by infection with wt CVB3 at an MOI of 1 and analysis for a cytopathic effect ( CPE ) 24 h p . i . by bright field microscopy . For transmission electron microscopy ( TEM ) analysis of purified CVB3 , a small aliquot ( 3–5 μl ) of virus was adsorbed for 1 min to the backside of a carbon-coated 400 mesh copper hexagonal grid that had been glow-discharged just before use . Excess sample solution was blotted away with a filter paper without letting the sample dry . Five μl of 2% uranyl acetate in water was applied to the grid and blotted off immediately . The staining was repeated for 1 min . After a final blotting of excess solution , the grid was air-dried and examined with an FEI Morgagni 268D ( FEI ) operated at 80 kV . Images were acquired using an 11 megapixel Morada CCD camera ( Olympus-SIS ) . In our hands , a concentration of ~ 1x109 virus particles/ml was normally required for proper visualization . For TEM of infected cells , 3x106 HeLa cells were seeded into 6 cm tissue culture dishes and inoculated with CVB3 in infection medium at an MOI of 10 . After adsorption for 1 h at 37°C , cells were rinsed with PBS and fresh medium was added containing 5 μM DDD85646 or 0 . 1% DMSO ( solvent control ) and incubation was continued at 5% CO2 and 37°C . Uninfected HeLa cells were treated identically . At 7 h p . i . cells were scraped from the dish , transferred to a benchtop centrifuge tube and fixed with 2 . 5% glutaraldehyde in 0 . 1 M sodium phosphate buffer , pH 7 . 2 overnight at 4°C . Cells were pelleted at 500 g for 5 min at RT , washed once with PBS , post-fixed in sodium phosphate-buffered 1% osmium tetroxide , dehydrated in a graded series of acetone , and embedded in Agar 100 resin ( Agar Scientific ) . Seventy nm sections were cut with a diamond knife ( Diatome ) and post-stained with 2% uranyl acetate and Reynolds lead citrate . Sections were examined with the FEI Morgagni as above . All experiments were done at least in duplicate for a total of ≥ 2 biological replicates . Data are displayed as means ± standard deviation ( SD ) . Statistical significance was determined using the unpaired two-tailed Student’s t test . The actual p value and sample size n of each experimental group are provided in the respective figure legends . | Picornaviruses are important human and animal pathogens . Protective vaccines are only available against very few representatives . Furthermore , antiviral drugs have not made it to the market because of serious side effects and viral mutational escape . We here show that pharmacological inhibition of cellular myristoyltransferases severely decreased myristoylation of enteroviral structural proteins as exemplified by coxsackievirus B3 , a prominent pathogen causing virus-induced acute and chronic heart disease . The drug DDD85646 substantially diminished virus yield and almost abolished the infectivity of the residual progeny virus . It is highly effective against several other picornaviruses , except those two included in our study that naturally do not process VP0 . Our work provides new insight into the role of myristoylation in the life cycle of picornaviruses and identifies the responsible cellular enzyme as a promising candidate for antiviral therapy . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"hela",
"cells",
"biological",
"cultures",
"microbiology",
"viral",
"structure",
"viruses",
"rna",
"viruses",
"cell",
"cultures",
"rhinovirus",
"infection",
"microbial",
"genomics",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"viral",
"genomics",
"genomics",
"proteins",
"viral",
"packaging",
"viral",
"replication",
"cell",
"lines",
"virions",
"biochemistry",
"post-translational",
"modification",
"virology",
"myristoylation",
"genetics",
"biology",
"and",
"life",
"sciences",
"cultured",
"tumor",
"cells",
"viral",
"diseases",
"organisms"
]
| 2018 | Cellular N-myristoyltransferases play a crucial picornavirus genus-specific role in viral assembly, virion maturation, and infectivity |
Complex gene expression patterns in animal development are generated by the interplay of transcriptional activators and repressors at cis-regulatory DNA modules ( CRMs ) . How repressors work is not well understood , but often involves interactions with co-repressors . We isolated mutations in the brakeless gene in a screen for maternal factors affecting segmentation of the Drosophila embryo . Brakeless , also known as Scribbler , or Master of thickveins , is a nuclear protein of unknown function . In brakeless embryos , we noted an expanded expression pattern of the Krüppel ( Kr ) and knirps ( kni ) genes . We found that Tailless-mediated repression of kni expression is impaired in brakeless mutants . Tailless and Brakeless bind each other in vitro and interact genetically . Brakeless is recruited to the Kr and kni CRMs , and represses transcription when tethered to DNA . This suggests that Brakeless is a novel co-repressor . Orphan nuclear receptors of the Tailless type also interact with Atrophin co-repressors . We show that both Drosophila and human Brakeless and Atrophin interact in vitro , and propose that they act together as a co-repressor complex in many developmental contexts . We discuss the possibility that human Brakeless homologs may influence the toxicity of polyglutamine-expanded Atrophin-1 , which causes the human neurodegenerative disease dentatorubral-pallidoluysian atrophy ( DRPLA ) .
The generation of complex spatial and temporal gene expression patterns during embryo development is achieved through gene regulatory networks in which broadly distributed transcriptional activators act in combination with repressors with a more restricted distribution ( reviewed in [1] ) . Repressors have an essential role in establishing gene expression boundaries . The mechanisms by which repressors act is not well understood , but may involve competition for DNA binding sites , inhibition of activator function ( quenching ) , and direct repression ( reviewed in [2–5] ) . Many activators and repressors require co-regulators for activity ( reviewed in [6 , 7] ) . One way that co-regulators work is to modulate the chromatin structure in order to facilitate or restrict transcription initiation complex assembly . However , co-regulators may have other functions as well , such as mediating an association between transcription factors and the basal transcription machinery . It is possible that the type of co-regulator that is recruited determines the mechanism of transcriptional control at use . For example , during Drosophila embryo development , repressors acting over a short range recruit the CtBP co-repressor , whereas several long-range repressors interact with the co-repressor Groucho ( reviewed in [8] ) . Yet , the mechanism by which several important transcription factors in the embryo work remains unknown . We therefore set out to isolate novel transcriptional regulators that are required for Drosophila embryo segmentation , and identified the Brakeless protein as a co-repressor that is required for function of the transcription factor Tailless . Segmentation of the Drosophila embryo is achieved through a hierarchy of transcriptional control ( reviewed in [9 , 10] ) . The maternal mRNAs bicoid ( bcd ) and nanos ( nos ) localize to the anterior and posterior poles of the embryo , respectively , from where they give rise to protein gradients in the syncytial embryo . Bcd activates transcription of the hunchback ( hb ) gene and represses translation of maternal caudal ( cad ) mRNA , whereas Nanos represses translation of maternal hb message ( reviewed in [11] ) . The resulting Bcd , Hb , and Cad protein gradients act in combination to turn on expression of the first zygotic patterning genes , the gap genes , in restricted domains in the embryo . The gap gene products in turn are transcriptional repressors that regulate the next level in the hierarchy , pair-rule gene expression . Pair-rule proteins are transcription factors that control the segment-polarity genes , which in turn specify the positions of the 14 segments of the animal . The positioning of gap gene expression domains relies on interpretation of the Bcd , Hb , and Cad activator gradients and on mutual repression by the gap gene products . Positional information originating from activator and gap gene repressor gradients is integrated by cis-regulatory DNA modules ( CRMs or enhancers , reviewed in [9] ) . For example , in the knirps ( kni ) CRM , binding sites for the activators Cad and Bcd are present in separate modules that are distinct from a module binding the repressors Hb , Krüppel ( Kr ) , Giant ( Gt ) , and Tailless ( Tll ) [12] . In a screen for novel maternal factors required for segmentation of the Drosophila embryo [13] , we isolated mutations in the brakeless ( bks ) gene . Bks was previously identified as a nuclear protein with unknown function required for axonal guidance in the Drosophila eye [14 , 15] , where it represses Runt expression in R2 and R5 photoreceptor cells [16] . It is also known as Scribbler , due to its behavioral locomotor phenotype in larvae [17] , and as Master of thickveins ( mtv ) because it is important for expression of the TGF-ß receptor thickveins in wing imaginal disks [18] . Related sequences can be found in deuterostome genomes ( echinoderms and chordates ) , indicating that Bks proteins may play equally important roles in other organisms . We show here that in embryos lacking maternal bks function ( from here on referred to as bks mutant embryos ) , the Kr and kni expression patterns expand despite the presence of the known transcriptional regulators . We find that Bks is recruited to the Kr and kni CRMs , represses transcription when bound to DNA , and functionally interacts with Tll . The Tll protein is a dedicated transcriptional repressor that belongs to the NR2 subfamily of orphan nuclear receptors [19 , 20] . It specifies the terminal embryonic structures by repressing transcription of genes such as Kr and kni [19] . Bks is required for Tll function , and so is Atrophin [21] , the homolog of human Atrophin-1 , which causes the neurodegenerative disease dentatorubral-pallidoluysian atrophy ( DRPLA ) when a polyglutamine stretch in the protein is expanded ( reviewed in [22] ) . We demonstrate a direct interaction between Bks and Atrophin that is conserved between their human homologs , and propose that these proteins work together as a co-repressor complex in many developmental contexts . We discuss the possibility that this interaction may be important for both the normal and pathological function of Atrophin-1 .
From a screen for new maternal genes involved in embryonic pattern formation [13] , we searched for mutant phenotypes that reflect defects in the transcriptional regulation of segmentation . We found a mutant , 2R-14 , that displayed severe segmentation defects in embryonic cuticle preparations ( Figure 1 ) . 2R-14 germline clone larvae show deletions of denticle belts to a variable extent . All of the abdominal segments , as well as terminal structures , can be affected , but there is no effect on dorsal-ventral patterning . Two additional alleles ( 2R-278 and 2R-339 ) were isolated that also show this phenotypic variability , and differ only in the frequency of the phenotypic classes . In contrast to the maternal phenotypes , patterning of 2R-14 zygotic mutant larvae is fully normal , but they die at later stages of development . We mapped the 2R-14 locus to an approximately 600-kilobase ( kb ) interval between 55B and 55E , uncovered by the deficiency Df ( 2R ) PC4 . We performed complementation tests with all available lethal mutants in this interval and found that the 2R-14 , 2R-278 , and 2R-339 alleles fail to complement the brakeless ( bks ) alleles l ( 2 ) 04440 , bks1 and bks2 ( described in [15 , 17] ) . Thus the lethality of the 2R-14 locus maps to the bks gene . We cleaned the 2R-14 chromosome by recombination and did complementation tests with the bks alleles , as well as generated germline clone embryos from the resulting recombinants . The phenotype of these embryos is essentially identical to embryos derived from the original 2R-14 chromosome ( Figure 1C ) . In order to confirm that mutations in the bks gene indeed cause a maternal segmentation defect , we recombined the independently isolated bks1 and bks2 alleles onto FRTG13 ( 42B ) chromosomes . bks1 and bks2 germline clone–derived larvae and early embryos ( Figure 1D and unpublished data ) are phenotypically indistinguishable from 2R-14 mutants . Thus , we named our alleles bks14 , bks278 , and bks339 . The bks locus encodes at least two proteins , Bks-A and Bks-B ( Figure 2A ) . Bks-A is a 929–amino acid ( aa ) -long protein , whereas Bks-B consists of 2 , 302 aa and has the first 929 aa in common with Bks-A [14 , 15 , 17 , 18] . The only sequence similarity to known functional domains is a single C2H2 zinc finger located in the unique region of Bks-B . One additional domain ( D2 ) is highly conserved and present also in sequences from deuterostome species ( Figure S1 ) . In vertebrates , a duplication has resulted in two genes with sequence similarity to Bks , encoding zinc-finger protein 608 ( ZNF608 ) and ZNF609 . In addition , we identified three domains ( D1 , D3 , and D4 ) that are highly conserved in insects and that contain limited similarity to vertebrate sequences ( Figure 2A ) . We sequenced our three bks alleles in order to identify the molecular lesions associated with the mutations ( see Figure 2A ) . For bks14 , we were unable to amplify the first exon using various primer combinations . No mutation was detected in the rest of the gene . Sequencing of bks278 revealed a 345–base pair ( bp ) large deletion after nucleotide 2 , 365 of the Bks-B cDNA . This deletion together with a 8-bp insertion causes a frame shift at aa 741 that results in addition of 79 novel amino acids . The weaker bks339 allele is due to a C to T transition at position 5 , 485 that converts Q1758 into a stop codon . This truncates the protein before the conserved D3 domain and shows that the maternal function of the Bks-B subtype is necessary for embryo development . We examined the bks expression pattern during embryogenesis by whole-mount in situ hybridization with a probe that recognizes both bks-A and bks-B . We found bks mRNA to be expressed in the egg and throughout all stages of embryogenesis . At the blastoderm stage , ubiquitous expression is caused by the maternal contribution of the mRNA ( Figure 2B ) . Following gastrulation , low levels of maternal transcripts remain , and bks is zygotically transcribed in neural cell precursors and the central nervous system ( CNS ) ( see [17] ) . In order to test whether both bks-A and bks-B are present during embryogenesis , we performed reverse transcription PCR ( RT-PCR ) on mRNA from early embryos ( 0–3 h ) . Using primers specific for the A and B isoforms , we detected both transcripts ( Figure 2C ) . An antibody raised against the conserved D2 region of Bks [15] stained all cells in the embryo ( Figure 2D ) , whereas in embryos derived from bks278 or bks14 germline clones , nuclear staining is absent , although we detected cytoplasmic background staining ( unpublished data ) . In summary , we found that Bks is maternally expressed and is present ubiquitously in early embryos . To understand the segmentation phenotypes seen in bks germline clone larvae , we analyzed gene expression patterns of developmental control genes . We tested expression of all gap genes in early blastula stage embryos and found severe expression phenotypes in bks embryos . In wild-type ( wt ) pre-cellular embryos , the gap gene kni is expressed in two domains , one in the anterior-ventral end of the embryo , the other one as a stripe in the posterior half ( Figure 3A ) . In bks mutants , this posterior domain is broadly expanded ( Figure 3B ) . Whereas in wt , the posterior kni domain extends from 27% to 43% egg length ( EL , where 0% is the posterior pole and 100% the anterior pole ) , in bks14 embryos , it extends from 15% to 41% EL . In cellularizing bks embryos , the posterior domain remains expanded , and additional ectopic expression is found in the posterior-ventral end of the embryo ( arrowhead in Figure 3D ) . To determine whether the effect on kni expression is transcriptional or post-transcriptional , we introduced a kni 4 . 4-kb CRM-lacZ transgene [23] into bks germline clone embryos . As shown in Figure 3F , lacZ expression expands towards the posterior as compared to wt embryos ( Figure 3E ) . The kni-lacZ pattern extends from 31% to 43% EL in wt embryos , and expands to 22%–41% EL in bks278 mutant embryos . We conclude that expansion of the kni pattern in bks mutant embryos is due to transcriptional deregulation . Kr is first expressed in a central domain ( CD ) of the embryo ( Figure 3G ) . Later , additional anterior and posterior domains are detectable ( unpublished data ) . In bks embryos , the CD is broadly expanded both in an anterior and a posterior direction ( Figure 3H ) . We measured the CD to 31%–62% EL in bks14 embryos , compared to 40%–57% EL in wt . Intensity of expression also appears enhanced and persists into later stages of embryogenesis . In addition , expression of the anterior domain is enhanced , expanded , and expressed earlier in bks embryos as compared to wt ( arrowhead in Figure 3H ) . Thus , Bks is necessary to restrict Kr expression . The gt expression pattern develops from a broad domain in the anterior half and one narrower domain in the posterior half in the early blastula embryo , to three anterior stripes and a posterior stripe in cellularizing embryos at mid-cycle 14 . In bks embryos , gt expression is variable , but in a majority of embryos , resolution of the anterior domain into stripes is delayed ( compare Figure 3J with 3I ) . The posterior domain is less affected , but in about 25% of the embryos , its expression is reduced ( unpublished data ) . The gap genes are activated by the maternal factors Bcd , Cad , and Hb . The terminal gene products Tll and Huckebein ( Hkb ) act as repressors that restrict gap gene expression together with mutual inhibition by gap gene products . Expression of these upstream regulators is mostly normal in bks mutants ( Figure S2 ) , and cannot be responsible for the gap gene phenotypes observed . In conclusion , three gap genes are de-repressed in bks mutants . Similar to previous findings [16 , 18] , absence of Bks leads to de-repression of transcription , indicating that Bks may normally be involved in transcriptional repression . Gap gene expression boundaries are set by repressor proteins . For example , Kr and Gt restrict each other's expression [24–27] , and Tll represses Kr and kni [23 , 28–30] . We therefore tested whether Bks is a co-repressor required for the activity of regulators of Kr and kni expression . We first investigated whether the activities of Tll and Hb are affected in bks embryos by misexpressing them in wt and bks mutant backgrounds , and compared their ability to repress transcription . We used a snail promoter construct that directs ectopic Tll or Hb expression in the ventral domain of the embryo ( described in Protocol S1 and in [31] ) . When Tll is misexpressed in wt embryos , kni expression becomes repressed in ventral cells ( Figure 4A , arrow ) . By contrast , in a majority of bks mutant embryos , the kni expression pattern is unaffected by misexpressed Tll ( Figure 4B , arrow; and Table S1 ) , suggesting that full Tll activity depends on wt Bks function . On the other hand , misexpression of Hb from the snail promoter causes repression of kni in the ventral half of the embryo in both wt ( Figure 4C ) and in a bks mutant background ( Figure 4D ) . Despite the enhanced and expanded levels of kni expression in bks embryos , ectopic Hb is sufficient to repress kni ventrally . The ectopic patch of kni expression in the posterior-ventral part of the embryo remains unaffected ( star in Figure 4D ) , presumably because snail expression does not extend to the very posterior of the embryo [32] . To examine the repressor activities of Kni , Kr , and Gt proteins , we introduced lacZ reporter gene constructs into bks mutant embryos . LacZ expression is driven by a modified rhomboid neuroectoderm enhancer ( NEE ) that is activated on the ventral side of the embryo by the Dorsal and Twist proteins . In addition , the enhancer constructs contain either Kni , Kr , or Gt binding sites ( described in [33–35] ) . In wt embryos , binding of the corresponding gap protein leads to repression of the reporter gene in the domain of gap gene expression ( Figure 4E , 4G , and 4I ) . Similarly , lacZ expression is repressed in the gap gene expression domains in a bks background ( Figure 4F , 4H , and 4J ) . Thus , in bks mutants , the three gap proteins Kni , Kr , and Gt are able to perform repression at least on the artificial enhancer constructs used , indicating that Bks is not required for the repressor activities of these proteins . We conclude that Tll-mediated repression is impaired in a bks mutant background , whereas the Hb , Kni , Kr , and Gt repressors are not affected under these conditions . We tested whether the dependence of Tll repressor function on Bks might be due to a molecular interaction between these proteins . Tll and Bks-A were expressed as GST-fusion proteins in bacteria , and mixed with radiolabeled in vitro–translated proteins . As shown in Figure 5A , in vitro–translated Tll interacts with GST-BksA , and in vitro–translated Bks-B interacts more strongly with full-length Tll than with a GST fusion lacking the DNA binding domain . This shows that Bks and Tll interact in vitro , and that the Tll DNA binding domain is important for the interaction . A functional interaction between Tll and Bks was demonstrated in vivo by genetic means . We found that lowering the dose of bks in a tll mutant background causes enhanced de-repression of kni expression . In embryos derived from a tll hypomorph , kni expression expands towards the posterior ( compare Figure 5D with Figure 5B ) . By contrast , embryos receiving half the dose of maternal bks have an essentially wt kni expression pattern ( Figure 5C ) . However , in tll mutant embryos with reduced amounts of maternal bks product , the kni expression pattern expands even further to the posterior ( Figure 5E ) . These results suggest that Bks and Tll cooperate to set the normal posterior boundary of kni expression . It was recently demonstrated that Tll also interacts with the Atrophin protein , and that Atrophin and Bks genetically interact in adult flies [21 , 36] . We therefore performed a GST pulldown assay to investigate whether Bks and Atrophin can interact in vitro . As previously published [21] , the C-terminus of Atrophin interacts with the ligand binding domain of Tll ( Figure 5F ) . We found that the Atrophin C-terminus interacts with GST-BksA as well . A truncated Bks protein ( Bks 1–780 ) lacking the evolutionarily conserved D2 region still binds to Atrophin , but a weaker , independent interaction was also found with a Bks portion consisting of the conserved D2 region and the zinc finger ( Bks 834–1 , 151 , Figure 5F ) . Thus , Atrophin can bind to at least two separate parts of the Bks protein . These results show that Tll can interact with both Bks and Atrophin , and that Bks and Atrophin can bind to one another as well . This suggests that a tripartite complex consisting of Tll , Bks , and Atrophin might form . We confirmed the interactions among Bks , Atrophin , and Tll in S2 cells expressing V5-tagged Bks-B and FLAG-tagged Tll proteins . Using an Atrophin antibody , we could co-immunoprecipitate V5-tagged Bks and FLAG-tagged Tll with endogenous Atrophin ( Figure 5G ) . We then tested whether this interaction is evolutionarily conserved . We made a GST-fusion protein consisting of the first 600 aa of the human Bks homolog ZNF608 ( including the conserved D2 domain and the zinc finger ) , and mixed it with radiolabeled C-terminus of human Atrophin-1 . A strong interaction between these proteins was observed ( Figure 5H ) . We conclude that the interaction between Bks and Atrophin has been conserved during evolution . An interaction with Tll is expected to bring Bks to the kni and Kr CRMs to directly regulate their expression . To determine if Bks is associated with the Kr and kni CRMs , we performed chromatin immunoprecipitations ( ChIP ) from S2 cells expressing V5-tagged Bks-B protein . We found a 23-fold and 4 . 7-fold enrichment at the kni and Kr CRMs , respectively , with the V5 antibody compared to a control green fluorescent protein ( GFP ) antibody ( Figure 6A and 6B ) . As a control , we performed ChIP from normal S2 cells lacking the tagged Bks protein . From these cells , the V5 antibody precipitated less kni and Kr CRM DNA than the control GFP antibody ( Figure 6A and 6B ) . A comparable amount of kni 5′ UTR DNA was precipitated with the V5 antibody from V5-tagged Bks-B–expressing cells as from normal S2 cells ( 2 . 8-fold and 2 . 7-fold compared to GFP antibody; Figure 6C ) . A locus on Chromosome 4 was precipitated at a similar efficiency with V5 and GFP antibodies ( 1 . 7-fold enrichment; Figure 6D ) . From these results , we conclude that Bks specifically associates with kni and Kr CRM sequences when expressed in S2 cells . We extended these results to Drosophila embryos using an affinity-purified antibody raised against Bks amino acids 450–620 . We found an enrichment of kni CRM sequences with Bks , Atrophin , and Tll antibodies compared to the control V5 antibody with chromatin prepared from wt embryos ( Figure 6E ) . No enrichment at the Chromosome 4 locus was observed ( Figure 6E ) . In summary , both Bks and Atrophin are recruited to a Tll-regulated target gene in vivo . Since bks genetically behaves as a repressor , we tested whether Bks proteins are capable of repressing transcription when tethered to a promoter . We fused bks coding regions to the DNA binding domain of the tetracycline repressor ( TetR-DBD ) and expressed the fusion constructs in Drosophila tissue-culture cells . We co-transfected a luciferase reporter construct driven by the actin5C enhancer that also contains tet operators , binding sites for the TetR-DBD ( described in [37] ) . We compared luciferase activity of cells that expressed TetR-Bks fusion proteins with those that expressed the TetR-DBD protein alone . We found that both Bks-A and Bks-B are able to repress transcription when tethered to DNA in mbn-2 as well as in S2 cells ( Figure 7A and unpublished data ) . We also investigated Bks repressor activity in a transgenic embryo assay . Bks-A coding sequence was fused to the Gal4 DNA binding domain and placed under control of the Kr CD enhancer , which directs expression in the CD of the early embryo . A lacZ reporter gene containing a modified rhomboid NEE lacking Snail repressor sites and containing three upstream activation sequence ( UAS ) sites was used to monitor Gal4-Bks repressor activity ( described in [34] ) . In a wt background , the reporter gene is expressed in ventral regions of the embryo ( Figure 7B ) . However , when crossed into transgenic embryos expressing the Gal4-BksA fusion protein , the reporter is repressed in central regions ( Figure 7C ) . In conclusion , our data show that Bks proteins are capable of repressing transcription when bound to a promoter . Taken together with our other results , we conclude that Bks acts as a transcriptional co-repressor .
Repression plays a pivotal role in establishing correct gene expression patterns that is necessary for cell fate specification during embryo development . For example , in the early Drosophila embryo , repression by gap and pair-rule proteins is essential for specifying the positions of the 14 segments of the animal . The mechanisms by which transcriptional repressors delimit gene expression borders are not well understood . However , many repressors require co-repressors for function . In the Drosophila embryo , the CtBP and Groucho co-repressors are required for activity of many repressors ( reviewed in [8 , 38] ) . More recently , Atrophin has been identified as a co-repressor for Even-skipped and Tll [21 , 39] . Still , co-regulators for several important transcription factors in the early embryo have not yet been identified . We therefore performed a screen for novel maternal factors that are required for establishing correct gene expression patterns in the early embryo . From this screen , we identified mutations in the bks gene that cause severe phenotypes on gap gene expression and embryo segmentation . The Bks protein is evolutionarily conserved between insects and deuterostomes , but has not been characterized in any species except Drosophila , in which it has been shown to repress runt expression in photoreceptor cells and thickveins expression in wing imaginal disks [16 , 18] . However , the molecular function of Bks was unknown . We show here that Bks interacts with the transcriptional repressor Tll , is recruited to target gene CRMs , and will repress transcription when targeted to DNA . Tll was recently shown to utilize Atrophin as a co-repressor [21] . Atrophin genetically interacts with Tll and physically interacts with its ligand binding domain . Atrophin binding is conserved in nuclear receptors within the same subfamily , such as Seven-Up in Drosophila as well as Tlx and COUP-TF in mammals [21 , 40] . When expressed in mammalian cells , Drosophila Atrophin and mouse Atrophin-2 interact with the histone deacetylases HDAC1 and HDAC2 [21 , 41] . Histone deacetylation may therefore be part of the mechanism by which Atrophin functions as a co-repressor . Another recent report described genetic interactions among bks and atrophin mutants in the formation of interocellar bristles in adult flies [36] . Furthermore , it was shown that atrophin mutants have virtually identical phenotypes as bks mutants , including de-repression of runt expression in the eye , thickveins expression in the wing , and Kr and kni expression in the embryo [21 , 36 , 42] . We now show that both proteins are recruited to the kni CRM , a Tll-regulated target gene , in the embryo . Importantly , we further demonstrate that Atrophin and Bks interact in vitro and that they can be co-immunoprecipitated from S2 cells . We propose that Bks and Atrophin function together as a co-repressor complex , and based on the similar bks and atrophin mutant phenotypes at several developmental stages , the complex may function throughout development . Our results are compatible with the existence of a tripartite complex consisting of Tll , Bks , and Atrophin . Bks binding to Tll is enhanced by the Tll DNA binding domain , whereas the interaction of Tll with Atrophin is mediated through the C-terminal ligand binding domain . Tll may therefore simultaneously interact with Bks and Atrophin . Alternatively , Tll interacts separately with Bks and Atrophin on the kni CRM . In either case , both Bks and Atrophin are required for full Tll activity . However , at high enough Tll concentration , Bks activity is dispensable . Some bks embryos misexpressing Tll still repress kni expression ( Table S1 ) , and overexpressing Tll from a heat-shock promoter can repress the posterior kni stripe in both wt and bks mutant embryos ( unpublished data ) . For this reason , we believe that Bks and Atrophin are cooperating as Tll co-repressors , so that Tll function is only partially impaired by the absence of either one . We found that Tet-Bks–mediated repression in cells is insensitive to the deacetylase inhibitor trichostatin A ( TSA; unpublished data ) . It is possible , therefore , that whereas Atrophin-mediated repression may involve histone deacetylation , Bks could repress transcription through a separate mechanism . Our results have not revealed any differences between the molecular functions of the two Bks isoforms . Both Bks-A and Bks-B repress transcription when tethered to DNA , and the sequences that mediated binding to Tll and Atrophin are shared between the two isoforms . However , the bks339 allele that selectively affects the Bks-B isoform causes a weaker , but comparable phenotype to the stronger bks alleles that disrupt both isoforms . Therefore , the C-terminus of Bks-B provides a function that is indispensable for embryo development and regulation of kni expression . This part of Bks-B contains two regions ( D3 and D4 ) that are highly conserved in insects and loosely conserved in deuterostome Bks sequences , but does not resemble any sequence with known function . The only sequence similarity to domains found in other proteins is a single zinc-finger motif in Bks-B . Preliminary results indicate that the zinc finger in isolation or together with the conserved D2 domain does not exhibit sequence-specific DNA binding activity ( unpublished data ) . Indeed , multiple zinc fingers are generally required to achieve DNA binding specificity ( reviewed in [43] ) . Instead , Bks is likely brought to DNA through interactions with Tll and other transcription factors . Atrophins are required for embryo development in Caenorhabditis elegans , Drosophila , zebrafish , and mice [39 , 41 , 42 , 44–47] . In vertebrates , two atrophin genes are present . Atrophin-1 is dispensable for embryonic development in mice , and lacks the N-terminal MTA-2 homologous domain that interacts with histone deacetylases [48] . However , the homologous C-termini of Atrophin-1 and Atrophin-2 can interact , and we found that this domain can also bind to the human Bks homolog ZNF608 ( Figure 5H ) . Atrophin-1 interacts with another co-repressor–associated protein as well , ETO/MTG8 , and can repress transcription when tethered to DNA [49] . These data are consistent with the emerging view that deregulated transcription may be an important mechanism for the pathogenesis of polyglutamine diseases ( reviewed in [50 , 51] ) . Recent evidence indicates that interactions with the normal binding partners may cause toxicity of polyglutamine-expanded proteins such as Ataxin-1 [52] . It will be interesting to investigate whether the interaction between human Bks homologs and Atrophin-1 is important for the neuronal toxicity of polyglutamine-expanded Atrophin-1 .
The bks alleles bks14 , bks278 , and bks339 were generated on an FRT2R-G13–containing chromosome in germline clone ethylmethane sulfonate ( EMS ) screens performed in Tübingen ( [13] and N . Vogt , unpublished data ) . Recombination mapping placed the 2R-14 locus on chromosome arm 2R between the markers curved ( 52D ) and plexus ( 58E ) . Complementation tests with deficiencies covering this area narrowed the 2R-14 locus down to approximately 600 kb between 55B and 55E , uncovered by the deficiency Df ( 2R ) PC4 . We performed complementation tests with all available lethal mutants in this interval and found that the 2R-14 , 2R-278 , and 2R-339 alleles fail to complement the bks alleles l ( 2 ) 04440 , bks1 , and bks2 . l ( 2 ) 04440 is a P element insertion described in [17] . The bks1 and bks2 EMS-induced alleles , kindly provided by Barry Dickson ( described in [15] ) , were recombined to an FRT2R-G13–containing chromosome ( using stock #1958 in [53] ) . The bks14 and bks278 alleles were outcrossed against an FRT2R-G13 c px sp/CyO hs-hid chromosome to clean the stock from additional mutations . Four different recombinants ( two from both sides of the bks locus ) were tested and showed no significant phenotypic differences from the parental chromosomes . The bks14 , bks278 , and bks339 alleles were balanced over CyO tubulin-GFP to enable isolation of homozygous mutant larvae . Genomic DNA was prepared and bks exonic sequences amplified by PCR , sequenced , and compared to an FRT2R-G13 chromosome derived from another mutant from the screen , 2R-91 . Females harboring bks germline clones were crossed with transgenic males to introduce various transgenes into bks mutant embryos . To determine if kni expansion in bks germline clone embryos is due to transcriptional control , we crossed males containing a lacZ reporter regulated by a 4 . 4-kb kni enhancer ( GO125 kni4 . 4lacZ , [23] ) to bks278 germline clone females and analyzed the resulting embryos for lacZ expression using in situ hybridization . To analyze the activity of Tll and Hb in a bks mutant background , males containing transgenes misexpressing tll or hb were crossed to bks278 or bks14 germline clone females or to wt females . Expression in a ventral domain of the embryo was achieved by use of the snail promoter ( sna:tll stocks x196 and x197 , described in Protocol S1 , and sna:hb stock x227 , described in [31] ) . The constructs contain transcriptional stop signals flanked by FRT sites downstream of the sna promoter to allow maintenance of transgenic lines . Ventral expression was activated by crossing in a ß2-tubulin-FLP transgene [54] . Male progeny containing both FLP and sna promoter transgenes ( in whose spermatocytes recombination occurred ) were crossed to virgins with bks germline clones or to wt virgins; embryos were then collected and processed for in situ hybridization with a kni probe . To test the repressor activities of Kni , Kr , and Gt proteins in a bks germline clone background , we crossed wt females or females with bks14 germline clones with males containing modified rhomboid NEE enhancers . The NEE-kni-lacZ transgene ( lab stock A45 ) is described in [33] and contains synthetic Kni binding sites , but lacks Snail sites . NEE-Kr-lacZ ( lab stock G5 . 5 ) contains synthetic Kr sites , but lacks Snail sites , and is described in [34] . The 2xgt-55 lacZ reporter gene is described in [35] . It is activated by the rhomboid NEE as well as the 2xPE twist enhancer and contains two Gt sites situated 55 bp upstream of the transcription start site . Embryos were collected and fixed 2–4 h after egg laying , and lacZ expression patterns were analyzed by in situ hybridization . Flies containing a modified rhomboid NEE-lacZ reporter gene with three UAS sites ( described in [34] ) were crossed to wt or Kreggy-BksA transgenic flies ( see Protocol S1 ) , embryos collected , and lacZ reporter gene expression analyzed by in situ hybridization . Genetic interactions between bks278 and tll1 were tested by crossing bks278/+; tll1/+ females with tll1/TM3 Sb males . Embryos from this cross were compared to embryos derived from bks278/+ females crossed to wt males , and with embryos derived from the tll1 stock . A detailed description of this procedure can be found in Protocol S1 . In brief , we established a stable S2 cell line expressing V5-tagged Bks-B , prepared sheared chromatin from this and a control S2 cell line , as well as from 2–4-h wt embryos , and performed ChIP essentially according to the Upstate ChIP assay kit protocol ( Upstate Biotechnology , http://www . upstate . com ) . Real-time PCR was performed on an ABI prism 7000 machine using Power SYBR Green reagent ( Applied Biosystems , http://www . appliedbiosystems . com ) . PCR was performed on 1 μl ( cells ) or 3 μl ( embryos ) template DNA in triplicate samples , and immunoprecipitated DNA was compared against standard curves from serial dilutions of input DNA . The values are plotted as percent input DNA from the corresponding extract , and the standard deviation within the triplicate samples indicated . Similar results were obtained in independent ChIP experiments .
The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank ) accession number for the Drosophila melanogaster Bks-B cDNA is AF242194 . | Nuclear receptors play important roles in embryonic development and cellular differentiation by regulating gene expression at the level of transcription . The functions of transcriptional repressors , including nuclear receptors , are often mediated by other proteins , so-called co-repressors . We performed a genetic screen in the fruit fly Drosophila melanogaster to search for novel co-repressor proteins . We isolated mutations in the brakeless gene that alter normal transcriptional repression in early fly embryos . Brakeless was already known to regulate axon guidance in the eye , larval behavior , and gene expression in wing imaginal discs . However , the molecular function of this protein was unknown . Here we show that Brakeless is a co-repressor required for function of the Tailless nuclear receptor . Tailless was previously shown to interact with another co-repressor , Atrophin . Here , we demonstrate that Brakeless and Atrophin can bind to one another and that this interaction is conserved between a human Brakeless homolog , ZNF608 , and human Atrophin-1 . A polyglutamine expansion in Atrophin-1 is the cause of the neurodegenerative disease dentatorubral-pallidoluysian atrophy ( DRPLA ) . It is possible that the interaction with ZNF608 could contribute to the pathogenesis of polyglutamine-expanded Atrophin-1 . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
]
| [
"developmental",
"biology",
"homo",
"(human)",
"drosophila",
"molecular",
"biology",
"genetics",
"and",
"genomics"
]
| 2007 | Drosophila Brakeless Interacts with Atrophin and Is Required for Tailless-Mediated Transcriptional Repression in Early Embryos |
Distal arthrogryposis type 2B ( DA2B ) is an important genetic disorder in humans . However , the mechanisms governing this disease are not clearly understood . In this study , we generated knock-in mice carrying a DA2B mutation ( K175del ) in troponin I type 2 ( skeletal , fast ) ( TNNI2 ) , which encodes a fast-twitch skeletal muscle protein . Tnni2K175del mice ( referred to as DA2B mice ) showed typical DA2B phenotypes , including limb abnormality and small body size . However , the current knowledge concerning TNNI2 could not explain the small body phenotype of DA2B mice . We found that Tnni2 was expressed in the osteoblasts and chondrocytes of long bone growth plates . Expression profile analysis using radii and ulnae demonstrated that Hif3a expression was significantly increased in the Tnni2K175del mice . Chromatin immunoprecipitation assays indicated that both wild-type and mutant tnni2 protein can bind to the Hif3a promoter using mouse primary osteoblasts . Moreover , we showed that the mutant tnni2 protein had a higher capacity to transactivate Hif3a than the wild-type protein . The increased amount of hif3a resulted in impairment of angiogenesis , delay in endochondral ossification , and decrease in chondrocyte differentiation and osteoblast proliferation , suggesting that hif3a counteracted hif1a-induced Vegf expression in DA2B mice . Together , our data indicated that Tnni2K175del mutation led to abnormally increased hif3a and decreased vegf in bone , which explain , at least in part , the small body size of Tnni2K175del mice . Furthermore , our findings revealed a new function of tnni2 in the regulation of bone development , and the study of gain-of-function mutation in Tnni2 in transgenic mice opens a new avenue to understand the pathological mechanism of human DA2B disorder .
Distal arthrogryposis type 2B ( DA2B ) ( also called Sheldon-Hall syndrome , or SHS ) is an autosomal dominant genetic disorder with marked genetic heterogeneity . The typical clinical phenotypes of DA2B patients include hand and/or foot contracture malformation and shortness of stature [1]–[3] . Approximately 20% of familial incidences of DA2B can be explained by mutations in TNNI2 that encodes a subunit of the troponin complex consisting of troponin C , troponin I and troponin T . Troponin complex is set of muscle proteins that are part of the contractile apparatuses for contraction of fast skeletal muscle . The previous knowledge shows that the two functional domains of TNNI2 are located at the N terminus ( residues 96 to 115 and residues 140 to 148 ) . In absence of Ca2+ , TNNI2 prevents muscle contraction through binding to actin and tropomyosin using the two domains [4] , [5] . However , to date , all of the etiological mutations reported in DA2B families [6]–[10] are mapped at the TNNI2 C terminus that appears to regulate , in part , its sensitivity to Ca2+ ions [11] , [12] . These TNNI2 mutations include p . R156X , p . E167del , p . R174Q , p . K175del and p . K176del [6]–[10] . They are highly conserved in eukaryotes from zebrafish to humans . Evidence from in vitro cell assays has suggested that the p . R156X and p . R174Q mutations facilitate muscle contractility through increasing the sensitivity of TNNI2 to Ca2+ [13] . To investigate pathogenesis of DA2B , we generated genetically modified mice with an endogenous Tnni2K175del mutation , which had been mapped by our previous study , using targeted gene editing [10] . Besides limb deformity , we observed that both the homozygous and heterozygous mutant mice consistently showed a small body size that was similar to the reduced size phenotype in human DA2B . However , this phenotype did not seem to be convincingly explained using the present knowledge of TNNI2 associated skeletal muscle contraction . Thus , we hypothesized that an additional function of TNNI2 might be responsible for this abnormal body development . To test this possibility , we focused on investigating the potential mechanism that leads to the small body size phenotype in DA2B mice .
Using site-directed mutagenesis and targeted gene editing , we generated genetically modified mice with an endogenous Tnni2K175del mutation ( DA2B mice ) ( Figure S1 ) , which has been mapped to a large Chinese DA2B family in our previous study [10] . The 175K amino acid is highly conserved from zebrafish to humans ( Figure 1A ) . Like DA2B patients , homozygous and heterozygous mutant mice exhibited an obviously smaller body size than their wild-type counterparts ( Figure 1B and Figure S2B ) , as well as forelimb contracture ( Figure S2A and C ) , supporting the idea that the mutation has a dominant effect on body development and limb formation . Newborn heterozygous ( n = 41 ) and homozygous mice ( n = 25 ) had lower birth weights ( by approximately 10% and 15% , respectively ) than wild-type littermates ( n = 122 ) ( P<0 . 001 ) ( Figure 1C ) . The growth rates of heterozygous ( Tnni2+/K175del ) mice from post-neonatal day ( P ) 5 to P20 were significantly slower than wild-type controls ( P<0 . 001 , n = 7 ) ( Figure 1D ) , and further analysis indicated that P12 Tnni2+/K175del mice ( n = 125 ) were smaller and lighter than wild-type controls ( n = 111 ) ( P<0 . 001 ) ( Figure 1E ) . The lengths of the spines , forelimbs and hindlimbs of P12 heterozygous mice were consistently shorter than those of their wild-type littermates ( by 13–28% in the spines; by 13–24% in the forelimbs; and by 12–24% in the hindlimbs ) ( n = 5 , Figure 1F ) . The length of the spine , fore limbs and hind limbs ( Figure 1F ) seemed to vary between the different pairs , and we considered that the difference between the different pairs of mutant mice could be due to individual variability in breastfeeding . However , the differences were not significant ( for spines P = 0 . 21; for forelimbs P = 0 . 34; for hindlimbs P = 0 . 53 ) . In addition , the respective lengths of the spines and forelimbs of Tnni2+/K175del mice were 9–14% and 7–14% shorter than wild-type littermates at 2 months ( n = 5 , Figure 1G ) . Except for the small body size and contracture in the forelimbs , we did not observe any limitations in movement of heterozygous or homozygous DA2B mice . Neither heterozygous nor homozygous embryos displayed differences in body size and weight relative to their wild-type littermates before embryonic day ( E ) 18 . 5 ( Figure S3 ) , indicating that mutant mice apparently experienced growth retardation during late embryonic development . Because the majority of homozygous Tnni2K175del/K175del mice always died within a few hours after birth and less than 2% of Tnni2K175del/K175del mice survived until weaning ( Figure S4 ) , subsequent studies were performed mainly using heterozygous mutants . Together , our data indicated that the phenotypes of Tnni2K175del mutant mice were similar to those in human . Given that the known function of TNNI2 in regulating muscle contraction does not seem to reasonably explain the characteristic small body size that was observed in the DA2B mice and DA2B patients [14]–[15] , we speculated that TNNI2 might have unknown roles in body development . Therefore , we initially performed in situ hybridization analyses to characterize the Tnni2 expression in different organs from E15 . 5 wild-type embryos ( Figure 2A , Figure S5 ) . Tnni2 mRNA was detected in the growth plates of long bones of radii and ulnae ( Figure 2A ) , oesophagus ( Figure S5B ) and skeletal muscle as well ( Figure S5G ) . Subsequent immunostaining analyses showed that tnni2 was present in the chondrocytes ( Figure 2B ) of growth plates and in the osteoblasts ( Figure 2C ) surrounding the border of the growth plates . Moreover , the expression of Tnni2 in the growth plates exhibited a particular spatial and temporal change in wild-type mice ( Figure 2D–F ) . In E15 . 5 wild-type embryos , tnni2 protein was observed in pre-hypertrophic and hypertrophic chondrocytes ( Figure 2D ) , whereas at E17 . 5 , tnni2 was mainly observed in proliferating cells zones ( Figure 2E ) . After birth , tnni2 was expressed by resting , immature-proliferating , pre-hypertrophic and hypertrophic chondrocytes , but no discernible tnni2 was observed in mature proliferative chondrocytes ( Figure 2F ) . In addition , we observed the expression of Tnni2 in articular chondrocytes ( Figure 2G ) and calvarial osteoblasts ( Figure 2H ) of wild-type mice . The similar expression pattern was also observed in Tnni2K175del mice ( Figure S6 ) . Thus , the expression in the mouse growth plate and cranium suggested that the tnni2 protein might be involved in bone development . We then investigated the bone morphology in DA2B mice by performing skeletal analyses . All skeletal elements of DA2B mice had a markedly smaller size and lower mineral content than those observed in wild-type mice . The degree of mineralization in the cranium was significantly reduced at E16 . 5 homozygous mutant embryos , P2 and P5 heterozygous mutant mice ( Figure 3A and Figure S7 ) , indicating that intramembranous ossification was delayed in Tnni2K175del mutants . At E14 . 75 , the long bones of the forelimbs of Tnni2K175del/K175del embryos were smaller , and the primary centers of ossification in radii , ulnae , humeri and scapulae were evidently shorter than those in their wild-type littermates ( Figure 3B and Figure S8 ) . The heterozygous mutants had delayed cartilage ossifications in the ribs ( Figure 3C ) , carpus and metacarpus ( Figure 3D ) , phalanx ( Figure 3E ) , growth plates of radii and ulnae ( Figure 3F ) and tail cartilage ( Figure 3G ) at P2 , P5 , P6 , and P10 . The primary and secondary ossification centers in the long bones of DA2B mice were impeded ( Figure 3B–G ) , and their skeletal structures were smaller ( Figure 3H ) , indicating that endochondral ossification was also impaired in Tnni2K175del mutants . Taken together , our data indicated that the Tnni2K175del mutation impeded the intramembranous ( Figure 3A and Figure S7 ) and endochondral ( Figure 3B–G ) ossification of DA2B mice . To characterize the molecules involved in the impeded bone development in Tnni2K175del mutant mice , we performed a microarray analysis using three pairs of intact radii and ulnas from 1-day-old Tnni2K175del/K175del and wild-type littermate mice . We tested differential gene expression using the SAM package . The Tnni2K175del mutation was associated with upregulation of 58 and downregulation of 6 probe sets ( Score ≥3 and fold ≥2 ) ( Figure 4A and Table S1 ) . Furthermore , gene set enrichment analysis ( GESA ) of these mutant and wild-type samples showed significant enrichment of a set of hypoxia induced genes [16] ( Normalized Enrichment Score = 2 . 30 , P = 0 . 0 , FDR q-value = 0 . 0 ) ( Figure 4B ) . Hypoxia signaling , which is mainly mediated by hypoxia-inducible factor 1α ( HIF1A ) and its downstream target genes , including VEGF and its receptor FLT1 , is important for bone development [17]–[20] . Notably , in our microarray dataset ( SAM analysis , Table S1 ) , Hif3a ranked in the top fifth of the significant differential expression genes . Of the genes that were observed to be dramatically upregulated in our microarray dataset , Hif3a had higher expression levels ( more than 3-fold ) in homozygous mutants than in the wild-type littermates ( Figure 4C ) . HIF3A negatively regulates HIF1A signaling by binding to HIF1A or HIF1B and antagonizing HIF1-induced gene expression [21]–[22] . Therefore , we further examined the expression patterns of Hif3a and Hif1a in mRNA levels by qPCR in 1-day-old wild-type and mutant radii and ulnas ( Figure 4D ) . Consistently , Hif3a and a main splicing variant of the Hif3a ( Ipas ) [23] were significantly increased by approximately 3 . 6-fold on average in homozygous mutant bones ( Figure 4D ) . Hif1a expression was no significant difference between mutant bones and wild-type controls ( Figure 4D ) . Furthermore , immunostaining showed that hif3a protein was significantly increased in the hypertrophic cell zones of mutant radii and ulnae compared to wild-type controls ( Figure 5A ) . However , hif1a did not seem to differ in the growth plate of radii at P12 mutant and wild-type controls ( Figure 5B ) . Thus , the marked increase in Hif3a expression in Tnni2K175del mutant bones suggested that Hif-signaling could be compromised in the mutant bones of DA2B mice . We next tested the expression of molecules involved in Hif-signaling using qPCR analyses in 1-day-old homozygous mutant radii and ulnae and wild-type controls . The results showed that many hypoxia-associated genes , including Hif1b , Vegf164/188 , and Flt1 , were downregulated ( approximately 1 . 85–2 . 2-fold ) in homozygous mice compared to the wild-type controls ( Figure 4D ) . Similarly , Aldoa and Eno1 , two Hif1a target genes , were slightly decreased in Tnni2K175del mutants compared to controls ( Figure 4D ) . In addition , the expression patterns of Hif3a/Hif1a/Vegf/Flt1 in calvaria were consistent with those of the radii and ulnae ( Figure S9 ) . Taken together , our data suggested that the Tnni2K175del mutation disturbed the expression of molecules involved in Hif-signaling in the mutant bones . Tnni2K175del mutation impeded the formation of primary and secondary ossification , and impaired chondrocyte differentiation and osteoblast proliferation in bones . Vegf is important for bone development [18]–[20] and was significantly decreased in the bones of Tnni2K175del mice ( Figure 4D and Figure S9 ) . In accordance with diminished transcription , the vegf protein level was dramatically lower ( 96% , P<0 . 001 , n = 4 ) at the hypertrophic cell zones of growth plates of radii and ulnae from heterozygous mutant mice than at those regions of wild-type mice ( Figure 5C ) . As a consequence , fewer blood vessels were observed in the trabecular bones of heterozygous mutant mice at P12 than wild-type littermates ( Figure 5D ) . Vegf plays important roles in the formation of ossification centers [18] , [24] . It is produced by hypertrophic chondrocytes , and then induces blood vessel invasion into the matrix of hypertrophic zones . In this study , we observed a delay in the formation of primary and secondary ossification centers in mutant mice ( Figure 5E–J ) . At E15 . 5 , apoptotic chondrocytes were replaced by osteoblasts in the cartilage of wild-type mice , but not in their Tnni2K175del/175K counterparts ( Figure 5E ) . The vegf expression was reduced in growth plates of radii and ulnae from E15 . 5 homozygous mutant mice compared to that of wild-type littermates ( Figure 5F and G ) . As a consequence , the recruitment of blood vessels into the cartilage plates was evidently fewer in E15 . 5 homozygous embryos than wild-type embryos ( Figure 5H and I ) . At P6 , early stage of the secondary ossification , the hypertrophy of growth plate chondrocytes and the invasion of vascular canals were normal in wild-type mice . However , these characteristics were not observed in heterozygous mutant mice ( Figure 5J ) . Vegf also promotes chondrocyte differentiation and stimulates osteoblast proliferation [24]–[26] . We next tested the effect of the decreased vegf on chondrocyte and osteoblast . We observed that the mineralization of the matrix and hypertrophic chondrocytes was absent at P6 nasal cartilage from Tnni2+/K175del mice compared to those of their wild-type littermates ( Figure 6A ) . At 1-day-old homozygous mutant radii , the mineralization of the matrix at hypertrophic cells was fewer than that of wild-type littermate bones ( Figure S10 ) . Notably , the lengths of the hypertrophic cartilage zones in homozygous mice were significantly shorter than those of wild-type controls ( Figure 6C and D ) . Furthermore , we tested the expression of differential chondrocyte markers using qPCR . The results showed that Ihh , Pthrp , Col10a1 and Comp were decreased in radii and ulnae from 1-day-old homozygous mutant mice ( Figure 6B ) . Together , these observations indicated a delay in chondrocyte differentiation in DA2B mice . Osteoblasts produce osteoid , which is mainly composed of type I collagen and is mineralized to form trabecular bone and compact bone [27]–[28] . We found that the type I collagen , trabecular bones and the number of osteoblasts surrounding the trabecular bones were significantly decreased in Tnni2+/K175del mutant radii and ulnae compared to those of wild-type littermates ( Figure 6E–H ) . The in vitro cultures of primary osteoblasts showed that the differentiation of osteoblasts of the mutants had no appreciable effect on alkaline phosphatase expression compared to that of control cells ( Figure S11 ) . However , BrdU incorporation experiments showed that the proliferation rate of the osteoblasts in heterozygous mutant mice was markedly lower ( by 2 . 4-fold ) than that of wild-type controls ( Figure 6I and J ) . These results suggested that a deficiency in proliferative osteoblasts in DA2B mice might , at least in part , contribute to a reduction in trabecular bone . Taken together , our data showed that Tnni2K175del mutant mice exhibited an impairment of angiogenesis , delay in endochondral ossification , and decrease in chondrocyte differentiation and osteoblast proliferation , which were similar to those bone defects observed in Vegf-deficient mice [18] , [29] . Subsequently , we administered recombinant vegf to heterozygous mutants from P4 to P12 . Notably , after supplementation with recombinant mouse vegf , the number of osteoblasts surrounding the trabecular bone was increased by approximately 30% on average ( Figure 6K and L ) , and the number and density of trabecular bones in the treated mice were increased by approximately 36% ( Figure 6K and M ) . These observations further suggested that the reduced vegf could contribute to defective bone development of the Tnni2K175del mice . We next determined how the Tnni2K175del mutation resulted in an increase in Hif3a expression . A proteomic analysis was employed to identify novel TNNI2 interaction partners ( Table S2 ) . Immunoprecipitation followed by mass spectrometry analysis was performed using the nuclear extract from 293T cells ectopically expressing tnni2-GFP fusion protein . Interestingly , a group set of nucleocytoplasmic transporters including KPNB1 , IPO5 , IPO8 , IPO9 , TNPO1 and TNPO2 , was identified as potential TNNI2 binding partner ( Table S2 ) , which suggested that tnni2 might be a nucleocytosolic shuttling protein and that its nuclear localization might be implicated in the regulation of Hif3a expression . To test this possibility , we examined the subcellular distribution of tnni2 in primary mouse osteoblasts . As shown in Figure 7A and Figure S12 , tnni2 was observed in the cytoplasm and the nucleus of osteoblasts , and tnni2 nuclear distribution was independent of environmental oxygen tension . Similar results were also observed in the primary Tnni2K175del mutant osteoblasts ( Figure S13 ) , and other cell types , such as 293T and HepG2 cells ( Figure S14 ) . To determine whether nuclear-localized tnni2 was involved in the transcriptional regulation of Hif3a , we performed chromatin immunoprecipitation ( ChIP ) using primary osteoblasts from the craniums of homozygous mutants and their wild-type littermates ( Figure 7B and C , and Table S3 ) . In Figure 7B , we identified a potential tnni2-binding fragment ( 471 bp , referred to as Part 1 ) residing in the Hif3a promoter region . The Part 1 consists of eight highly conserved sequences ( Figure 7C ) . Our data suggested that wild-type and mutated tnni2 protein associate with the endogenous Hif3a promoter at similar strengths ( Figure 7B ) . The binding of wild-type and mutated tnni2 to the Hif3a promoter was further verified by two in vitro experiments , DNA pull-down assays ( DPDA ) ( Figure 7D ) and electrophoretic mobility shift assays ( EMSA ) ( Figure 7E ) . DPDA showed that the Part 1 was capable of pulling down both wild-type and mutant tnni2 from nuclear extracts of wild-type tnni2-GFP 293T stable line and tnni2K175del-GFP stable line . In addition , EMSA and Super-EMSA verified the physical interaction between tnni2 and the Part 1 fragment . Consistent with the ChIP results , the in vitro binding assays confirmed that wild-type and mutated tnni2 exhibited a comparable abilities to bind the Hif3a promoter . Furthermore , to test whether tnni2 was capable of activating the transcription of Hif3a through the Part 1 , we constructed the whole length fragment ( −2200 to 0 Hif3a promoter ) , the truncated fragment ( −1673 to 0 Hif3a promoter ) and Part 1 fragment ( −2200 to −1674 Hif3a promoter ) luciferase reporter vectors , and performed luciferase reporter assays . Our data showed that the whole length fragment and the Part 1 fragment both had significantly stronger luciferase activity compared with the truncated fragment ( Figure 7F ) , and Tnni2K175del mutant exhibited a much stronger transactivation capacity ( more than 2 folds ) than its wild-type counterparts under normoxia ( Figure 7F ) and hypoxia ( Figure S15 ) . This finding indicated that the Tnni2K175del mutation enhanced the transactivation of tnni2 for Hif3a expression . To further validate the effect of Tnni2k175del mutant protein on expression of Hif3a , Hif1a and Vegf in mutant bone , we performed western blot analyses using growth plates of radii and ulnae from P12 Tnni2+/K175del mutant mice and their wild-type littermates , primary osteoblasts from calvariae of newborn Tnni2K175del/K175del mutants and wild-type littermates , and primary chondrocytes from radii and ulnae of newborn Tnni2K175del/K175del mutants and wild-type littermate mice , respectively . The results consistently showed that the expression of hif3a was significantly increased , vegf was reduced and hif1a was not evidently altered in Tnni2K175del mutant growth plate and cells ( Figure 7 G , H and Figure S16 ) . Next , to show in vivo the potential interaction between tnni2 , hif3a , hif1a and vegf , we performed a set of immunostaining on consecutive sections of radii and ulnae from P12 heterozygous mutant mice and wild-type controls . The staining displayed that the four proteins all were localized in resting , immature-proliferating , pre-hypertrophic and hypertrophic chondrocytes of the growth plates ( Figure 7L ) . Moreover , to assess skeletal cell-specific role of the mutant tnni2 , we conducted MTT and ELISA assays using primary osteoblasts harboring the Tnni2K175del/K175del mutant protein . Our data showed that the Tnni2K175del/K175del protein led to a reduction in proliferation and in secreted vegf levels of mutant osteoblasts ( Figure 7I and J ) . Von Kossa staining displayed that no appreciable effect on mineral deposition in mutant cells compared to that of control cells ( Figure 7K ) . Furthermore , we tested the expression of differential osteoblast markers using qPCR . The data showed that Ocn , Spp1 and Ibsp were no significant difference between mutant and wild-type osteoblasts ( Figure S17 ) . These data suggested that the mutant tnni2 did not seem to affect differentiation of mutant osteoblasts . Taken together , our data indicated that the mutant tnni2 protein had a higher capacity to transactivate Hif3a than the wild-type protein , resulting in a reduction in vegf expression in mutant bone .
In this study , we presented genetic evidence that the Tnni2K175del mutation impaired bone development and disturbed hif-vegf signaling in DA2B mice . Hypoxia-inducible factors ( HIFs ) belong to a group of heterodimeric transcription factors consisting of HIF1A , HIF2 A , HIF3A and HIF1B HIF1A or HIF2A binds to HIF1B and induces the transcription of targeted genes , such as VEGF , under hypoxic conditions . Unlike HIF1A and HIF2A , HIF3A can negatively regulate HIF1-mediated gene expression by combination with HIF1A or HIF1B [21] , [23] . In the human renal cell carcinoma cells and osteosarcoma cells , overexpressed HIF3A evidently inhibited HIF1-induced VEGF expression [30]–[31] . In mouse , ipas , a main isoform of hif3a , is induced under hypoxic condition [23] . Makino et al . reported that ectopic expression of Ipas in hepatoma cells reduced vegf production under condition of hypoxia and significantly decreased tumor vascular density . Moreover , obvious angiogenesis was observed in mouse corneas treated with IPAS antisense oligonucleotide [23] . The results strongly suggested that HIF3A suppressed angiogenesis through negative regulation of VEGF expression . In this study , we found that the growth plates of mice expressed Hif3a . And the hif3a was significantly elevated in Tnni2K175del mutant growth plates , primary osteoblasts and chondrocytes . However , the hif1a expression was not different between mutant groups and their wild-type littermate groups . Together , our results suggested that hif3a might inhibit hif1a-induced Vegf expression in the mutant growth plates and cells . The suggestion was further supported by the data that tnni2 , hif3a , hif1a and vegf being localized in the same chondrocytes zones of the growth plates ( Fig . 7L ) . Our data indicated that tnni2 regulates Hif3a expression at transcriptional level . The Tnni2K175del mutant protein had a higher capacity to transactivate Hif3a expression than wild-type protein . Our results showed that both wild-type and Tnni2K175del mutant protein had a comparable capacity of attaching the Part 1 fragment ( Figure 7C ) locating at the promoter of mouse Hif3a gene . HIF1A-VEGF signaling mediates angiogenesis in osteogenesis and regulates the differentiation of chondrocytes and the proliferation of osteoblasts [17] , [18] , [24]–[27] , [32] . In HIF1A or VEGF deficient mice , abundant researches have reported the impairment of bone development including diminished angiogenesis , delayed primary ossification centers , decreased mineralized extracellular matrix in chondrocytes , reduced osteoblasts proliferation and reduced trabecular bone formation [17] , [18] , [24]–[27] , [32] . In the current study , Tnni2K175del mice showed bone development deficiencies that were similar to mice lacking Hif1a or Vegf . These data strongly suggested that the impairment of bone development in the Tnni2K175del mutant mice was mediated by the reduced Vegf expression . This suggestion was supported by the partial recovery of trabecular bone density and the number of osteoblasts when the Tnni2K175del mutant mice were treated with recombinant mouse vegf ( Figure 6 K , L and M ) . Taken together , our data suggested that the abnormal bone development associated with the Tnni2K175del mutation could be partly attributed to the disturbed Hif-Vegf signaling . In addition , Vegf and its receptor Flt1 are crucial for osteoclast activity [19] , [26] . However , we did not observe an evident difference in the number or size of osteoclasts between E16 . 5 homozygous mutant radii and wild-type controls using the tartrate-resistant acid phosphatase ( TRAP ) activity test ( Figure S18A ) . We also tested expression of Trap and Mmp9 , two marker genes of activated osteoclasts , in the radii and ulnae of homozygous mutant mice and their wild-type littermates . However , their expression had not obvious change between mutant bones and wild-type controls ( Figure S18B and C ) . Besides HIF1A , CBFA1 , also known as RUNX2 , is important for regulating VEGF expression in bone development [33] . However , our microarray and immunostaining analyses did not show significant difference in the Runx2 expression in the radii and ulnae from 1-day-old and E16 . 5 homozygous mutants compared to their wild-type controls ( Figure S19 ) . In the study , we described that tnni2 expressed in the chondrocytes and osteoblasts of long bone growth plate of mouse . Furthermore , we observed a decreased angiogenesis in the cartilage plate of Tnni2K175del bones . The observation was in the agreement with the report from Marsha , et al [34] . Their findings showed that TNNI2 was present in human cartilage and angiogenesis in corneas was inhibited through administration of TNNI2 . Notably the extent of suppression of angiogenesis in corneas which was administrated with TNNI2 was similar to that of ipas in corneas [23] , suggesting that TNNI2 inhibited angiogenesis in corneas might be mediated by HIF3A Interestingly , three independent reports have shown that TNNI2 acted as an inhibitor of angiogenesis to inhibit tumor growth and metastasis [35]–[37] . However , the molecular mechanism by which TNNI2 suppresses angiogenesis in tumors remains unclear to date . Our findings may provide a clue for the pathological process . In this study , we focused on the mechanism causing the small body size phenotype of DA2B mice . Our findings showed that the Tnni2K175del mutation led to small body size in DA2B mice by impairing bone development . However , it needs to further investigate whether the small body size phenotype is also resulted from other abnormal tissue function . Besides TNNI2 , the patients with mutations of MYH3 also show the short stature phenotype [38] . Yet , the molecular mechanism is unknown . Recently a study reported that myosin directly interacted with runx2 in rat osteoblasts and that the expression of myosin was associated with osteoblast differentiation in vitro [39] . These data suggests that MYH3 was possibly involved in bone development as well . In summary , our data indicated that the disease-associated Tnni2K175del mutation resulted in an abnormal increase in hif3a and a reduction in Vegf expression in bone . Furthermore , the decreased vegf impeded bone development and contributed to the small body size phenotype of DA2B mice . Our findings demonstrated , for the first time , a novel role of Tnni2 in the regulation of bone development in mice . These findings might provide an insight into the short stature in human DA2B disorder .
All animal procedures were approved by the Institutional Animal Care and Use Committee of National Institute of Biological Sciences ( NIBS ) ( Beijing , China ) . Mouse R1 genomic DNA was used as a DNA template to amplify the Tnni2 gene . The mutation of c . 523–525del-aag ( K175del ) in Tnni2 was produced by overlapping-PCR . The 5′ homology arm ( long arm ) consists of a 2-kb promoter region and the entire Tnni2 genomic DNA sequence . The 3′ homology arm ( short arm ) begins 400 bp downstream of the Tnni2 . 3′ UTR and extends to 2-kb-downstream sequence . The ploxp vector was used as the targeting vector . An EcoR V site was added to the reverse primer for the short arm of recombination to facilitate Southern blot analysis . The targeting construct was introduced into 129Sv embryonic stem cells by electroporation . Positive ESC lines were selected following G418 screening and were confirmed by PCR-sequencing . Correctly targeted clones were injected into blastocysts of C57BL/6 mice and transferred into uteri of pseudopregnant females . Chimeric males were mated with ICR mice , and PCR-sequencing and Southern blot were used to determine germ line transmission of the targeted allele . cDNA from the muscle tissue of Tnni2K175del/K175del mice was sequenced to validate the Tnni2 transcript with the del-aag mutation . Expression of Tnni2 in the brains , hearts , lungs , stomachs , spleens , skeletal muscles , kidneys , radii , and ulnae at E15 . 5 was detected using in situ hybridization as previously described [40] . A Tnni2 cDNA fragment of 546 bp was subcloned into pBluescript . Sense and antisense 35S-labeled RNA probes were generated by T7 and SP6 polymerases , respectively . The activity of the probes was approximately 2×109 disintegrations/minute/µg . Frozen sections ( 10 µm ) of organs of wild-type mice at E15 . 5 were mounted onto poly-L-lysine-coated slides and fixed in 4% paraformaldehyde diluted in PBS at 4°C for 15 minutes . After hybridization and washing , the slides were incubated with RNase A ( 20 µg/mL ) at 37°C for 15 minutes . The hybrids with RNase A resistance were detected after 1 week of autoradiography using Kodak NTB-2 liquid emulsion . The slides were post-stained with hematoxylin and eosin . To observe the development of the whole skeleton , we consecutively stained skeletons of wild-type and mutant mice from E14 . 5 to E18 . 5 and from P0 to P20 for 24 hours with Alcian blue ( pH 5 ) . The skeletons were fixed for 6 hours in 95% ethanol followed by staining with Alizarin Red S . The clarification of soft tissue was performed using 1% potassium hydroxide . Each bone staining experiment was performed at least using two pair of samples . For histological examination , the tibiae , radii , ulnae , and skulls from mutant mice and wild-type littermates were fixed in 4% paraformaldehyde . Paraffin sections and ice sections ( 4 µm thickness ) were stained with HE for histological structures . Von Kossa method was used to stain mineralized trabecular bones and cartilage matrix . Masson's trichrome and Picrosirius-polarization staining was performed to test type-I collagen at osteoid at P6 , P7 , and P12 mice . Alkaline phosphatase staining showed osteoblasts in E15 . 5 mice ( Alkaline Phosphatase kit , Sigma-Aldrich , USA ) . Alcian blue was used to stain chondrocytes at P12 , P13 , and P19 mice . Immunohistochemical analyses were conducted using anti-runx2 ( 1/50 , Santa Cruz Biotechnology , USA ) , anti-CD31 ( 1/100 , Santa Cruz Biotechnology , USA ) , anti-hif3a ( 1/50 , Abcam Biotech , Cambridge , UK ) , anti-hif1a ( 1/50 , Abcam Biotech , Cambridge , UK ) , anti-vegf ( 1/50 , Abcam Biotech , Cambridge , UK ) , anti-tnni2 ( 1/40 , Merck Chemicals , Nottingham , UK ) at E15 . 5 to E18 . 5 , P0 , P6 to P8 , and P12 mutant mice and wild-type littermates . Whole radii and ulnae from E16 . 5 homozygous mutant and wild-type littermates were stained for TRAP activity using Sigma-Acid Phosphatase and Tartrate Resistant Acid Phosphatase Kit . TRAP-positive cells on bone surfaces that contained more than two nuclei were counted as osteoclasts . These histological and immunohistochemical analyses were performed using at least three biological repeats . Osteoblasts were obtained from calvariae of newborn Tnni2K175del/K175del mutants and wild-type littermates . The experiment was performed as previously described [41] . Briefly , calvariae from newborn homozygous mutants and their wild-type littermates were carefully removed . They were digested for 2 hours at 37°C in DMEM containing 0 . 2% collagenase type I . Cells were isolated and cultured in DMEM containing 10% FBS . The primary osteoblasts from neonatal mice were inoculated at a density of 105 cells per dish in 6-cm dishes . The primary osteoblastic cells were placed at a density of 5×104 cells per well in a 24-well plate and cultured for 48 hours and ALP activity were assessed using the Sigma-Aldrich Alkaline Phosphatase kit . The primary wild-type and homozygous mutant osteoblasts were cultured in the presence of 10 mM β-glycerophosphate and 50 µg/mL ascorbic acid to 14 days , and then were stained using Von Kossa method to test minerization and differential osteoblast markers . Primary osteoblasts were grown on cover glasses under 1% O2 or 20% O2 conditions for 30 hours . The cells were fixed with 4% paraformaldehyde for 2 hours . After incubation with 0 . 5% Triton X-100 and 2 . 5% BSA for 4 hours , the osteoblasts were stained with an antibody against tnni2 ( Santa Cruz Biotechnology , USA ) overnight at 4°C followed by incubation with a fluorescent antibody for 1 hour . DNA was stained using 1 µg/mL DAPI . P12 and P13 heterozygous mutants and wild-type littermates were injected intraperitoneally with BrdU at 200 µg/g of body weight for 2 hours and were then sacrificed for BrdU staining ( n = 4 ) . BrdU labeled osteoblasts were assessed ( BrdU positive osteoblast counts/area ) using Image-Pro Plus 6 . 0 software . A subcutaneous injection of recombinant mouse vegf165 ( 50 ng/g , PeproTech , USA ) was performed on Tnni2+/K175del mice ( n = 4 ) from day 4 to day 12 . Heterozygous littermate controls were treated with NaCl ( 50 ng/g ) . Treated heterozygous mice and their heterozygous littermate controls were sacrificed at 13 days of age , and the radii and ulnae were collected for H&E staining . The trabecular bone density and number of osteoblasts was assessed using Image-Pro Plus 6 . 0 software . The trabecular bones/area was defined as the density of trabecular bones and the number of osteoblasts/density of trabecular bone was used to assess relative osteoblast number . Total RNA from the entire radii and ulnae of newborn Tnni2K175del/K175del mice ( n = 3 ) and wild-type littermates ( n = 3 ) was extracted using TRIZOL reagent . The mRNA was labeled , hybridized and scanned by CaptialBio ( Beijing , China ) . Affymetrix Mouse Gene 1 . 0 ST Array was used to generate mRNA expression profile data . Differentially expressed genes associated with the Tnni2K175del mutation were analyzed using the SAM software 3 . 02 . The GEO ID of the expression profile data is GSE32163 . Gene enrichment analyses were performed on our Affymetrix Mouse Gene 1 . 0 ST microarray expression dataset of wild-type and mutant bones using GSEA 2 . 0 software [42] . An enrichment score and nominal P value were obtained for the genes that are be induced in human arterial endothelial cells following exposure to hypoxia condition [16] . Total RNA was extracted from the radii and ulnae of newborn Tnni2K175del/175K mice and wild-type littermates using TRIZOL reagent . Reverse transcription was performed using the M-MLV reverse system ( Promega , USA ) . mRNA expression of Hif-associated signaling molecules , Ihh , Pthrp , Col10a1 , Comp , Ocn , SPP1 and Ibsp were quantified using qPCR . Data were analyzed using the 2 ( -Delta Delta C ( T ) ) method . Primary osteoblasts from newborn Tnni2K175del/K175del mice ( n = 3 ) and wild-type littermates ( n = 3 ) were cultured as described above for ChIP-PCR . ChIP was performed as previously described [43] . Approximately 1×107 osteoblasts from homozygous mice and wild-type littermates were prepared . Osteoblasts were cross-linked in 1% formaldehyde solution at 37°C for 10 min followed by 1/20 volume of 2 . 5 M glycine to stop the crosslinking . The cells were rinsed twice with ice-cold PBS , transferred to 15 mL conical centrifuge tubes , and spun for 4 min at 3000 rpm . Cell pellets were re-suspended in 10 mL of LB1 ( 50 mM Hepes–KOH , pH 7 . 5; 140 mM NaCl; 1 mM EDTA; 10% glycerol; 0 . 5% NP-40; 0 . 25% Triton X-100 ) . The samples were rocked at 4°C for 10 min followed by a spin at 2000 rpm at 4°C for 4 min . The pellets were re-suspended in 10 mL of LB2 ( 10 mM Tris–HCl , pH 8 . 0; 200 mM NaCl; 1 mM EDTA; 0 . 5 mM EGTA ) . The samples were gently rocked at 4°C for 5 min . The nuclei were pelleted and collected by spinning at 2000 rpm at 4°C for 5 min . The pellets were resuspended in 3 mL LB3 ( 10 mM Tris–HCl , pH 8; 100 mM NaCl; 1 mM EDTA; 0 . 5 mM EGTA; 0 . 1% Na–deoxycholate; 0 . 5% N-lauroylsarcosine ) . The samples were sonicated using the following settings: 30 cycles of 15 s ON and 59 s OFF , power-output 31 watts . 10% Triton X-100 was added to sonicated lysate , followed by a 10 min spin at 12 , 000 rpm at 4°C . The supernatants were combined and eluted with 3-fold volume LB3 . A subsample of 100 µL of cell lysate from sonication was saved as an input control and stored at −20°C . A total of 100 µL magnetic beads ( Invitrogen , California , USA ) was added to 30 µg of TNNI2 goat polyclonal IgG antibody ( Santa Cruz , Biotechnology , USA ) and 1 mg of goat serum , and incubated overnight at 4°C . The beads were washed twice in 1 mL block solution ( 0 . 5% BSA ( w/v ) in PBS ) . The beads were then collected using a magnetic stand and re-suspended in 100 µL block solution . A total of 100 µL of antibody/magnetic bead mix was added into cell lysates and gently mixed overnight at 4°C . The beads were collected with a magnetic stand and washed with 1 mL RIPA buffer ( 50 mM Hepes-KOH , pH 7 . 5; 500 mM LiCl; 1 mM EDTA; 1% NP-40; 0 . 7% Na-deoxycholate ) four times , followed by one wash with 1 mL TBS ( 20 mM Tris-HCl , pH 7 . 6; 150 mM NaCl ) . The beads were then spun at 1500 rpm for 3 min at 4°C to remove any residual TBS buffer using the magnetic stand . A total of 200 µL of elution buffer ( 50 mM Tris-HCl , pH 8; 10 mM EDTA; 1% SDS ) was added into each IP beads and 50 µL input control to reverse crosslink at 65°C for 16 hours . A total of 200 µL TE was added with 8 µL of 1 mg/mL RNaseA to each IP bead and input control DNA at 37°C for 30 min . A total of 4 µL 20 mg/mL proteinase K was added and incubated at 55°C for 2 hours . DNA was extracted with 400 µL phenol-chloroform isoamyl alcohol and precipitated DNA with cold ethanol . PCR was performed to test potent tnni2 binding sequences . PCR reactions contained 0 . 15 ng IP DNA and input control DNA . The primers for PCR reactions were shown in Table S3 . The PCR products were tested using electrophoresis and Sanger sequencing . cDNA fragments of wild-type and Tnni2K175del mutation were subcloned into an EGFP-N1 vector to produce wild-type tnni2-GFP and mutated tnni2-GFP proteins , respectively . 293T cells were transfected with 20 µg mutated vector and the same amount of Tnni2 wild-type vector . 1 mg/mL of G418 was used to screen anti-G418 cells for 2 weeks . GFP-positive cells were sorted by Flow Cytometry . Wild-type and mutated Tnni2 cDNA fragments were subcloned into pcDNA 3 . 1 vector . The truncated fragment ( −1673 to 0 Hif3a promoter ) , the whole length fragment ( −2200 to 0 Hif3a promoter ) and the Part 1 fragment ( −2200 to −1674 Hif3a promoter ) were obtained by PCR amplification from mouse genomic DNA and subcloned into pGL3 luciferase reporter vector . 293T cells were respectively transfected with 100 ng mutated tnni2K175del vector , wild-type tnni2 vector , hif3a ( −1673 to 0 ) pGL3 vector , hif3a ( −2200 to 0 ) pGL3 vector , hif3a ( −2200 to −1674 ) pGL3 vector and pEGFP vector ( as negative control ) , combined with 5 ng pRL-TK reporter vector and incubated under 1% or 20% O2 conditions for 30 hours . Dual reporter experiments were performed using the Dual-Glo luciferase assay system ( Promega , Wisconsin , USA ) . The luminescence signals were detected using a GloMax-96 Microplate Luminometer ( Promega , Wisconsin , USA ) . Firefly Luciferase was measured 30 hours later and normalized by Renilla Luciferase ( pRP-TK ) for transfection efficiency . Total cytoplasm extracts ( CE ) and nuclear extracts ( NE ) from 293T tnni2 wild-type-GFP stable cells were prepared according to the following method: Buffer A ( 10 mM HEPES , pH 7 . 9 , 1 M HEPES , 10 mM KCL , 2 M KCL , 0 . 1 mM EDTA , 0 . 5 M EDTA , 0 . 1 mM EGTA ) was added with 1 mM PMSF and 1 mM DTT into the cell pellet and then suspension was vortexed at 4°C for 10 min . After a spin at 3000 rpm for 5 min at 4°C , supernatant ( CE ) was collected and 5-fold volume buffer C ( 10 mM HEPES pH 7 . 9 , 500 mM NaCl , 0 . 1 mM EDTA , 0 . 1 mM EGTA , 0 . 1% NP40 , 10% glycerol ) was added with 1 mM PMSF and 1 mM DTT to isolate the nuclear extract . The lysate was rotated for 20 min at 4°C and centrifuged at 12 , 000 rpm for 20 min at 4°C . The supernatant ( NE ) was transferred to a new tube . 20 mg of CE or NE was incubated with 200 µL agarose beads ( Invitrogen , California , USA ) that were coupled with an antibody against GFP protein at 4°C for 4 hours , followed by a spin at 1000 rpm for 5 min at 4°C . The supernatant was removed , and the agarose beads were resuspended with 1 mL Buffer C three times . The agarose beads were boiled in 1× protein loading buffer at 95°C for 30 min to separate the protein from the beads . After centrifugation at 12 , 000 rpm for 20 min at room temperature , the supernatant was collected . The IP products were used for mass spectrometry analyses . Nuclear extracts of wild-type tnni2-GFP 293T stable cell lines and tnni2K175del-GFP stable cell lines were prepared as described above . The protein-DNA binding reactions were performed using the Thermo Scientific LightShift Chemiluminescent EMSA Kit . 20 mg of nuclear extract and 10 fmol of biotin-labeled probe were used for each binding reaction . For the competition assay , 5 pmol of cold probe was incubated with the samples at room temperature for 10 min prior to the addition of biotin-labeled probes . For the super-shift assay , 2 µg anti-GFP monoclonal antibody ( Sigma ) was used for each reaction . Incubated samples were separated by 6% non-denaturing PAGE at 100 V on ice for 80 min and transferred to a nitrocellulose filter membrane at 350 mA on ice for 40 min . The sequences of the 5′ biotin-labeled probe used for EMSA and Super-EMSA was: forward , 5′-ccctgtctagaatcccccagtgaggggctgggggcgtggctcagtggtagagc-3′ , reverse , 5′-gctctaccactgagccacgcccccagcccctcactgggggattctagacaggg- 3′ . 10 mg of the 5′ biotin-labeled probe ( the same as the probe used in EMSA assay ) and 100 µL avidin magnetic beads ( Invitrogen , Dynabeads ) were added into a 1 . 5 mL centrifuge tube with 0 . 5 mL buffer C ( refer to IP assays above ) and incubated at room temperature for 1 hour . The beads were collected using a magnetic stand . 20 mg of NE from wild-type tnni2-GFP and tnni2K175del-GFP 293T stable lines were added into the centrifuge tubes with a magnetic bead-probe and they were incubated at room temperature for 1 hour . The supernatant was removed , and the magnetic beads were collected with a magnetic bead stand . The beads were carefully washed with buffer C three times . The magnetic beads were boiled using 80 µL 1×protein loading buffer at 95°C for 30 min to separate the protein from the beads . Following centrifugation at 12 , 000 rpm for 20 min at room temperature , the supernatant was collected . 40 mL of IP products from wild-type and mutated NE were examined by western blotting using an anti-GFP monoclonal antibody ( Sigma ) . The products isolated from beads that were incubated only with wild-type and mutant NE were used as negative controls . The NE from wild-type TNNI2-GFP were used as Input . The experiment was performed using four independent biological repeats . Protein concentrations of soluble vegf were determined using the Valukine ELISA kit for mouse VEGF ( R&D Systems ) . Cell culture supernatants ( Tnni2K175del/K175del and wild type primary osteoblast cells ) from triplicates of three different experiments were harvested after 18 hours , 24 hours and 36 hours respectively , centrifuged with 12 , 000 rpm for 5 min and stored at −20°C . Each corresponding well was subsequently trypsinized and the numbers of live cells were counted to permit the appropriate correction of secreted vegf . ELISA assays were performed according to the manufacturer's instructions . Cell proliferation assays were performed with a CellTiter 96 AQueous Proliferation Assay kit ( Promega , Wisconsin , USA ) following the manufacturer's instructions , and the plates were measured with a Molecular Devices plate reader . Proteins were separated by SDS-PAGE and transferred to a PVDF membrane ( Millipore , Billerica , USA ) . The membrane was blocked with 5% non-fat milk and incubated with rabbit anti-HIF3A ( 1/500 , Abcam Biotech , Cambridge , UK ) , rabbit anti-HIF1A ( 1/500 , Abcam Biotech , Cambridge , UK ) , rabbit anti-VEGF ( 1/1000 , Abcam Biotech , Cambridge , UK ) , rabbit anti-TNNI2 ( 1/500 , Merck Chemicals , Nottingham , UK ) , Rabbit ant-histone H3 ( 1/800 , Cell Signaling Technology , Danvers , USA ) or mouse anti-β-actin ( 1/10000 , Sigma , St Louis , USA ) antibodies . Data were processed using SPSS 17 . 0 and denoted as the mean±s . d . A comparison between the quantitative data of the two groups was performed using Student's t-test ( two-tailed distribution ) . Statistical significance was indicated by P<0 . 05 . | Distal arthrogryposis type 2B ( DA2B ) is an autosomal dominant genetic disorder . The typical clinical features of DA2B include hand and/or foot contracture and shortness of stature in patients . To date , mutations in TNNI2 can explain approximately 20% of familial incidences of DA2B . TNNI2 encodes a subunit of the Tn complex , which is required for calcium-dependent fast twitch muscle fiber contraction . In the absence of Ca2+ ions , TNNI2 impedes sarcomere contraction . Here , we reported a knock-in mouse carrying a DA2B mutation TNNI2 ( K175del ) had typical limb abnormality and small body size that observed in human DA2B . However , the small body did not seem to be convincingly explained using the present knowledge of TNNI2 associated skeletal muscle contraction . Our findings showed that the Tnni2K175del mutation impaired bone development of Tnni2K175del mice . Our data further showed that the mutant tnni2 protein had a higher capacity to transactivate Hif3a than the wild-type protein and led to a reduction in Vegf expression in bone of DA2B mice . Taken together , our findings demonstrated that the disease-associated Tnni2K175del mutation caused bone defects , which accounted for , at least in part , the small body size of Tnni2K175del mice . Our data also suggested , for the first time , a novel role of tnni2 in the regulation of bone development of mice by affecting Hif-vegf signaling . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"mutation",
"genetics",
"biology",
"and",
"life",
"sciences",
"human",
"genetics",
"genetics",
"of",
"disease"
]
| 2014 | A Gain-of-Function Mutation in Tnni2 Impeded Bone Development through Increasing Hif3a Expression in DA2B Mice |
Over the past several years fungal infections have shown an increasing incidence in the susceptible population , and caused high mortality rates . In parallel , multi-resistant fungi are emerging in human infections . Therefore , the identification of new potential antifungal targets is a priority . The first task of this study was to analyse the protein domain and domain architecture content of the 137 fungal proteomes ( corresponding to 111 species ) available in UniProtKB ( UniProt KnowledgeBase ) by January 2013 . The resulting list of core and exclusive domain and domain architectures is provided in this paper . It delineates the different levels of fungal taxonomic classification: phylum , subphylum , order , genus and species . The analysis highlighted Aspergillus as the most diverse genus in terms of exclusive domain content . In addition , we also investigated which domains could be considered promiscuous in the different organisms . As an application of this analysis , we explored three different ways to detect potential targets for antifungal drugs . First , we compared the domain and domain architecture content of the human and fungal proteomes , and identified those domains and domain architectures only present in fungi . Secondly , we looked for information regarding fungal pathways in public repositories , where proteins containing promiscuous domains could be involved . Three pathways were identified as a result: lovastatin biosynthesis , xylan degradation and biosynthesis of siroheme . Finally , we classified a subset of the studied fungi in five groups depending on their occurrence in clinical samples . We then looked for exclusive domains in the groups that were more relevant clinically and determined which of them had the potential to bind small molecules . Overall , this study provides a comprehensive analysis of the available fungal proteomes and shows three approaches that can be used as a first step in the detection of new antifungal targets .
There has been a significant rise in the incidence of fungal infection over the last few years . This has been partially due to an increase in the susceptible population as the result of blood cancer , intensive care , solid organ transplantation , or chronic granulomatous disease , in addition to a growing number of patients receiving high doses of corticosteroids or other immunosuppressive treatments [1] , [2] . The most common cause of invasive fungal infection ( IFI ) is Candida spp . followed by Aspergillus spp . [3] , [4] . The mortality of candidemia ranges from 44% to 75% [1] , [5] and that of aspergillosis is around 60% [4] . IFI constitutes even a more important problem for immunocompromised patients , with mortality reaching up to 90% [6] . It has also been observed that new fungal species are part of the etiology of these infections , the common characteristic being an enhanced resistance to known antifungal agents , which further complicates the management of these infections . The success of available antifungal therapies is limited by the drug toxicity for the host ( resulting potentially in severe side effects ) and the late detection and treatment of the infection [7] . The cost of patient hospitalization , due to the usual long period needed for the currently available drugs to be efficient , is another factor to consider . Therefore , the discovery of new drug targets and antifungal drugs with a broader spectrum of activity is one active field of research . The rationale in the search for new antifungal targets in the pre-genomics era was based on the molecular study of genes associated with fungal viability or virulence . The advent of massive parallel sequencing technologies and their progressively reduced cost has enabled the sequencing of large numbers of genomes in a short period of time . As a result , it is now possible to look for broad-spectrum antifungal targets by detecting homologous proteins present in most of the available fungal proteomes [8] . Approaches based on sequence similarity ( e . g . BLAST [Basic Local Alignment Search Tool] ) have some difficulties when trying to detect homologous proteins present in distantly related species , when sequence similarity is less than 30% [9] . However , protein domain composition is likely to be conserved throughout evolution due to functional constraints [10] , so it can be used for this purpose . In the late 90s , the concept of protein domain was coined by Branden and co-workers to define an independent , compact and stable protein structural unit that folds independently of other such units [11] . Most proteins in eukaryotes are composed of different domains , which are independent units with potentially different biological functions . The term domain has also been used to name blocks of protein sequences highly conserved throughout the course of the evolution . The order in which these domains are arranged within the protein sequence constitutes its domain architecture . Proteins having the same domain architecture are likely to have similar structures and therefore preserve the same cellular function [12] . The emergence of proteins with new domain and/or domain combinations is thought to be a major mechanism of evolution since it can give rise to new functions for the organisms [13] , [14] . In this context , the term ‘merology’ has been proposed to describe the fact that the different units of multidomain proteins have different evolutionary history . Correspondingly , ‘merologous’ proteins refer to non-homologous proteins that display the same domain organization [15] . Whereas the domain repertoires of different organisms can be relatively similar , there can be a high variation in the number of the domain combinations . Therefore , the number of different domain architectures is more variable and is related to the organism complexity and lifestyle [12] , [16] . In fact , the emergence of animals and vertebrates has been associated with the appearance of novel domain combinations [13] , [17] . Domain promiscuity is a measurement of the capability of a protein domain to combine with other domains . For instance , it has been shown that the high mobility of protein structural units embraces the capacity to interact with other proteins and might hold important roles in Protein-Protein Interaction ( PPI ) networks [18] . Intuitively , a protein domain to be considered promiscuous must appear in several domain combinations within a proteome . However , different studies have observed that the number of domain combinations correlates positively with the degree in which such domain is spread over a genome [11] . Nevertheless , most approaches used to analyse domain promiscuity normalize the number of distinct combinations by the number of times a domain is found in the genome . Protein domain information can be used for many different purposes . For instance , the domain architecture can determine the overall protein function and be used to transfer genomic annotations in newly sequenced genomes [19] , using automatic annotation approaches . Domain configurations can also be exploited to determine the phylogeny of complete genomes , achieving comparable results than those using more sophisticated methods [20] . Furthermore , protein domains can be studied in the context of their ability to bind small molecules , since they can represent targets for biologically important ligands including potential drugs . One example of this approach is a recent study [21] that predicted in silico the potential of ligands and small molecules included in the ChEMBL database [22] to bind protein domains included in Pfam [7] . The underlying assumption was that ligands and small molecules could bind structurally conserved protein regions , supported by the reported observation that 88 . 4% of annotated binding sites from UniProtKB/Swiss-Prot rested entirely within the boundaries of a given Pfam domain [21] . In this manuscript , we first perform an analysis of the protein domains and domain architectures derived from all the available complete fungal proteomes . We identify the core and exclusive domains and domain architectures at different levels of the fungal taxonomical classification ( phylum , subphylum , order , genus , species ) and characterize which domains are found to be promiscuous . Then , we use the obtained protein domain information to explore in silico approaches to detect potential candidate drug targets , using information coming from different public resources and previous studies . In addition , we also classify some fungal species according to their occurrence in clinical samples and provide the related exclusive domain/domain architecture related information .
Proteins from all the available complete fungal proteomes ( 137 organisms ) were retrieved from UniProtKB ( UniProt Knowledgebase , release 2013_01 ) . The 137 fungal proteome sets represented 111 species of the following four fungal phyla: Ascomycota , Basidiomycota , Chytridiomycota and Zygomycota . In addition , the phylum Microsporidia was also considered . Microsporidia include a controversial group of eukaryotic organisms of fungal origin , without mitochondria and peroxisomes . Microsporidia have been considered in this study since they are considered the earliest-diverging clade of sequenced fungi [23] and some of them are causative agents of disease in immunocompromised humans [24] . Specifically , the vast majority of organisms analysed belonged to the Ascomycota phylum ( 112 proteomes , containing 88 different species ) followed by Basidiomycota ( 18 proteomes , containing 16 species ) , Microsporidia ( 5 proteomes coming from the same number of species ) , and Chytridiomycota and Zygomycota , each of them with only one representative species . Proteomes from different strains of the same species were included in the study e . g . Saccharomyces cerevisiae ( 13 strains ) or Ajellomyces capsulata ( 4 strains ) . For a complete list of species and strains , see Table S1 . Overall , 1 , 191 , 070 proteins were analysed , with an average of 8 , 694 proteins per proteome . The related domain information for each protein was obtained from Pfam ( release 27 . 0 ) . On average , 67 . 0% of the proteins had at least one Pfam domain assigned . However , the Pfam coverage was not evenly distributed among the studied species . On the low coverage side , there were plant pathogens like Melampsora laricis-populina ( strain 98AG31/pathotype 3-4-7 ) , the poplar leaf rust fungus , and Puccinia graminis sp . tritici ( CRL 75-36-700-3/race SCCL ) , the causal agent of wheat and barley stem rust . These fungi only had a 36 . 5% and 39 . 5% of Pfam coverage , respectively . The likely explanation for this lower coverage is that some of the domains are not found in the traditional model organisms and therefore not included in Pfam . On the contrary , the fission yeast Schizosaccharomyces pombe ( strain 972/ATCC 24843 ) and its close relative species S . japonicus ( strain yFS275/FY16936 ) showed a coverage over 87 . 5% . The complete coverage information per organism can also be found in Table S1 . On average , each fungal protein in the studied set had 1 . 55 domains , a proportion that did not vary significantly between species . Overall , 5 , 279 different Pfam domains were found among the studied proteomes ( 35 . 6% of the total Pfam domains ) . Among them , 131 domains ( 2 . 5% ) were found in every fungus analysed ( they are ‘core’ domains ) , whereas 612 domains ( 11 . 6% ) were found only in one organism . Interestingly , when Microsporidia organisms were not considered in the analysis , the number of fungal ‘core’ domains rose to 268 . Overall , 70 . 8% of all domain types ( 3 , 740 ) appeared in both single and multidomain proteins . In addition , 964 domains ( 18 . 3% ) were not seen in combination with any other domain , whereas 575 domains ( 10 . 9% ) appeared exclusively in multidomain proteins ( see Table S2 for the complete lists of domains in each category ) . The domain repertoire found in the different fungal species was similar , with the exception of some Microsporidia species , which contained proportionally fewer domain types ( Table S1 ) . No linear correlation was found between the number of distinct domains and the size of the proteome . From an evolutionary perspective this should not be surprising since larger proteomes can be explained by an increasing number of duplication events , creating new copies of the same domain [25] . Protein domain architectures were defined throughout this study as the ordered tuples of Pfam domains , listed from the protein N-terminus to C-terminus . Architectures with either different domain counts or order were computed as independent domain combinations , even if they had the same domain types . For instance , the representation of three potential different example domain architectures with just two domains could be: 1 ) D1∼D1∼D2; 2 ) D1∼D2; and 3 ) D1∼D2∼D1 . As indicated , all three would be considered different domain architectures because either the number of domains , the order of the domains and/or the number of domains were different . In total , 21 , 853 unique domain architectures were identified in our set of 137 fungal proteomes , with 10 , 206 of them ( 46 . 7% ) appearing exclusively in a single proteome and only 56 ‘core’ architectures ( 0 . 2% ) that were present in all the fungi . However , when Microsporidia species were not taken into account , the number of ‘core’ architectures almost doubled: 107 domain architectures were found . Gene Ontology ( GO ) term annotations were added to give an insight of their molecular function and/or biological activity ( see ‘Methods’ ) . Approximately one third of all Pfam domains ( 32% ) had one or more GO terms associated . Most of these 56 ‘core’ domain architectures ( Table S3 ) were composed of only one domain: there were 50 single domain and 6 multidomain ‘core’ architectures . As expected , the corresponding proteins are related with essential and well-conserved regulatory mechanisms such as control of vesicle formation ( Ras family ) , DNA binding , modelling patterns of anatomical development ( Homeobox domain ) , RNA metabolism , nuclear transcription and pre mRNA splicing ( DEAD/DEAH box helicases ) . DEAD/DEAH box proteins are highly conserved among a wide variety of species . While DEAD proteins participate in the translation initiation of the spliceosome , DEAH proteins are required in further stages of the splicing process like the transesterification , releasing the mRNA and degradation of some spliceosome complexes [26] . The effect of removing Microsporidia from the analysis of ‘core’ domains and architectures highlights how different Microsporidia organisms can be , when compared with the rest of the organisms analysed . To measure how significant this increase is , we compared it to the effect induced by removing the species from other phylum ( Zygomycota ) . Note that the only strain from Zygomycota ( Rhizopus delemar ) had more proteins ( 16 , 968 ) than the total amount of proteins from all the Microsporidia strains together ( 11 , 973 ) . When Zygomycota were removed , the number of ‘core’ domains rose just from 131 to 132 , whereas the number of ‘core’ architectures did not change . The same pattern of modest variance in the number of domains shown in the previous section was also observed at the level of the domain architectures ( Table S1 ) . On the contrary , the ratio between unique domains and unique domain architectures in a given species is more diverse across the fungal kingdom . We analysed the counts of distinct domains and domain architectures per proteome ( Figure 1 , for a more detailed representation see Figure S1 ) . It was observed that filamentous fungi ( such as R . delemar , Fusarium oxysporum , Emericella nidulans or Neosartorya fumigata ) tend to have a larger ratio of architectures per domain , whereas yeasts ( such as Rhodotorula glutinis , S . cerevisiae or S . pombe ) in general hold less combinatorial power to create new architectures . Interestingly , the thermo-dimorphic ( found in both filamentous and unicellular form ) fungal pathogens A . capsulata ( Histoplasma capsulatum ) , Paracoccidioides brasiliensis Pb01 ( recently reclassified as P . lutzii [27] , [28] ) and Candida albicans ( unicellular yeast also able to produce true hypha ) were found in the middle section of the figure with an architecture/domain ratio value close to 1 . At the species level , 637 domains and 10 , 582 domain architectures were identified as exclusive for one of the 111 fungal species analysed . In this case , proteomes from multiple strains belonging to the same species were grouped . Throughout the rest of the manuscript , organisms belonging to the phylum Microsporidia were considered as any other fungus . In any case , the different analyses were also performed without Microsporidia . However , the results are not shown for simplification purposes since we believe they do not add much novel information , unlike the information about ‘core’ domains and architectures , included in the previous sections . The average number of exclusive architectures per species was 96 , ranging from 10 to 467 ( see below ) . All the exclusive architectures for each species are provided as Table S4 . In terms of subphyla , the 58 species from the Pezizomycotina subphylum contained an exceptional amount of exclusive domains ( 24 . 6% ) and domain architectures ( 60 . 0% ) , as can be observed in Figure 2 . Saccharomycotina , with 28 species , was also a very prominent group with a similar fraction of exclusive domains ( 7 . 4% ) and domain architectures ( 10 . 6% ) . Batrachochytrium dendrobatidis , the only representative species of the Chytridiomycota subphylum had most of the exclusive domains ( 36 . 6% ) , but with a modest combinational power ( only 4 . 2% of exclusive architectures ) . The organisms with the least number of exclusive architectures were those from the primitive Microsporidia species Encephalitozoon intestinalis and E . cuniculi , with only 10 species-specific domain architectures . Then , there were two Saccharomycotina species: Candida dubliniensis ( 12 architectures ) and C . glabrata ( 15 architectures ) . Cryptococcus gattii serotype B was the species with the lowest number of exclusive architectures in the phylum Basidiomycota ( 32 architectures ) . As a simple observation , species without other closely related species in the dataset tended to have larger numbers of exclusive domain architectures . The highest numbers were found in R . delemar ( 467 architectures ) and Batrachochytrium dendrobatidis ( 443 ) , which were the only representatives of the Zygomycota and Chytridiomycota phyla , respectively . Remarkably , two Ascomycetes: the pathogen F . oxysporum ( 424 ) and A . capsulata ( 301 ) , the causative agent of histoplasmosis [29] , showed also a high number of exclusive architectures . Another prolific species in terms of exclusive architectures was Mixia osmundae ( 262 ) , a Basidiomycete genetically highly divergent compared with other members of the subphylum Pucciniomycotina [30] . The repertoire of exclusive protein domains looked slightly different . For a complete list of exclusive domains per species see Table S5 . On average , only 6 domains were found to be exclusive for one single species . Most species ( 94 out of 111 ) showed a smaller number and sixteen of them did not even account for a single exclusive domain . Among the species with the highest amount of exclusive domains were the already mentioned unique representatives for the phylum Chytridiomycota ( Batrachochytrium dendrobatidis , 233 domains ) and Zygomycota ( R . delemar , 48 ) . In addition , the Microsporidia member with the highest number of exclusive domains was E . bienueusi , with 68 . Finally , the Ascomycete model organism Sordaria macrospora had 23 , and the tree pathogen Basidiomycete Moniliophthora perniciosa had 11 exclusive domains . A similar analysis of the exclusive domain architectures was done at the genus level . In this case , all the proteomes from the species belonging to the same genus were grouped ( Figure 3 ) . Domain architectures exclusive to one genus or shared by members of the same genus were considered clade-specific . Overall , 14 genera ( seven of them having more than one species analysed ) had more than 200 clade-specific domain architectures . Aspergillus , the best represented genus with 7 species and 9 strains in total , had the largest set of exclusive architectures ( 943 ) , containing 104 architectures shared by at least two Aspergillus species ( Figure 3 ) . A list containing the results of the analysis for all genera including more than one species is included in Table 1 . A comprehensive collection of the exclusive and core domains and domain architectures at different taxonomical levels ( phylum , subphylum , order , genus and species ) can be found in the Tables S6 and S7 , respectively . The information is split by taxonomical levels and allows the screening of conserved domains and domain architectures in the studied organisms . In particular , an interesting case is that of the Domains of Unknown Function ( DUFs ) and their related domain architectures . The specificity and exclusiveness or these domains to certain taxonomical groups could help to elucidate the underlying biological functions . Since we were interested in knowing which domains were essential for the different fungi , next we identified the list of promiscuous domains per organism , as explained in ‘Methods’ . Table 2 shows a list of protein domains commonly found among the 25 top-ranked promiscuous domains in the individual organisms . Important functions for the cell survival and interaction with the environment were associated with most of the detected promiscuous domains . The promiscuous domain found in a highest number of organisms was the ‘ATPases Associated with cellular Activities’ ( ‘AAA’ or ‘AAA+’ ) domain , found to be promiscuous in 133 out of the 137 organisms studied . It belongs to a large and intensively studied protein superfamily . ‘AAA’ domains usually have a ring shaped oligomeric complex that conveys them several activities through the energy-dependent unfolding of macromolecules [31] . Phylogenetic analyses of ‘AAA’ proteins have suggested the existence of six main clades of ‘AAA’ domains [32] , covering all kingdoms of living organisms . The second promiscuous domain most commonly found was ‘GATase’ ( 123 times ) , which is the principal component of the homonym enzyme glutamine amidotransferase , enabling the catalysis of the ammonia group from glutamine . Interestingly , after ‘AAA’ and ‘GATase’ , the most promiscuous domains identified were ‘SH3_1’ ( found in 122 organisms ) and ‘PX’ ( 117 organisms ) , two related domains which can interact with each other . ‘SH3 ‘domains are typically 40–60 amino acids long [33] and have a hydrophobic pocket that can bind proteins containing peptides rich in proline [34] . ‘SH3’ domains may modulate interactions with the cytoskeleton and the membrane . In contrast , the ‘PX’ domain has an average length of 130 amino acids and is characterized by: i ) a PxxP motif ( referenced by the ‘PX’ acronym ) recognized by the ‘SH3’ domains; and ii ) the basic residues that form a phospholipid binding pocket . Frequently , ‘PX’ domains co-appear with dimerization-related domains such as coiled-coil domains , increasing the affinity of the ‘PX’ domains to the cell membrane [35] . Both the ‘PX’ and ‘SH3’ domains have been identified in intracellular signalling pathways via protein-protein interactions [36] . Apart from the domains mentioned above , other 12 domains , including ‘PH’ , ‘SNF2_N’ , ‘Helicase_C’ , ‘MMR_HSR1’ , ‘DEP’ , ‘UBA’ , ‘TPR_1’ and ‘zf-RING_2’ , ‘C1_1’ , ‘JmjC’ , ‘UCH’ and ‘BRCT’ , were the ones found to be promiscuous in more than 50% of the studied fungi . It has been previously pointed out that protein domain promiscuity is a volatile feature throughout evolution [37] . In fact , only a few domains are consistently classified as promiscuous in organisms from all major taxonomical domains . We think that this observation fits with our results since , overall , only 20 domains were found to be promiscuous in at least 25 organisms ( out of the 137 studied , Table 2 ) . So far , in this study we have performed a detailed analysis of the protein domain and architecture content of the available fungal proteomes . The resulting information can be used with different purposes in mind . In our case , we decided to further mine the data and explore approaches to detect in silico potential targets for antifungal agents . It is well-known that undesired side effects are one of the main issues of the currently used antifungal drugs [38] , so we first aimed at finding targets that were exclusive for fungi . Since protein domains often represent different three-dimensional structures and these structures determine the binding potential of small molecules and drugs , we first tried to detect which of the domains were exclusively present in fungi and were not found in the human proteome . Protein domain information of the human reference proteome ( UniProtKB release 01_2013 ) was then analysed and compared with the information available for all the fungal organisms . At least one Pfam domain was retrieved for 72 . 8% of the human proteins , with an average of 2 . 06 domains per protein . The number of distinct Pfam domains found ( 5 , 519 , 37 . 2% of all the Pfam domains ) was slightly higher than the sum of the different domains available for all the fungal proteomes analysed ( 5 , 279 ) . However , on the other hand , almost three times more architectures were found in the combined fungal proteomes when compared with human ( 17 , 469 vs . 6 , 741 , respectively ) . The number of domains and domain architectures either exclusive or shared between human and fungi are represented in Figure 4 . The number of exclusive domains and architectures for fungi were 1 , 786 and 13 , 111 , respectively . In order to have a second view on the human and fungal domain contents , we again analysed the protein domain information , but this time at the level of the Pfam domain clans . Pfam clans consist of a series of evolutionary related Pfam families which are believed to share a common ancestor [39] . Two conditions are checked to assess whether Pfam domains are grouped in the same clan: i ) to have a related three-dimensional structure; and ii ) to have a quality matching of the HMM ( Hidden Markov Model ) profiles . In total , 469 clans comprising 4 , 563 Pfam domains ( 30 . 7% of all the Pfam domains ) were found for both human and fungi . Overall , 48 clans ( 10 . 2% of the total number of clans ) were found to be composed by Pfam domains exclusive for fungi , not present in any human protein ( Figure 4 and Table S8 ) . Based on the criterium of drug side effects , all proteins containing protein domains belonging to those 48 clans ( corresponding to 77 Pfam domains ) are in our opinion , potentially good initial candidates to be considered as antifungal targets . Using the already available information about the exclusive domains per taxonomic level , the list of potential targets could then be tailored to e . g . particular species , genera or phyla . Just as one example , one of those detected 77 Pfam domains was the domain ‘DinB’ ( PF05163 , part of the clan CL0310 ) , which was exclusively found in proteins from species from the genus Aspergillus . The ‘DinB’ domain was named after dinB , one member of the DNA damage-inducible din genes . The three din loci ( dinA , dinB and dinC ) were originally described in Bacillus subtilis as components of a global SOS-like regulatory network called SOB system [40] . It has been shown that the SOS pathway could be essential in the acquisition of bacterial mutations which lead to resistance to some antibiotic drugs [41] . This domain was only found in monodomain proteins from Aspergillus and might then represent a good potential target for anti-Aspergillus drugs . The second approach was to search for biological pathways that could be potential targets for antifungal compounds , focusing on proteins with promiscuous domains , since they might play a role in maintaining network stability [42] . In this case , we decided to use the identified promiscuous domains ( described in section 4 ) that were exclusively found in fungi ( described in the previous section ) , since promiscuous domains are believed to play a major role in cellular signalling networks . From the set of 219 fungal promiscuous domains ( present in the compiled list of the 25 top-ranked for each species , Table 2 ) , eight of them were not found in the human proteome: ‘Cas_Cas4’ ( PF01930 ) , ‘CBM_1’ ( PF00734 ) , ‘Glyco_hydro_72’ ( PF03198 ) , ‘HisKA’ ( PF00512 ) , ‘Hom_end’ ( PF05204 ) , ‘KR’ ( PF08659 ) , ‘PH_10’ ( PF15411 ) and ‘Sirohm_synth_M’ ( PF14824 ) . Overall , 3 , 675 fungal proteins in UniProtKB contained at least one of these 8 domains . We then looked for representation of these proteins in fungal pathways available in the public domain . Table 3 shows detailed information about the matched pathways in the UniPathway resource and the corresponding proteins and domain architectures found . As a result , five proteins were detected which contained one of the above mentioned eight domains , and which were represented in at least one metabolic pathway in UniPathway ( Table 3 ) . In fact , only three ( out of the eight ) domains were represented in those proteins: ‘CBM_1’ , ‘KR’ and ‘Sirohm_synth_M’ . The ‘CBM_1’ domain is a fungal cellulose-binding domain which can interact with specific structural polysaccharides [43] . The ‘Sirohm_synth_M’ domain is part of the enzyme which catalyzes the biosynthesis of siroheme , a cofactor essential for the assimilation of nitrogen and sulfur [44] . Finally , ‘KR’ is an enzymatic domain , part of the multimodular polyketide synthases , responsible of the growth of the polyketide chain . Overall , three pathways present in UniPathway , containing these five proteins , were identified as potential targets for antifungals: Unfortunately , the amount of pathway information from fungal organisms in the public domain is quite limited at present . It was not possible to find comparable ready-to-use information in other resources apart from UniPathway ( see ‘Methods’ ) . This clearly limits the applicability of this approach at present . For instance the xylan degradation pathway may not be potentially interesting for human pathogens , but it could definitely be for plant pathogens . Therefore , independently of the three pathways detected here , we think that the described approach will be more usable as more pathway related information becomes publicly available . Since annotation of fungal pathways in resources like UniPathway is limited to the most popular species , the protein domain architecture of the five proteins identified was searched in the rest of fungal organisms to know how common these three pathways were throughout the fungal kingdom . Assuming that proteins with the same domain architecture can have a similar function and can be involved in the same or similar pathways , the most ubiquitous pathway among the three was the “biosynthesis of siroheme” , present in 96 out of the 111 fungal species studied ( Table 3 ) . Analogous approaches can be used to determine the desired scope of a given antifungal potential . The 111 fungal species used in this study were classified in five groups taking into account the frequency in which they had been found in clinical samples ( see ‘Methods’ and Table 1 for the list of species and groups ) . Species belonging to the same group were merged and the distribution of domains and domain architectures per group was examined as before . A complete list with all exclusive domain architectures found in at least one member of each group is provided in Table S9 . Group 4 ( the one including the species which are most commonly found in clinical samples ) was characterized by 140 exclusive architectures . Among those , 11 architectures contained DUFs . Group 3 contained 674 unique domain architectures . The least clinically relevant groups , groups 2 and 1 , had 1 , 801 and 3 , 048 architectures , respectively . They were the ones with the highest number of exclusive architectures . When the analysis was performed at the level of the protein domains , 11 exclusive domains were found for at least one species in Group 3 , whereas only two domains were found to be exclusive of Group 4 . Table 4 shows the domains exclusively associated with the species of the Groups 3 and 4 . Among those domains , three of them can be highlighted: ‘HI0933_like’ ( PF03486 ) , ‘TIR_2’ ( PF13676 ) , and ‘Keratin_B2_2’ ( PF13885 ) . The first one is found in redox enzymes . Redox metabolism has already been proved to be a useful target for drug development in other microorganisms [49] , [50] . The ‘TIR_2’ domain belongs to a family of bacterial ‘Toll-like’ receptors . Recently , novel potential drug targets related to the ‘Toll-like’ receptor have been described in Mycobacterium tuberculosis [51] . Finally , the ‘Keratin_B2_2’ domain is rich in cysteine and it is present in proteins associated with keratin . These proteins belong to the same Pfam clan ( CL0291 ) than the AMP protein , a small protein with microbial activity . Defensins from plants , insects and mammals constitute the most prominent group within the known antifungal AMPs . The number of defensin-like antifungal AMPs of fungal origin studied is increasing and they show similar structural features to the defensins . In fact , AMPs might provide to their hosts the advantage to successfully compete with organisms that possess similar nutritional and ecological requirements . It is believed that the activity of many of these fungal AMPs is based on their interaction with proteins from the cell wall or the membrane of target fungi implicated in the signalling cascade . AMP related studies constitute a promising and fast growing field , which could trigger the identification of novel related antifungal drugs in the coming years [52] . Extended information about all domains exclusively found in different at least one species from the described groups according to their occurrence in clinical samples can be found in Table S10 . With this approach we aimed at looking for drug targets in some of the most common human pathogens . However , the classification used here has clearly some limitations . The main drawback is that it is based on two epidemiological studies performed only in Spain . Although some worldwide information was also taken into account to perform the classification , the scenario can be slightly different in other parts of the world . However , the spectrum of species isolated in the Spanish studies is comparable with other studies performed in other countries ( such as the TransNet study in the US ) . It is also important to highlight that both studies are focused mainly in deep infections and therefore the highly prevalent subcutaneous infections could not be classified according to these criteria . However , this last issue can be assumed since the focus in the development of new antifungal drugs is usually put in the type of diseases where the mortality rate is particularly high and where multi-resistant fungi are mainly present . As a third approach to identify potential targets for drugs , the list of protein domains produced by Kruger and colleagues in a previous study [21] was used to assess the ligand-binding capability for each of the fungal domains . They generated a list of 215 Pfam domains where some evidence of ligand-binding capabilities to small molecules was found . When we compared directly this list with the generated list in the previous section ( according to the occurrence of the fungi in clinical samples ) , we found that none of the exclusive domains from proteins of the Groups 3 or 4 were identified to have small-molecule binding potential . Only seven domains exclusive for organisms belonging to the less relevant group 2 appeared to have small-molecule binding potential: ‘BH3’ ( PF15285 ) , ‘Cons_hypoth698’ ( PF03601 ) , DUF2146 ( PF10220 ) , ‘Gb3_synth’ ( PF04572 ) , ‘NCD2’ ( PF04905 ) , ‘bact-PGI_C’ ( PF10432 ) and ‘TTKRSYEDQ’ ( PF10212 ) . Interestingly , these exclusive domains were present mainly in proteins of the multi-resistant species F . oxysporum and R . delemar ( closely related to Rhizopus oryzae ) . Based on the structural similarities , we decided to include in the comparison all the Pfam domains belonging to the same clans than the list of 215 domains initially studied ( extending the initial number to 1 , 193 ) . As a result of this second analysis , three domains were exclusively found in proteins exclusive of species of the Group 3: ‘CTP_transf_3’ ( PF02348 ) , ‘HI0933_like’ ( PF03486 , also highlighted in the previous section ) and ‘Uma2’ ( PF05685 ) . ‘CTP_transf_3’ is a domain found in cytidylyl-transferase membrane proteins , which are important regulatory enzymes in the synthesis of phospholipids in eukaryotic cell membranes [53] . As mentioned before , the ‘HI0933_like’ domain can be found in redox enzymes . It belongs to the ‘NADP_Rossmann’ clan whose domains typically bind nicotinamide adenine dinucleotide ( NAD+ ) [54] . Finally the ‘Uma2’ domain family , frequently found in bacteria and archaea , is believed to be part of restriction endonucleases , which are enzymes which cleave DNA and allow DNA recombination [55] . ‘CTP_transf_3’ ( unlike ‘HI0933_like’ and ‘Uma2’ ) was found in human proteins . The ‘HI0933_like domain’ was found only in one protein ( Q0CDT4 ) of A . terreus whereas ‘Uma2’ was exclusively identified in Trichophyton rubrum ( protein F2SXP0 ) . Both are putative uncharacterized poorly annotated proteins . Q0CDT4 was predicted to have a signal peptide using SignalP [56] , indicating that it could be secreted . To summarize , these three domains have potential ability to bind small molecules and domains from the same clan have been found exclusively in fungal organisms with clinical interest , so they could be further studied as potential antifungal targets . In this study , we have characterized the protein domain and domain architecture content of the available fungal proteomes ( including the phylum Microsporidia ) and we have shown how that information can be used in silico to detect potential candidate targets for antifungal drugs . Throughout the study , we have considered and taken into account information coming from both individual domains and complete protein domain architectures . In the case of the latter , independent architectures were defined as those ones in which the order or number of domains were different ( taking into account domain duplications ) . In analogous previous studies , protein domain order and domain repetitions were considered in a different way , assigning the same domain arrangement regardless of the number of consecutive copies of a single given domain [10] , [13] , [37] , [57] . That means that some of the results reported here cannot be directly compared with the ones reported previously . We decided to follow this approach since recent studies had noted the importance of domain recurrence ( repetitions of the same domain within a protein ) in the overall domain function [58] . Furthermore , domain repetition is a predominant mechanism for protein diversity and evolution [59] . For instance , in some cases it has been shown that the repetition of a domain in a multidomain protein is essential for the protein function . One good example is the glutamate receptor interacting protein ( GRIP ) consisting in 7 ‘PDZ’ domains , two of them interacting with an α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid ( AMPA ) receptor , but only in the presence of the adjacent copies [58] . Another example is the existence of the multi-modular enzymes non-ribosomal peptide synthetases ( NRPS ) and polyketide synthases ( PKS ) [60] , [61] , which are responsible of the synthesis of multiple biologically active products produced by bacteria and fungi , among other organisms . The number and exact order of these domains will determine the resulting synthesized compound/s . Another reason for the consideration of the architectures as they were , is that part of the approach developed in this project is planned to be used to improve automatic protein sequence annotation in UniProtKB [62] . UniProtKB has two sections: UniProtKB/Swiss-Prot , which is manually annotated and reviewed , and UniProtKB/TrEMBL , which is automatically annotated using different bioinformatics approaches . The general method used to annotate proteins coming from new genomes is to transfer the annotation from homologous proteins identified by sequence-based or protein domain-based approaches . It is also possible to connect the information of the domain architectures specific to a given taxonomical level with the detected promiscuous domains . This information could be used to retrieve more information on the unknown functions of certain domains . For example , knowing that the fungal promiscuous domain ‘PX’ is usually involved in targeting proteins to cell membranes , domain architectures including ‘PX’ domains might be of interest . Looking into this particular case , we identified the domain architecture ‘PXB∼PX∼DUF3818’ , present in 106 proteins of all fungal phyla . The only protein manually annotated in UniProtKB with this domain architecture is Q06839 , a peripheral membrane protein believed to be involved with cell communication process ( GO:0007154 ) in S . cerevisiae ( strain ATCC 204508/S288c ) . From this observation two applications of the work presented here could take place: on one hand , proteins containing the same domain composition could be similarly annotated . On the other hand , the importance of these particular proteins and the corresponding domains ( like DUF3818 ) could be further investigated given their membrane localization and the important cell communication process in which these proteins are believed to play a role . As a result of this comprehensive study , we provide access to the full list of ‘core’ and exclusive domains and domain architectures at different levels of the taxonomic classification and also identify the promiscuous domains . This information can be a very valuable resource for researchers interested in comparative studies between different fungal organisms . Here , we have only highlighted some examples of how this information could be used , but it is clear for us that more focused studies could be performed on particular groups of organisms , using all the generated information here . This information could also be combined with genome features such as gene clusters ( very frequently found in fungi , e . g . for genes involved in secondary metabolism ) or synteny . However , in this study we decided to focus on the possible application of this information in the detection of antifungal targets . We then followed three different approaches . First of all , we identified those protein domains and domain architectures that were present in fungi but not found in the human proteome . Secondly , using the promiscuous domains , we identified three pathways whose components could be targeted . Last , we created five groups of organisms depending on their occurrence in clinical samples and then inferred small-molecule protein domain binding information obtained in a recent study involving small molecules stored in the ChEMBL database . The results coming from these three approaches constitute just a first step and should be taken with caution , since they have different inherent limitations . It is also expected that these approaches will provide new information when new data ( e . g . pathway related information ) is made available in the public domain . Analogous studies where the interaction between protein domains and small molecules is assessed , are becoming more popular in the last few years . For instance , recently , using data from Protein Data Bank ( PDB ) structures , more than thirteen thousand physical interactions between small molecules and protein domains were identified [63] . The authors found that the capability of the protein domains to bind particular small molecules did not depend only on the protein sequence and protein structure . For example , protein domains distributed in different domain families and biological pathways were able to bind the same or similar small molecules . Some very heterogeneous domains were able to bind hundreds of small molecules . In addition , they found that 12% of the small molecules were able to bind multiple domains . It is also important to highlight that it is envisioned that small-molecule and protein domain binding information could be used in the future for finding the optimal combination of drugs in treatments . Throughout this study we have used domain information coming from Pfam-A domains , so all the conclusions are limited to that context . One fact to consider is that protein sequence coverage in Pfam for fungi is lower than for other organisms , especially for species like M . laricis-populina ( coverage of 36 . 5% ) and P . graminis sp . tritici ( coverage of 39 . 5% ) . In the future , this study could be extended by using Pfam-B domains and protein domain information coming from other databases , for instance from other protein domain/family resources that are also part of the InterPro consortium [64] . However , integrating protein domain information obtained using different computational methods or coming from different resources makes direct comparisons troublesome . In addition , another way to continue this study would be to use the sequences that currently lack protein domain annotation as an input to perform local protein sequence alignments with different methods and thresholds , and then generate the corresponding HMM-based profiles , in order to increase the detectability of homologous proteins . Overall , this manuscript provides a comprehensive analysis of protein domain and domain architectures in the available fungal proteomes and shows three approaches that can be used as a first step in the detection of new antifungal targets . These approaches could also be used for organisms with clinical interest other than fungi e . g . bacteria . Therefore , analogous analyses could be performed for different groups of pathogenic bacteria using as a starting point the scripts provided ( available at https://github . com/alexbarrera/fungidomDB ) . We hope that some of these strategies can be refined and applied in the practice , followed by in vivo validation .
The proteomes used in this study were obtained from UniProtKB ( http://www . uniprot . org/ ) [62] , release 2013_01 . Only fungal complete and reference proteomes were included . In total , 137 fungal proteomes belonging to 111 species were analysed . All complete proteomes available for species with more than one strain ( 12 species ) were also added to the collection ( Table S1 ) . The human “reference proteome” ( including protein isoforms ) was also retrieved for comparison purposes ( 64 , 677 proteins ) . “Reference proteomes” in UniProtKB are defined as complete proteomes targeted for manual annotation by curators , to emphasize the reliability of their annotations . Domain information was obtained from Pfam [7] release 27 . 0 ( June 2012 ) . The Pfam database is divided in a collection of manually curated families known as Pfam-A and a set of automatically generated families named Pfam-B . Pfam-A domains and Pfam clans were considered . The mapping of Pfam domains to proteins was obtained via FTP ( ftp://ftp . sanger . ac . uk/pub/databases/Pfam/releases/Pfam27 . 0/ ) . Only 425 protein sequences ( 0 . 035% ) from UniProtKB release 2013_01 were not present in Pfam 27 . 0 . For those proteins , Pfam domain information was obtained directly from UniProtKB using a tailored extended version of UniProtJAPI , the Java Application Programming Interface ( API ) to access UniProtKB [65] . Additional functional annotation for Pfam domains was retrieved using Gene Ontology ( GO ) terms ( http://www . geneontology . org ) [66] . Information available in the file ‘Pfam2GO . txt’ ( version 2013/07/24 ) was used to map GO terms to the Pfam domains . These expert manual annotations are originally assigned by curators in the InterPro team based on the function of particular domains rather than the function of domain families [67] . A local MySQL database was developed to store the protein sequence and domain information . The analysis of domain and architectures was performed using R ( http://www . r-project . org ) version 2 . 15 . 3 and scripts developed in Python . The charts and diagrams were produced using the following R packages: ‘ggplot2’ [68] for the histograms and ‘treemap’ [69] to generate the distribution of domains and architectures per genus . Finally , standard functions were used to plot the pie charts and Venn diagrams . All the scripts used in this study and related documentation are available at https://github . com/alexbarrera/fungidomDB . The nomenclature of the fungal organisms followed in this study was the same one used in UniProtKB ( release 01_2013 ) . The taxonomical classification used was the one provided by the National Centre for Biotechnology Information ( NCBI ) [70] . One limitation of the NCBI taxonomy is that not every taxonomical category ( e . g . phylum , subphylum , genus , etc ) is defined for every species . In that case , the strategy followed was to propagate the parent taxonomical level in those taxonomical categories that were lacking a specific one . Of note , P . brasiliensis Pb01 has been recently reclassified as P . lutzii [27] , [28] . However , we kept the original nomenclature used in UniProtKB to facilitate the traceability and reproducibility of the results . A method to measure the weighted bigram frequency ( WBF ) , introduced by Basu and colleagues [37] , was used to perform the analysis . This method is an adaptation of the Kullback-Leibler information gain formula . To reduce the impact of the over-abundance of domains , this algorithm divides the number of distinct domain combinations by the frequency in which the domain is found in the proteome . The rationale behind is to minimize the impact of highly abundant domains that would otherwise account for the majority of the distinct domain combinations . Information in Table 2 was computed applying this metric for each domain in each organism . To consider one domain as promiscuous , its WBF value had to be greater than the WBF value of a domain appearing only one time in the given proteome in combination with another domain . Promiscuous domains were ranked according to their ‘promiscuity’ . Metabolic pathway information was retrieved from the public resource UniPathway ( http://www . unipathway . org/ ) [71] . The mapping between the UniProtKB entries and the UniPathway pathways was obtained from UniProtKB ( release 07_2013 ) . UniPathway was chosen over other pathway resources as KEGG ( http://www . genome . jp/kegg ) or MetaCyc ( http://metacyc . org ) due to the comprehensive mapping to UniProtKB entries and the hierarchical representation of this information . In addition , entities from UniPathway are cross-linked to other major pathway resources such as KEGG , MetaCyc and ‘The SEED’ ( http://www . theseed . org ) . Based on the frequency of appearance in clinical samples ( according to two recent epidemiological prospective studies carried out in Spain [72] , [73] ) , the fungi were classified in five groups using the following criteria ( Table S1 ) : A list of in silico predicted protein domains that can bind small molecules was collected from a recent study [21] . The authors explored the binding capability of ligands and small molecules included in the ChEMBL database [22] to Pfam domains . The resulting list of domains ( available in http://www . ebi . ac . uk/~fkrueger/mapChEMBLPfam/ ) was used to cross-check this capability for the fungal domains included in this study . | Some fungi have become pathogenic to plants and in a lesser extent to animals . Under certain conditions their presence in the human body can prove a threat for human health , especially for immunocompromised patients . Yet , some fungi can also infect healthy individuals . The low sensitivity of the antifungal drugs available together with the clinically observed resistance of some fungi raises the demand for new alternative treatments . Proteins are biological molecules which perform essential functions within the living organisms . Many of those functions are attributed to the varying folded structure of each protein . These configurations are composed of functional units -also called domains- each one independently responsible for a fraction of the overall biological function . Understanding how the different block combinations are distributed across members of the same or similar families of organisms is important . For instance , exclusive domain combinations can hold particular acquired functions . Blocks displaying a high mobility can play major roles for the organism's survival . The biological goal of this study was to analyse the functional implications of protein domains and domain combinations in the available fungal proteomes . This information can be used to highlight proteins and pathways that could be potentially used as drug targets . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
]
| [
"biotechnology",
"biomacromolecule-ligand",
"interactions",
"sequencing",
"techniques",
"medicine",
"and",
"health",
"sciences",
"functional",
"genomics",
"small",
"molecules",
"molecular",
"biology",
"techniques",
"pharmacology",
"sequence",
"analysis",
"genome",
"complexity",
"proteins",
"molecular",
"biology",
"drug",
"discovery",
"biochemistry",
"drug",
"research",
"and",
"development",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"computational",
"biology"
]
| 2014 | Analysis of the Protein Domain and Domain Architecture Content in Fungi and Its Application in the Search of New Antifungal Targets |
DJ-1 is one of the causative genes for early onset familiar Parkinson’s disease ( PD ) and is also considered to influence the pathogenesis of sporadic PD . DJ-1 has various physiological functions which converge on controlling intracellular reactive oxygen species ( ROS ) levels . In RNA-sequencing analyses searching for novel anti-oxidant genes downstream of DJ-1 , a gene encoding NADP+-dependent isocitrate dehydrogenase ( IDH ) , which converts isocitrate into α-ketoglutarate , was detected . Loss of IDH induced hyper-sensitivity to oxidative stress accompanying age-dependent mitochondrial defects and dopaminergic ( DA ) neuron degeneration in Drosophila , indicating its critical roles in maintaining mitochondrial integrity and DA neuron survival . Further genetic analysis suggested that DJ-1 controls IDH gene expression through nuclear factor-E2-related factor2 ( Nrf2 ) . Using Drosophila and mammalian DA models , we found that IDH suppresses intracellular and mitochondrial ROS level and subsequent DA neuron loss downstream of DJ-1 . Consistently , trimethyl isocitrate ( TIC ) , a cell permeable isocitrate , protected mammalian DJ-1 null DA cells from oxidative stress in an IDH-dependent manner . These results suggest that isocitrate and its derivatives are novel treatments for PD associated with DJ-1 dysfunction .
Parkinson’s disease ( PD ) is the second most common neurodegenerative disease and is characterized by typical movement disorders and selective loss of dopaminergic ( DA ) neurons in the substantia nigra pars compacta ( SNpc ) [1] . Accumulated evidence has firmly linked the death of these neurons to oxidative stress , the state of imbalance between generation and elimination of reactive oxygen species ( ROS ) [2] . Postmortem brain analysis showed that markers of oxidative damage to lipids , proteins , and nucleic acids are substantially elevated in the SNpc of PD patients [2] . High levels of somatic mitochondrial DNA ( mtDNA ) deletion are also found in the SNpc neurons from PD patients [3] , suggesting a vicious cycle of oxidative damage to mtDNA and other mitochondrial components , thus increasing ROS production in the course of DA neuron degeneration . The link between oxidative stress and DA neuronal loss is further supported by modeling parkinsonism in various animals using oxidative stress-inducing agents , such as 1-methyl-4-phenyl-1 , 2 , 3 , 6-tetrahydropyridine ( MPTP ) , rotenone , paraquat , and 6-hydroxydopamine ( 6-OHDA ) [4–8] . In addition to PD , other neurodegenerative diseases including Alzheimer’s disease ( AD ) , Huntington’s disease ( HD ) , and amyotrophic lateral sclerosis ( ALS ) are also associated with oxidative stress , further strengthening the correlation between oxidative stress and neurodegeneration [9] . However , the molecular mechanisms resulting in DA neuron degeneration under oxidative stress have not been fully elucidated . Although PD mainly occurs in a sporadic manner , it could also occur by monogenic mutations [10] . Because familial forms of PD are often clinically and pathologically indistinguishable from sporadic ones , they are likely to have common pathogenic mechanisms [11] . Moreover , recent genome-wide association studies ( GWAS ) have revealed variations in several familial PD genes as significant risk factors for the development of sporadic PD [12] . Therefore , investigating how PD gene mutations cause familial PD could potentially reveal the molecular pathogenesis of sporadic PD . Among PD-linked genes , DJ-1 is most closely associated with oxidative stress [2] . DJ-1 was first identified as an oncogene that transforms mouse NIH3T3 cells in cooperation with activated Ras [13] . Later , Bonifati et al . found that DJ-1 is associated with an autosomal recessive early onset type of familial PD [14] . DJ-1-deficient animal models showed hypersensitivity to oxidative stress [15–18] , and further cell biological studies revealed that DJ-1 is a multifunctional protein that participates in transcription regulation , anti-apoptotic signaling , protein stabilization and degradation , and mitochondrial regulation to respond to oxidative stress [19] . DJ-1 is sequentially oxidized on its cysteine residues , and its activity and subcellular localization are regulated by its oxidative status [20–23] . Excessive oxidation of DJ-1 inactivates it , and this oxidized form is observed in the brains of patients with sporadic PD and AD [24 , 25] , suggesting that DJ-1 participates in the pathogenesis of sporadic PD as well as familial PD . In Drosophila , there are two homologues of mammalian DJ-1; DJ-1α and β . DJ-1α is predominantly expressed in the testes , whereas DJ-1β is expressed throughout the whole body [16 , 18 , 26] , similar to the expression pattern of mammalian DJ-1 [13] . DJ-1β loss-of-function mutants show locomotive dysfunction and loss of DA neurons , resembling the phenotypes seen in PD patients [18 , 27] . In this study , we found that DJ-1 is critical for maintaining transcription of NADP+-dependent isocitrate dehydrogenase ( IDH ) under oxidative stress induced by pesticides like rotenone that have been associated with onset of PD in recent epidemiologic studies [28] . IDH catalyzes decarboxylation of isocitrate into α-ketoglutarate and CO2 , and also produces NADPH , which provides a reducing power to antioxidant processes scavenging ROS [29] . Indeed , our Drosophila IDH mutants showed decreased NADPH levels with increased ROS production and hyper-sensitivity to oxidative stress . Moreover , loss of IDH induced age-dependent mitochondrial defects and DA neuron degeneration , very similar to the phenotypes of Drosophila PD models [30] . Consistently , overexpression of IDH in DJ-1 mutants successfully enhanced their survival rates and ameliorated DA neuron loss under oxidative stress . Further genetic analysis revealed that DJ-1 maintains IDH expression by regulating the Kelch-like ECH-associating protein 1 ( Keap1 ) -nuclear factor-E2-related factor2 ( Nrf2 ) pathway . In addition , trimethyl isocitrate ( TIC ) , a cell permeable form of isocitrate , markedly restored oxidative stress-induced decrease of NADPH level and inhibited subsequent cell death in mammalian DA cells with DJ-1 deficiency . These results consistently support that the activity of NADP+-dependent IDH is critical in protecting neurons from oxidative stress and DJ-1 mutation .
To find out a new molecular mechanism in which DJ-1 protects cells from oxidative stress , we treated rotenone , a well-known ROS inducer associated with PD [28] , to wild type and DJ-1β-deficient flies , and investigated gene expression in both of them through RNA-sequencing ( RNA-seq ) analysis . Based on the role of mitochondria as a center for generating and controlling ROS [9] , we hypothesized that a ROS controlling protein located in mitochondria would act downstream of DJ-1 . We looked over the RNA-seq result and found that oxidation-reduction process gene ontology was the most changed in biological process terms ( S1 Table ) and oxidoreductase activity gene ontology was the most changed in molecular function terms ( S2 Table ) between wild type and DJ-1β-deficient flies , consistent with the role of DJ-1 in oxidative stress responses . We further looked into the gene list falling into two groups: oxidation-reduction process ( GO: 0016491 ) and mitochondrion ( GO: 0005739 ) . As a result , 173 genes in oxidation-reduction process ontology and 237 genes in mitochondrion ontology were found to be different between wild type and DJ-1ß null mutants in RNA-seq analysis . Interestingly , 34 genes were included in both ontologies ( S1A Fig ) , and most of the genes diminished their mRNA expression in DJ-1β null mutants ( S1A and S1B Fig ) . Among them , 3 genes showed statistically significant difference in false discovery rate ( FDR ) < 0 . 05 and satisfied fold change > 1 . 5 at the same time . As we expected , IDH , which encodes a protein regulating intracellular ROS level by producing NADPH , was one of the 3 genes ( Fig 1A and Table 1 ) . In quantitative RT-PCR , IDH gene expression was decreased in DJ-1β null mutants compared to wild type controls under oxidative stress , confirming the RNA-seq data ( Fig 1B ) . The reduction in IDH expression in DJ-1β null mutants was observed in heads , thoraces and abdomens , indicating that it may be not tissue-specific ( S1C–S1E Fig ) . In mammalian organisms , IDH1 and IDH2 are located in cytosol and mitochondria , respectively , and they are expressed from independent genes [29] . However , in Drosophila , a cytosolic isoform ( IDHc ) and mitochondrial isoforms ( IDHm1 and IDHm2 ) are expressed from the single gene IDH , although Drosophila IDHs are highly homologous to human counterparts ( S2A and S2B Fig ) . IDHP is a Drosophila mutant line containing a P-element in an exon that is shared by all isoforms of IDH ( S2C Fig ) . IDHP flies successfully developed into adults ( Fig 1C ) , but they showed decreased IDH mRNA expression , IDH enzyme activity , and NADPH/NADP+ ratio compared to wild type controls ( Fig 1D–1F ) . All these phenotypes were rescued in the revertant ( RV ) generated by precise P-element excision ( Fig 1D–1F ) . These results demonstrated that the P-element insertion successfully inhibits gene expression , activity , and function of IDH in the IDH mutants . In addition , the inserted P-element slightly increased expression of CG17352 , a gene located next to IDH ( S2C and S2D Fig ) . However , compared to control flies , CG17352 transgenic flies showed no difference in survival rates under oxidative stress ( S2E Fig ) , ruling out the possibility that this small increase of CG17352 expression affects the oxidative stress-related phenotypes of IDHP mutants that we examined in following experiments . To understand physiological functions of IDH , we checked the lifespan of IDHP flies that exhibited notably shorter life spans than WT and RV controls ( S2F Fig ) . Furthermore , ROS markers such as dihydroethidium ( DHE ) and 5- ( and-6 ) -chloromethyl-2′ , 7′-dichlorodihydrofluorescein diacetate ( CM-H2DCFDA ) demonstrated that in vivo ROS accumulated highly in IDH mutants ( Fig 1G and 1H ) . This finding was further confirmed by measuring the level of hsp22 mRNA expression which increases according to mitochondrial oxidative stress ( S2G Fig ) [31] . Consistently , the survival rate of IDHP flies radically declined under rotenone treatments showing that IDH is indispensable for sustaining resistance to oxidative stress ( Fig 1I ) . To further investigate the effect of the in vivo ROS accumulation , the phenotypes of 3- and 30-day-old IDHP adult flies were compared to those of RV control flies without oxidative insults . While ATP level in the muscles , which indicates the function of mitochondria , and mtDNA content , which shows the amount of mitochondria , did not present any significant changes in 3-day-old flies , those of 30-day-old IDHP mutants were markedly dropped compared to RV control flies ( S3A–S3D Fig ) . Based on the fact that oxidative stress and mitochondrial defects are vital factors in onset and deterioration of a variety of human diseases , particularly neurodegenerative diseases including PD [2] , the phenotypes related to PD in IDH mutants were analyzed . The results showed that climbing ability and the number of DA neurons of 30-day-old IDH mutants substantially decreased ( ~30% for climbing ability , ~10% for DA neuron number ) compared to those of RV control flies ( Fig 1J and 1K , and S3E and S3F Fig ) . Specifically , among the four major DA neuron clusters [dorsolateral clusters 1 ( DL1 ) , dorsomedial clusters ( DM ) , posteriomedial clusters ( PM ) , and dorsolateral clusters 1 ( DL2 ) ] [32] , only the DA neurons in DL1 and DM clusters were degenerated ( Fig 1J and 1K ) , as previously shown in PINK1 and Parkin mutant flies [32] . These results showed that DA neuronal degeneration induced by mutating IDH gene can be clearly distinguished from non-specific neuronal degeneration caused by simply increasing ROS . Drosophila phenotypes related to PD symptoms , such as loss of climbing ability and DA neuronal degeneration , were also observed in IDHP flies . Consistent with these results , when IDHm1 , IDHm2 , and IDHc isoforms were overexpressed in DJ-1β mutant flies , survival rates of DJ-1β mutant flies increased dramatically in a rotenone-induced oxidative stress condition ( Fig 2A ) . IDH overexpression rescued the reduced climbing ability caused by DJ-1β mutation under rotenone treatment ( Fig 2B ) , and suppressed the DA neuronal degeneration caused by rotenone or H2O2 treatment in DJ-1β mutants ( Fig 2C , 2D and S4 Fig ) . Moreover , IDHP mutation decreased survival rates and the number of DA neurons under oxidative stress , but the genetic combining of IDHP mutation did not have an effect on DJ-1β mutants ( S5 Fig ) . From these results , we concluded that IDH is an important antioxidant enzyme that protects cells downstream of DJ-1 under oxidative stress . Especially , expressing IDHm1 and IDHm2 more effectively restored the phenotypes of DJ-1β mutants induced by oxidative stress ( Fig 2A–2D and S4 Fig ) , indicating that mitochondrial IDH plays an important role in DJ-1-regulated cell protection against oxidative stress . Consistent with this idea , in DJ-1β deficient flies stressed with rotenone , the mRNA expression of mitochondrial IDH isoforms ( IDHm1 and IDHm2 ) was severely decreased , but that of cytosolic IDH isoform ( IDHc ) showed no significant change ( Fig 2E–2G ) . In addition , when we overexpressed IDH genes in PINK1 mutants which showed reduced levels of ATP and mtDNA in muscles , loss of climbing ability , and DA neuronal degeneration in the absence of oxidative stress [32] , none of these phenotypes were rescued ( S6 Fig ) . These results suggested that IDH is specifically related to DJ-1 in PD pathology . The transcription factor Nrf2 induces expression of various antioxidant proteins in response to oxidative stress [33] . Under normal conditions , Nrf2 is degraded by Keap1 , an E3 ligase . Upon exposure to oxidative stress , Nrf2 is stabilized and translocated to the nucleus [34–37] . Interestingly , DJ-1 stabilizes Nrf2 by preventing its association with Keap1 and thereby preventing the subsequent degradation [38] . Previous data showed that Nrf2 also controls IDH mRNA expression [39] . Thus , we suspected Nrf2 to be a possible transcription factor that mediates IDH mRNA expression regulated by DJ-1 . When we checked mRNA levels of each IDH isoform , the expression levels of IDHm1 and IDHm2 were rescued by Keap1 mutation in DJ-1β null flies under oxidative stress ( Fig 3A and 3B ) . In contrast , Keap1 mutation failed to change IDHc mRNA expression in the stressed DJ-1β mutants , suggesting that DJ-1 specifically induces mitochondrial IDH isoforms via Keap1-Nrf2 pathways ( Fig 3C ) . In addition , Keap1 mutation also increased survival rates of DJ-1β mutants on rotenone- or H2O2-containing media ( Fig 3D and 3E ) . Consistently , Cap’n’Collar C ( CncC ) , a fly orthologue of mammalian Nrf2 [40] , could induce IDHm1 mRNA expression ( Fig 3F ) , but failed to change IDHc mRNA expression ( Fig 3G ) . When Keap1 was co-overexpressed , IDHm1 mRNA expression decreased , and increased again with the addition of DJ-1β overexpression ( Fig 3F ) . Moreover , when CncC was overexpressed in DJ-1β mutants , DA neuronal death induced by rotenone or H2O2 was suppressed ( Fig 3H–3K ) . These results confirmed that DJ-1 accelerates mRNA expression of mitochondrial IDH isoforms through the Keap1-Nrf2 pathway to protect the cells from oxidative stress . In the process of investigating the mechanism of Nrf2-induced IDH expression , we found that a putative Nrf2-binding element , also known as Antioxidant Response Element ( ARE ) , ( TGACGGGGC ) [33] is located on the promoter region of the IDH gene ( S2C Fig ) , and cloned this IDH promoter region in a luciferase reporter plasmid . We co-transfected CncC cDNA and this ARE reporter plasmid in Drosophila S2 cell line , and found that the promoter activity increased with CncC expression and decreased with ARE mutation ( Fig 3L ) . Therefore , we concluded that DJ-1 controls IDH expression through Nrf2 and ARE in an evolutionarily conserved manner . In the above experiments , we showed that IDH complements DJ-1β mutation in Drosophila . Using mouse DA cell line SN4741 [41] and CRISPR/Cas9-mediated genome editing , we generated DJ-1 null DA cells to further investigate the relationship between DJ-1β and IDH in mammalian systems ( S7A and S7B Fig ) . When we examined IDH1 and IDH2 mRNA expression in SN4741 cells under oxidative stress , there was no significant difference in IDH1 mRNA expression between wild type and DJ-1 null SN4741 cells with or without H2O2 treatment ( Fig 4A ) . In contrast , IDH2 mRNA expression in wild type SN4741 was elevated with H2O2 treatment , but it was almost nullified in DJ-1 null SN4741 cells ( Fig 4B ) . Moreover , Keap1 knockdown completely rescued IDH2 mRNA expression in H2O2-treated DJ-1 null cells , showing that Nrf2 also mediates between mitochondrial IDH and DJ-1 in mammalian cells ( Fig 4B and S7C Fig ) . In MTT assays , cell viability increased when IDH1 and IDH2 were expressed in DJ-1 null SN4741 cells under oxidative stress ( Fig 4C ) . We performed both annexin V and propidium iodide ( PI ) staining to observe cell death , but only PI staining was positive in DJ-1 null SN4741 cells in H2O2-containing media , indicating necrotic cell death ( Fig 4D ) . Overexpression of IDH1 or IDH2 rescued this necrotic cell death upon H2O2 treatment in DJ-1 null SN4741 cells ( Fig 4D and 4E ) . These results in mammalian DA cells were highly similar to the DJ-1 mutant phenotype recovery by IDH overexpression in Drosophila except that IDHc overexpression failed to protect DA neurons in some DA neuron clusters of DJ-1β mutants ( Fig 2 and S4 Fig ) . Moreover , CM-H2DCFDA showed that intracellular ROS level dramatically increased in DJ-1 null SN4741 cells when H2O2 was treated , but IDH1 and IDH2 overexpression significantly decreased the ROS level ( Fig 4F and 4G ) . Furthermore , when mitochondrial ROS level was measured by MitoSOX , mitochondrial ROS drastically increased by oxidative stress in DJ-1 null SN4741 . However , the mitochondrial ROS level also decreased with overexpression of IDH1 and IDH2 in DJ-1 null SN4741 ( Fig 4H and 4I ) . These results suggested that the relationship between DJ-1 and IDH genes is evolutionarily well conserved and is important in controlling intracellular ROS level , especially mitochondrial ROS level . In addition , a mitochondrial-specific ROS scavenger , MitoTEMPO , successfully inhibited oxidative stress-induced cell death in DJ-1 null SN4741 cells ( Fig 4J ) , further supporting the roles of DJ-1 in controlling mitochondrial ROS . To further investigate the antioxidant effect of IDH , we searched for a compound that increases IDH activity , but found no such existing activator . To increase NADPH production by IDH , we tried to deliver isocitrate , a substrate of IDH , into the cells . We decided to attach 3 methyl groups to isocitrate to make it permeable to the cells . When treated with H2O2 , NADPH/NADP+ ratio in DJ-1 null SN4741 cells decreased dramatically compared to wild type cells ( Fig 5A ) . This result showed that DJ-1 is indispensable for maintaining NADPH/NADP+ ratio under oxidative stress . Surprisingly , trimethyl isocitrate ( TIC ) treatment substantially restored NADPH/NADP+ ratio ( ~40–50% ) and cell viability ( ~70–80% ) dose-dependently in DJ-1 null SN4741 cells under oxidative stress ( Fig 5A and 5B ) , suggesting that TIC successfully crossed through the plasma membrane and increased NADPH production under oxidative stress . In an additional experiment using annexin V and PI , PI uptake was decreased when cells were pre-treated with TIC ( Fig 5C and 5D ) . These results indicate that TIC inhibited necrotic cell death caused by H2O2 treatment . When SN4741 cells were stained with ROS indicators , CM-H2DCFDA and MitoSOX , intracellular and mitochondrial ROS levels were increased in H2O2-treated DJ-1 null SN4741 cells . However , when cells were pre-treated with TIC , ROS levels immensely decreased in H2O2-treated DJ-1 null SN4741 cells ( Fig 5E–5H ) . These results implied that TIC raises NADPH/NADP+ ratio which detoxifies ROS and ultimately protects DJ-1 null DA cells from oxidative stress . To see whether IDH is indeed responsible for the cell protecting effect of TIC , we suppressed IDH expression in DJ-1 null SN4741 cells with siRNA technology . We observed that when gene expression of IDH1 and IDH2 was blocked simultaneously , TIC had no effect in protecting cells against oxidative stress , confirming that TIC functions through IDH ( Fig 5I ) .
Recently , it has been reported that oxidative stress and mitochondrial maintenance are highly correlated with PD [2] . In this study , we performed RNA-seq to find a novel gene which provides resistance against oxidative stress at the downstream of DJ-1 . We assumed that there are important target genes related to the mitochondrion , which is the main organelle of ROS generation . Through RNA-seq analysis ( Fig 1A and S1A and S1B Fig ) , we found that IDH was a gene included in mitochondrion ontology downstream of DJ-1 . Loss-of-function flies for IDH were generated , whose in vivo ROS levels were increased ( Fig 1G–1H ) . These IDH mutants displayed reduced survival rates against oxidative stress and decreased DA neurons and climbing ability ( Fig 1I–1K , and S3E and S3F Fig ) . IDH mutant flies showed decreased ATP generation and mtDNA contents ( S3A–S3D Fig ) that were also observed in IDH2 knockout mouse [42] . These phenotypes are very similar with the PD-related phenotypes of PINK1 , Parkin and DJ-1β mutant flies [18 , 27 , 32 , 43 , 44] , supporting that IDH is highly implicated in PD pathology . Consistently , under MPTP treatments , IDH2 KO mouse showed more severe PD-related phenotypes than those of wild type controls [45] . More interestingly , IDH overexpression in DJ-1β null mutants rescued the PD-related phenotypes of DJ-1β mutants ( Fig 2 ) , but not those of PINK1 null flies ( S6 Fig ) . This specific rescue of the phenotypes in DJ-1β mutants implies that IDH is a unique downstream regulator of DJ-1-dependent and/or oxidative stress-induced PD pathologies . In searching for the molecular link between DJ-1 and IDH , overexpression of Nrf2 , a well-known transcription factor that responds to oxidative stress [33] , suppressed the loss of DA neurons in DJ-1β mutants under oxidative stress ( Fig 3H–3K ) . Further genetic analysis demonstrated that Keap1 loss-of-function mutation restores decreased IDH expression and survival rates of DJ-1β mutants under oxidative stresses ( Fig 3A–3E ) , supporting Nrf2 as the molecular mediator that links DJ-1 and IDH under oxidative stress . Interestingly , only the expression of mitochondrial IDH isoforms was suppressed by DJ-1 mutation in Drosophila and mouse SN4741 DA cells under oxidative stress ( Figs 2E–2G , 4A and 4B ) . This specific regulation of mitochondrial IDHs is mediated by the Keap1-Nrf2 pathway in Drosophila ( Fig 3A–3C , 3F and 3G ) and mammalian system ( Fig 4B ) . When these mitochondrial IDHs were overexpressed , oxidative stress-induced DJ-1 null defects , including DA neuron loss in Drosophila brain , decreased Drosophila survival rates , and increased intracellular and mitochondrial ROS levels in SN4741 DA cells , were successfully rescued ( Figs 2 and 4 ) . These data established the role of the DJ-1-Nrf2-IDH pathway in DA neuronal protection against oxidative stress and implied that DJ-1 effectively eliminates intracellular ROS by reducing ROS generation in mitochondria through mitochondrial IDHs . Although most data converge on this conclusion , there are some results that may raise further questions . In Drosophila and SN4741 cells , cytosolic IDHs can also ameliorate the DJ-1 null defects mentioned above , including increased mitochondrial ROS levels . Is regulating mitochondrial ROS levels indeed important to protect cells ? Shin et al . reported that reducing cytosolic ROS level decreases mitochondrial ROS levels [46] . Consistently , in our experiment , overexpression of IDH1 , a cytosolic mammalian IDH , substantially reduced mitochondrial ROS accompanying decreased cellular ROS in SN4741 cells ( Fig 4I ) . However , IDH2 , the mitochondrial counterpart , reduced more mitochondrial ROS than IDH1 in SN4741 cells ( Fig 4I ) . In Drosophila brains , IDHc failed to protect DA neurons in some DA neuron clusters ( Fig 2C , 2D and S4 Fig ) . These results indicated that indirect elimination of mitochondrial ROS by cytosolic IDHs is not sufficient to protect cells compared to direct elimination via mitochondrial IDHs , especially in DA neurons that may have more complex and varied stresses compared to cultured cells . Furthermore , mitochondria-specific antioxidant MitoTEMPO strongly inhibited cell death induced by oxidative stress in DJ-1 null SN4741 cells , further confirming the importance of mitochondrial ROS reduction in DJ-1-mediated anti-oxidative stress responses ( Fig 4J ) . In addition , although all Drosophila IDHs are encoded from the same gene locus , only the expression of mitochondrial isoforms is regulated by DJ-1 and CncC ( Fig 3 ) . In DNA sequence analyses , we found a putative CpG island between the transcription start sites of IDHc and IDHm1 and 2 , raising the possibility that DNA methylation inhibits CncC to induce IDHc expression ( S2C Fig ) . However , we could not find the experimental evidence that DNA methylation is involved in the mitochondrial IDH-specific expression , so the molecular mechanism of this expression will be a future topic . After confirming IDH as a downstream target of DJ-1 , we investigated recent issues on IDH research based on our findings . It has been reported that IDH1 and IDH2 neomorphic mutations are prevalent in various cancers [29] . These mutant IDH enzymes convert α-ketoglutarate into D-2-hydroxyglutarate , which promotes tumorigenesis [47] . Therefore , we tested whether these neoenzymes can affect DJ-1 mutant phenotypes . Overexpression of Drosophila IDHm1 R166K and R134Q mutants , which correspond to the cancer-associated human IDH2 R172K and R140Q mutants , respectively [29] , failed to increase IDH activity in flies and restore the decreased survival rates and DA neuron numbers of DJ-1β mutants ( S8 Fig ) . This is consistent with the reports that the cancer-associated IDH mutants do not produce , but consume NADPH [47] . In contrast , wild type IDHm1 consistently increased in vivo IDH activity and rescued the PD-related phenotypes ( S8 Fig ) . Thus , these results implicate that the activity of wild type IDH , not the cancer-related one , is linked to DJ-1-associated PD , suggesting that the drugs developed to target cancer-related IDH mutant enzymes are not appropriate to treat DJ-1-associated PD . To overcome this limitation , we hypothesized that excess concentration of cell-permeable isocitrate , the substrate of IDH , would help treat DJ-1-associated PD by raising NADPH/NADP+ ratio to increase the reducing power in the cell . As expected , TIC substantially elevated NAPDH/NADP+ ratio and strongly reduced intracellular and mitochondrial ROS levels in H2O2-treated SN4741 cells ( Fig 5A and 5E–5H ) . TIC also increased survival rate and lowered necrotic cell death against oxidative stress ( Fig 5B–5D ) . IDH expression knockdown inhibited this TIC-mediated cell protection , supporting the idea that TIC protects cells via IDH ( Fig 5I ) . Since TIC treatment significantly increased cell viability in DJ-1 null SN4741 cells , we expected a similar degree of increase in NADPH/NADP+ ratio . However , the observed ratio was less than expected , indicating that NADPH is being rapidly used to resist oxidative stress ( Fig 5A ) . Overall , these results confirmed the protective role of IDH against oxidative stress , and also suggested cell-permeable isocitrates as putative drug candidates for the treatment of DJ-1 deficiency-associated human pathology including PD ( Fig 6 ) .
da-GAL4 , hs-GAL4 , and elav-GAL4 strains were obtained from the Bloomington Stock Center . IDHP mutants ( G9298 ) were obtained from KAIST-GenExel Drosophila library and backcrossed to w1118 controls for 6 generations to remove genetic background effects . The insertion site of the P-element in IDHP is located at +301 of IDHm1 ORF , +244 of IDHm2 ORF , and +205 of IDHc ORF . A revertant ( IDHRV ) was generated by precise excision of the P-element in IDHP after backcrossing . In DNA sequencing analysis , IDHRV showed a precise excision of the P-element with no insertion or deletion of nucleotides . IDHm1 , IDHm2 , IDHc , and CG17352 cDNAs were sub-cloned from GH01524 , RE70927 , AT04910 and GH02239 BDGP cDNA clones , respectively . IDHm1 R134Q and R166K mutant cDNAs were generated by QuikChange™ site-directed mutagenesis kit ( Agilent Technologies ) using following primer pairs: IDHm1 R134Q F ( gcc caa cgg tac cat cca aaa cat ctt ggg agg aac ) , R134Q R ( gtt cct ccc aag atg ttt tgg atg gta ccg ttg ggc ) , R166K F ( gaa gcc tat tgt gat cgg taa aca tgc cca cgc cga tca gt ) and R166K R ( act gat cgg cgt ggg cat gtt tac cga tca caa tag gct tc ) . The IDH cDNAs were inserted into the pUAST vector with C-terminal HA-tag and microinjected into w1118 embryos . The CG17352 cDNA was inserted into the pACU2 vector and microinjected into y1 w1118; PBac{y+-attP-3B}VK00001 embryos . PINK1B9 , DJ-1βex54 and UAS-DJ-1β flies were generated as previously described [18 , 32] . The tyrosine hydroxylase ( TH ) -GAL4 fly was a gift from Dr . S . Birman . The Keap1EY5 , UAS-Keap1 and UAS-CncC lines were provided by Dr . D . Bohmann . For oxidative stress assay , three or four groups of 3-day-old 30 male flies ( n = 90 or 120 ) were starved for 6 h and transferred to a vial containing a gel of phosphate-buffered saline ( PBS ) , 5% sucrose and an oxidative stress agent ( 5 mM rotenone or 1% H2O2 ) as indicated in figure legends . Dead flies were counted at the indicated time points . For life span assay , three or four groups of 30 male flies ( n = 90 or 120 ) were transferred to fresh fly food vials and scored for survival every 3 or 4 days . To check climbing activity of IDHP mutants , groups of fifteen 3- or 30-day-old males grown on normal media were transferred into climbing ability test vials and incubated for 1 h at room temperature for environmental acclimatization . After tapping the flies down to the bottom , the number of climbing flies was counted for 10 seconds . For each group , ten trials were performed , and the climbing score ( percentage ratio of the number of climbed flies against the total number ) was obtained . To check climbing ability of IDH expressing DJ-1β mutant males under oxidative stress , groups of 3-day-old 30 males were starved for 6 h and transferred to a vial containing a gel of phosphate-buffered saline ( PBS ) , 5% sucrose and 0 . 5 mM rotenone . After 4 days , they were re-grouped in size of fifteen and tested according to the procedures above . The average climbing score with standard deviation was calculated for five independent tests . For mtDNA PCR , total DNA from five thoraces of 3- or 30-day-old male flies was extracted . Then , quantitative real-time PCR was performed as previously described [32] . Genomic DNA levels of rp49 were measured for internal controls . The results were expressed as fold changes relative to the control . For ATP assay , five thoraces from 3-day-old male flies were dissected , and ATP concentration was measured as previously described [32] . The relative ATP level was calculated by dividing the measured ATP concentration by the total protein concentration . Protein concentration was determined by a bicinchoninic acid ( BCA ) assay ( Sigma ) . In the mtDNA PCR and ATP assay , the average value with standard deviation was obtained from three independent experiments . S2 cells were cultured and transiently transfected with IDHm1 , IDHm2 , and IDHc plasmids used to generate UAS-IDH flies as described previously [48] . To induce IDH protein expression in pUAST vector , we co-transfected pMT-GAL4 plasmids that contained GAL4 gene with metallothionein promoter . Twenty-four hours before cell staining , CuSO4 was treated to induce expression of GAL4 and IDHs . Cells were pre-incubated with 5 μg/mL MitoTracker Red CMXRos ( Molecular Probes ) for 1 h at 25°C and then subjected to the standard immunocytochemistry using anti-HA antibody ( Invitrogen ) . To check the change of the DA neuron numbers in DJ-1β mutants under oxidative stress , 30 male flies ( 3-day-old ) were starved for 6 h and incubated for 3 days in a vial containing a gel of phosphate-buffered saline ( PBS ) , 5% sucrose and an oxidative stress agent ( 0 . 2 mM rotenone or 1% H2O2 ) . To check the change of the DA neuron numbers in IDH or PINK1 mutants without oxidative insults , 30 male flies were transferred to a fresh normal media vial every 3 or 4 days for the time points indicated in figure legends . To stain DA neurons , adult brains from ten randomly chosen flies were fixed with 4% paraformaldehyde and stained with anti-TH rabbit antibody ( 1:50 , Pel-Freez , P40101-150 ) as previously described [32] . Brains were observed and imaged by LSM 700 confocal microscope ( Zeiss ) . For imaging ROS production in fly tissues , the indirect flight muscles from 3-day-old males were dissected in Schneider’s medium ( Sigma ) and incubated for 5 min in Schneider’s medium containing 30 μM dihydroethidium ( DHE , Invitrogen ) . Muscles were observed and imaged by BX-50 microscope ( Olympus ) . IDHRV ( IDHRV/IDHRV ) ; IDHP ( IDHP/IDHP ) ; hs ( hs-GAL4/+ ) ; hs>CG17352 ( hs-GAL4/UAS-CG17352 ) ; hs DJ-1βex54 ( hs-GAL4/+; DJ-1βex54/DJ-1βex54 ) ; hs>IDHm1 DJ-1βex54 ( hs-GAL4/UAS-IDHm1; DJ-1βex54/DJ-1βex54 ) ; hs>IDHm2 DJ-1βex54 ( hs-GAL4/UAS-IDHm2; DJ-1βex54/DJ-1βex54 ) ; hs>IDHc DJ-1βex54 ( hs-GAL4/UAS-IDHc; DJ-1βex54/DJ-1βex54 ) ; elav ( elav-GAL4/+ ) ; elav DJ-1βex54 ( elav-GAL4/+; DJ-1βex54/DJ-1βex54 ) ; elav>IDHm1 DJ-1βex54 ( elav-GAL4/UAS-IDHm1; DJ-1βex54/DJ-1βex54 ) ; elav>IDHm2 DJ-1βex54 ( elav-GAL4/UAS-IDHm2; DJ-1βex54/DJ-1βex54 ) ; elav>IDHc DJ-1βex54 ( elav-GAL4/UAS-IDHc; DJ-1βex54/DJ-1βex54 ) ; DJ-1βex54 ( DJ-1βex54/DJ-1βex54 ) ; Keap1EY5/+ ( Keap1EY5/+ ) ; DJ-1βex54 Keap1EY5/+ ( DJ-1βex54Keap1EY5/DJ-1βex54 ) ; WT ( +/Y ) ; elav>CncC DJ-1βex54 ( elav-GAL4/UAS-CncC; DJ-1βex54/DJ-1βex54 ) ; hs>CncC ( hs-GAL4 UAS-CncC/+ ) ; hs>CncC Keap1 ( hs-GAL4 UAS-CncC/UAS-Keap1 ) ; hs>CncC Keap1 DJ-1β ( hs-GAL4 UAS-CncC/UAS-Keap1; UAS-DJ-1β/+ ) ; IDHP DJ-1βex54 ( IDHP/IDHP; DJ-1βex54/DJ-1βex54 ) ; da ( da-GAL4/+ ) ; B9 da ( PINK1B9/Y;; da-GAL4/+ ) ; B9 da>IDHm1 ( PINK1B9/Y; UAS-IDHm1/+; da-GAL4/+ ) ; B9 da>IDHc ( PINK1B9/Y; UAS-IDHc/+; da-GAL4/+ ) ; TH ( TH-GAL4/+ ) ; B9 TH ( PINK1B9/Y;; TH-GAL4/+ ) ; B9 TH>IDHm1 ( PINK1B9/Y; UAS-IDHm1/+; TH-GAL4/+ ) ; B9 TH>IDHc ( PINK1B9/Y; UAS-IDHc/+; TH-GAL4/+ ) ; hs>IDHm1 ( hs-GAL4/UAS-IDHm1 ) ; hs>IDHm1RQ ( hs-GAL4/UAS-IDHm1R134Q ) ; hs>IDHm1RK ( hs-GAL4/UAS-IDHm1R166K ) ; hs>IDHm1RQ DJ-1βex54 ( hs-GAL4/UAS-IDHm1R134Q; DJ-1βex54/DJ-1βex54 ) ; hs>IDHm1RK DJ-1βex54 ( hs-GAL4/UAS-IDHm1R166K; DJ-1βex54/DJ-1βex54 ) ; elav>IDHm1RQ DJ-1βex54 ( elav-GAL4/UAS-IDHm1R134Q; DJ-1βex54/DJ-1βex54 ) ; elav>IDHm1RK DJ-1βex54 ( elav-GAL4/UAS-IDHm1R166K; DJ-1βex54/DJ-1βex54 ) . To measure transactivation activity of CncC on the IDH gene , the promoter and 5’ untranslated region ( S1C Fig ) were subcloned into pGL3 reporter plasmid ( Promega ) using following primers: IDH promoter F ( gcg ggt acc cag tta ttc gct gcg tct gat tgg ) and IDH promoter R ( gcg gga tcc gaa ccg acc gac gac tgg aaa cg ) . For generating the IDH reporter with ARE mutation , the first five bases ( TGACG ) of the putative ARE ( TGACGGGGC ) were deleted by QuikChange™ site directed mutagenesis kit ( Agilent Technologies ) . S2 cells were transfected with wild type or ARE mutant IDH reporter , pUAST-CncC , pRL-TK Renilla reporter , and pMT-GAL4 plasmids . Two days later , CncC expression was induced by CuSO4 treatment . After 24 h , luciferase assays were performed using Dual-Luciferase™ reporter assay kit ( Promega ) according to the manufacturer's instructions . The average luciferase activity with standard deviation was obtained from three independent experiments . Total RNA from heads , thoraces , abdomens , or whole bodies of 3-day-old flies or SN4741 cells was extracted and reversely transcribed as previously described [49] . To check the inhibition of IDH expression in IDHP mutants , 5 whole bodies were used ( Fig 1D and S2D Fig ) . To check the expression change of IDH and its isoforms in DJ-1β mutants , 5 heads and , thoraces , reported to be predominantly damaged in PD-gene-defected flies were used ( Figs 1D and 2E–2G ) [32] . To check whether the expression change of IDH is tissue-specific , 10 heads , 10 thoraces , or 10 abdomens were used ( S1C–S1E Fig ) . To confirm the gene expression of each isoform , 5 whole bodies were used ( Fig 3A–3C , 3F and 3G ) . SN4741 cells were seeded in 6-well plates at a density of 1 × 106 cells per well . Then , quantitative real-time PCR was performed using SYBR Premix Ex Taq ( Takara ) on Prism 7000 Real-Time PCR System ( ABI ) . rp49 levels or mouse actin levels were measured for internal control of Drosophila or SN4741 samples , respectively . The results were expressed as fold changes relative to the control . The average mRNA level with standard deviation was obtained from three independent experiments . For primer pairs , we used rp49-F ( gct tca aga tga cca tcc gcc c ) and rp49-R ( ggt gcg ctt gtt cga tcc gta ac ) , IDH-F ( cct tcc tgg aca ttg agc tg ) and IDH-R ( gta ccg ttg ggc gac ttc cac ) , CG17352-F ( cac atc tcg ttg aga gtg gat gac ) and CG17352-R ( cga atg tag tag cca ttg agg atg ) , hsp22-F ( gtc ctg acc atc agt gtg c ) and hsp22-R ( cca gtc tgc tcg atg gtc ac ) , IDHm1-F ( cat cag cgc cgc gat gg ) and IDH-R ( gta ccg ttg ggc gac ttc cac ) , IDHm2-F ( gtg agc gag atg gcc cag aag ) and IDH-R ( gta ccg ttg ggc gac ttc cac ) , IDHc-F ( gta tgc tct ccc gaa cag atg g ) and IDH-R ( gta ccg ttg ggc gac ttc cac ) , mouse IDH1-F ( cct ggg cct gga aaa gta ga ) and mouse IDH1-R ( tcc tgg ttg tac atg ccc at ) , mouse IDH2-F ( cta tga cgg gcg ttt caa gg ) and IDH2-R ( cct tga gcc agg atg tca ga ) , mouse actin-F ( ttc ttt gca gct cct tcg tt ) and mouse actin-R ( tgg atg gct acg tac atg gc ) , and mouse Keap1-F ( tgc ccc tgt ggt caa agt g ) and mouse Keap1-R ( ggt tcg gtt acc gtc ctg c ) . SN4741 cells were established from the substantia nigra region of wild type and DJ-1 knock out mouse embryos , and were characterized for expression of the neuronal markers including TuJ1 and NeuN , and the DA cell marker TH as previously described [41] . The SN4741 cells were grown in RF medium ( DMEM supplemented with 10% fetal bovine serum , 1% glucose , and 2 mM L-glutamine ) at 33°C in a humidified atmosphere with 5% CO2 . pCMV14 vector , pCMV14 FLAG-IDH1 , or pCMV14-IDH2 was transfected using Lipofectamine Plus Reagent ( Invitrogen ) according to the manufacture’s protocol . siRNAs for control ( Bioneer , #SN-1003 ) , mouse IDH1 ( Bioneer , #1371568 ) , mouse IDH2 ( Bioneer , #1371576 ) , or mouse Keap1 ( Bioneer , #1367293 ) was transfected to SN4741 cells using the RNAiMAX reagent ( Invitrogen ) according to the manufacture’s protocol . CRISPR genome editing technique was used for the deletion of DJ-1 . The guide RNA sequence ( gtg gat gtc atg cgg cga gc ) was cloned into the px459 vector . The plasmid was transfected into SN4741 cells . 48 h after transfection , transfected cells were selected by 5 μg/mL puromycin for 3 days and then single colony was transferred onto 96-well plates with one colony in each well . The clones were screened by immunoblot with anti-DJ-1 antibody ( 1:1 , 000 , Novus Biology , #NB100-483 ) . For detection of IDH1 , IDH2 , β-tubulin , and HA- or FLAG-tagged protein , S2 or SN4741 cells were lysed with Lysis Buffer [48] . The lysates were purified by centrifugation and boiled in SDS sample buffer . The samples were subjected to SDS-PAGE and proteins were transferred to nitrocellulose membrane . The membrane was incubated for 30 min in Blocking Solution and further incubated with anti-IDH1 antibody ( 1:1 , 000 , Bethyl , #A304-162A-T ) , anti-IDH2 antibody ( 1:1 , 000 , Bethyl , #A304-096A-T ) , anti-FLAG antibody ( 1:1 , 000 , MBL , #M185-3L ) , anti-HA antibody ( 1:1 , 000 , Invitrogen , #26183 ) , anti-DJ-1 antibody ( 1:1 , 000 , Novus Biology , #NB100-483 ) , or anti-β-tubulin antibody ( 1:1 , 000 , DSHB , Clone E7 ) as described previously [48] . Membrane-bound antibodies were detected with ImageQuant LAS 4000 system ( GE Healthcare Life Sciences ) . Cells were seeded in 12-well plates at a density of 6 × 105 cells per well . After pre-treatment of TIC ( LegoChem Biosciences ) or MitoTEMPO ( Sigma , 10 nM ) at the indicated concentrations for 1 h , cells were treated with 1 . 5 mM H2O2 . After 6 h incubation , the culture medium was removed and replaced with a medium containing 0 . 5 mg/mL of MTT dissolved in PBS ( pH 7 . 2 ) . After 4 h , the formed formazan crystals were dissolved in 400 μL of DMSO , and the absorbance intensity was measured at a wavelength of 595 nm using Infinite 200 pro ( TECAN ) . The relative cell viability was expressed as a percentage relative to the untreated control cells . The average viability with standard deviation was obtained from three independent experiments . SN4741 cells were seeded in 6 well plates with cell density of 1 × 106 cells per well . Treatment of TIC ( 5 mM ) and H2O2 ( 1 . 5 mM ) was performed as described above . The cells were stained using the Annexin V-FITC Apoptosis Detection kit ( BD Biosciences ) according to the manufacturer's protocol . Stained cells were analyzed by flow cytometry using BD FACSCanto II ( BD sciences ) . A total of 10 , 000 events was analyzed for each sample , and the necrotic cell death rates obtained from three independent experiments were presented as the mean values with standard deviations . SN4741 cells were pre-treated with TIC ( 5 mM ) for 1 h . Following 2 h treatment of 1 . 5 mM H2O2 , cells were incubated with 5 μM of 5- and 6-chloromethyl-2′ , 7′-dichlorodihydrofluorescein diacetate ( CM-H2DCFDA , Invitrogen ) for 30 min at 33°C . The cells were trypsinized , washed with PBS , suspended in PBS , and analyzed with BD FACSCanto II ( BD sciences ) . A total of more than 5 , 000 events was analyzed for each sample , and the results obtained from three independent experiments were presented as the mean values with standard deviations . SN4741 cells were pre-treated with TIC ( 5 mM ) for 1 h . Following 2 h treatment of 1 . 5 mM H2O2 , cells were incubated with 1 μM MitoSOX ( Invitrogen ) for 10 min at 33°C . The cells were trypsinized , washed with PBS , suspended in PBS , and analyzed with BD FACSCanto II ( BD sciences ) . A total of more than 5 , 000 events was analyzed for each sample , and the results obtained from three independent experiments were presented as the mean values with standard deviations . SN4741 cells were pre-treated with TIC ( 5 mM ) for 1 h . Following 6 h treatment of 1 . 5 mM H2O2 , the cells were lysed with 0 . 2 N NaOH with 1% dodecyl trimethyl ammonium bromide ( DTAB , Sigma ) . To measure NADPH/NADP+ ratio in flies , five 3-day-old male flies were homogenized in 0 . 2 N NaOH with 1% DTAB . Samples were centrifuged to obtain supernatants . NADP+ and NADPH levels of the lysates were individually measured by using NADP/NADPH-glo™ assay kit ( Promega ) according to the manufacturer's instructions , and NADPH/NADP+ ratio was calculated . The average NADPH/NADP+ ratio with standard deviation was obtained from three independent experiments . Ten 3-day-old male flies were homogenized in 40 mM Tris buffer ( pH 7 . 4 ) . Supernatants from samples were each added to the Tris buffer-containing NADP+ ( 2 mM ) , MgCl2 ( 2 mM ) , and isocitrate ( 5 mM ) . IDH activity was determined by monitoring the kinetics of NADPH production at 340 nm at 25°C with SpectraMax M2 multi-mode microplate reader ( Molecular Devices ) . The average relative IDH activity with standard deviation was obtained from three independent experiments . For quantification of DA neurons , four major DA neuron clusters from more than 15 brains of each genotype were observed in a blind fashion to eliminate bias ( n = 30~40 ) . To compare three or more groups , we used one-way ANOVA with Sidak correction . For two-group comparison , we used Student’s two-tailed t test . The Kaplan-Meier estimator and the log-rank test were conducted on the survival data to determine whether each treatment had any effect on the longevity of individuals using Online Application Survival Analysis Lifespan Assays ( http://sbi . postech . ac . kr/oasis ) . All n values defined in the figure legends refer to biological replicates unless otherwise indicated . The experiments were not randomized . To obtain consistent results , we incubated flies for at least three days after eclosion and excluded dead or malformed flies before any fly assay in this report . 20 male flies ( 3-day-old ) were starved for 6 h and transferred to a vial containing a gel of PBS , 5% sucrose and 5 mM rotenone . 16 h later , total RNA from ten heads and thoraces of ten randomly chosen stressed flies was extracted . Indexed RNA-seq libraries were constructed using Illumina TruSeq RNA Sample Prep Kit version 2 . Each library was sequenced in paired end using Illumina HiSeq2500 platform . Raw reads ( n = 3 ) were aligned to the Ensembl Drosophila melanogaster reference genome ( BDGP6 ) using Tophat2 . The read alignments were assembled into transcriptome assembly . Fragments per kilobase of transcripts per million reads ( FPKM ) as normalized expression levels were calculated using Cufflinks . The assemblies for each replicate were merged together using Cuffmerge . Differentially expressed gene ( DEG ) analysis was performed using Cuffdiff workflow to screen DEGs with false discovery rate ( FDR ) adjusted by P-value of < 0 . 05 and fold change of > 1 . 5 . Gene ontology ( GO ) analysis was performed for term enrichment using g:Profiler and Amigo2 . We filtered GO tree hierarchy and statistical significance threshold was FDR < 0 . 05 . A volcano plot and hierarchical clustering in a heat map were generated by statistical package R . | The molecular pathogenesis of Parkinson’s disease ( PD ) is still elusive even though many causative genes for the disease have been identified . In this study , we demonstrated that isocitrate dehydrogenase ( IDH ) , the enzyme responsible for converting isocitrate into α-ketoglutarate , is critical for the pathogenesis of PD by providing NADPH as a reducing power in the cell . IDH mutant animals showed increased reactive oxygen species ( ROS ) levels and phenotypes related to PD including dopaminergic ( DA ) neuron degeneration and locomotor defects . Conversely , elevating IDH function either by overexpression or treating a cell-permeable derivative of isocitrate , trimethyl isocitrate ( TIC ) , made DA cells resist oxidative stress and reduce ROS level , thereby suppressing PD phenotypes induced by DJ-1 mutations . These results demonstrate that IDH protects DA neurons from ROS at the downstream of DJ-1 and cell-permeable isocitrates can be novel treatments for PD . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"oxidative",
"stress",
"neurodegenerative",
"diseases",
"neuroscience",
"animals",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"genome",
"analysis",
"mitochondria",
"bioenergetics",
"cellular",
"structures",
"and",
"organelles",
"drosophila",
"research",
"and",
"analysis",
"methods",
"genomics",
"animal",
"cells",
"gene",
"expression",
"gene",
"ontologies",
"insects",
"arthropoda",
"movement",
"disorders",
"biochemistry",
"cellular",
"neuroscience",
"cell",
"biology",
"phenotypes",
"neurology",
"neurons",
"genetics",
"parkinson",
"disease",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"energy-producing",
"organelles",
"computational",
"biology",
"organisms"
]
| 2017 | Isocitrate protects DJ-1 null dopaminergic cells from oxidative stress through NADP+-dependent isocitrate dehydrogenase (IDH) |
Collective cell migration plays an important role in development . Here , we study the posterior lateral line primordium ( PLLP ) a group of about 100 cells , destined to form sensory structures , that migrates from head to tail in the zebrafish embryo . We model mutually inhibitory FGF-Wnt signalling network in the PLLP and link tissue subdivision ( Wnt receptor and FGF receptor activity domains ) to receptor-ligand parameters . We then use a 3D cell-based simulation with realistic cell-cell adhesion , interaction forces , and chemotaxis . Our model is able to reproduce experimentally observed motility with leading cells migrating up a gradient of CXCL12a , and trailing ( FGF receptor active ) cells moving actively by chemotaxis towards FGF ligand secreted by the leading cells . The 3D simulation framework , combined with experiments , allows an investigation of the role of cell division , chemotaxis , adhesion , and other parameters on the shape and speed of the PLLP . The 3D model demonstrates reasonable behaviour of control as well as mutant phenotypes .
In this section , we describe additional biological background for our investigation . The PLLP consists of more than 100 cells ( ≈4-5 cells wide ) when migration is initiated , which migrate about 2 . 3 mm down the horizontal myoseptum during the time interval 22-48 hours post-fertilization [1] . Cells in the trailing region re-organize into rosettes , are periodically deposited and later develop into neuromasts , to form sensory organs of the posterior lateral line . The cells at the leading region are flatter and have a mesenchymal morphology [2] . The changes in cell morphology and behaviour across the PLLP is thought to be regulated by antagonistic Wnt-FGF signalling network [2–4] . The leading zone of the PLLP is characterized by high-levels of WntR activity while the trailing zone is dominated by FGFR activity . It is not known which Wnt ligand is involved although recent experimental work by members of our group ( DDN and AC ) implicates Wnt10a . More research has been done on the downstream Wnt signalling network [4] . WntR activity at the front of the PLLP promotes the expression of FGF3 and FGF10 ligands , as well as sef , considered to inhibit FGFR activity . Consequently , FGF pathway activation is inhibited in the leading cells , and the secreted FGF ligand can redistribute towards the rear . Activation of FGF signalling in the trailing domain subsequently drives the expression of Dkk1 , a diffusible inhibitor of Wnt signalling [4] . In this way , Wnt and FGF signalling domains are established in the leading and trailing zones of the PLLP . Migration is dependent on chemokine signalling across the primordium , and is guided by a strip of CXCL12a , a chemokine that is produced by superficial muscle pioneer cells of the horizontal myoseptum , has limited diffusion on the surface of this layer , and determines the PLLP’s migration path [5–7] . Without CXCL12a the PLLP stalls . However , directional information is not encoded by CXCL12a , whose expression level is uniform . Instead , the differential expression of two chemokine receptors , CXCR4b and CXCR7b , across the length of the PLLP guides migration . Similar to the Wnt-FGF polarity , the expression of CXCR4b dominates in the leading end while expression of CXCR7b dominates in the trailing zone with some overlap [5 , 7] . Mechanisms that determine the differential expression of chemokine receptors in the PLLP are linked to the development of the polarized Wnt-FGF signalling , but are as yet incompletely understood [4 , 10] . Cells expressing CXCR4b respond to CXCL12a with protrusive and migratory behaviour . The cells expressing CXCR7b rapidly internalize and degrade CXCL12a . It has been proposed that this sets up a local gradient of CXCL12a that ensures unidirectional migration of the PLLP [9] . It was also suggested that the absolute level of CXCL12a serves as cue [8] . Moreover , recent laser ablation and cutting experiments in [8] revealed that the trailing cells undergo chemotaxis toward FGF ligand secreted by the leading PLLP cells . Evidence for this stems from several experiments described in [8]: ( 1 ) when trailing cells are physically separated from the leading cells by laser ablation , trailing cells move toward and eventually join with leading cells despite unpolarized expression of the chemokine receptor CXCR4b in the trailing fragment . ( 2 ) This movement is inhibited by treatment with the FGF inhibitor SU5402 , suggesting that FGF signaling is essential for this process . ( 3 ) Furthermore , an isolated trailing cell cluster generated by laser ablation can be induced to migrate directionally by placing a bead soaked in recombinant FGF3 either ahead of or behind the cells . This suggests that while the leading and trailing cells coordinate migration by inducing a local CXCL12a gradient , the leader cells respond to chemokine cues and the trailing cells migrate toward the leading cells [8] .
Our initial step was to explore mechanisms for the subdivision of the tissue into a Wnt-signalling leading domain and a FGF-signalling trailing domain . We specifically sought to determine how a spatially graded parameter , such as the steady state level of the Wnt receptors ( W1 ) , across the length of the primordium can set up this domain subdivision , and where the domain boundary would be positioned . We first explore this from a mathematical perspective , linking the outcome to underlying biochemistry , and then show the results numerically . We employ a 1D model simplification followed by the full 3D simulation for the PLLP . Having identified conditions for the formation of leading and trailing zones , we next asked what further conditions are required for collective migration of the PLLP . To probe this question , we assembled basic facts from the biology ( surveyed above ) supplemented , where needed , by reasonable minimal assumptions . In actual PLLP experiments , it is known that aside from the Wnt-FGF polarization of the tissue , the front and rear portions are also distinct in levels of chemokine receptors . Cells in the leading 60% of the PLLP express high levels of CXCR4b , whereas those in the remaining rear portion express high CXCR7b , [4] . This does not exactly coincide with the growth-factor zones: the leading 1/3 of the PLLP is the WntR activity zone [18] . However , here we made the simplification that from the standpoint of interactions with CXCL12a , the WntR-FGFR subdivision serves as reasonable proxy for the chemokine zone subdivision . One reason for this simplifying assumption is that we thereby avoid the further complications of untangling the signalling circuits that set up the zones of differential chemokine receptor expression . A second reason is that little is to be gained by this intermediate step currently , when the detailed biology of this step is largely unknown . It is well known that cell growth and division take place during the course of PLLP migration with the total number of cells doubling throughout the process [18 , 25] . This fact led us to inquire how such cell proliferation contributes to the zone formation , migration , and overall shape of the PLLP . While it was originally thought that leading zone cells proliferate faster [18 , 25] , more recent experiments by two of us ( AC and DDN ) revealed a relatively constant and spatially uniform cell cycle length ( 539 ± 127 minutes ) across the PLLP [26] . In the simulations the cell’s rate of growth in each time step is chosen from a uniform distribution between 0 and 0 . 001 min−1 ( corresponding to a cell cycle between 0 and 1000 minutes ) . Cell division occurs when the volume of a given cell has doubled . The daughter cells have the same parameter values as the mother cell . Results with a range of growth rate values are shown in Fig 10 . For the higher growth rate , we find that as the PLLP migrates and elongates , the cells at the back no longer detect FGF , and eventually break away from the rest of the PLLP . The spatiotemporal FGF and Wnt ligand distribution is hard to quantify experimentally . Hence , direct comparisons of model predictions to experimental observations are not possible for the chemical distributions of FGF and Wnt ligands . For this reason , we sought simple experimental comparisons that could ( in ) validate the computational model . We consequently compared our model behaviour to the results of laser ablation ( cutting ) experiments , even though such experiments have already been successfully recapitulated in a computational model [8] . We further simulated several mutants to check whether our predictions agree with what has been observed . In the experiments , staining has been done for Wnt signalling cells , using lef1 as a marker , and for FGF signalling cells , using pea3 as a marker . Hence we color the cells in our simulations based on signalling cell behaviour .
Here , we investigated the lateral line primordium development and migration in zebrafish . In this intriguing system , a cluster of 100-200 cells maintains cohesion and directionality while migrating over a distance one order of magnitude longer than the cluster length . We first showed that the mutual inhibition of Wnt and FGF signalling can , on its own , set up the initial polarization of the PLLP into distinct leading and trailing domains under appropriate assumptions . Our model links ligand-receptor binding ( in the Michaelis-Menten regime ) to the inhibitory effect ( of bound receptors on the antagonist receptors ) . Thereby , we could predict how the boundary between Wnt and FGF domains depends on parameters such as steady state receptor levels ( in absence of inhibition ) and sensitivity to inhibition by the antagonist bound receptors . The theoretical model discussed here has more universal applicability , and variations on this theme could be at play in other differentiation processes of tissue subdivision . In particular , “crossed gradients” ( i . e . gradients of opposite slopes ) in the two ratios of receptor-ligand properties that we identified as ϕ and ω could produce a tissue with a central zone of overlapped influence ( transition across the bottom panels of Fig 4 ) , or with a sharp boundary ( transition across the top panels of this figure ) . For our model of the PLLP specifically , such crossed gradients do not exist because the spatial variation of ϕ and ω stem from the FGF and Wnt ligand distributions , which are both graded in the same direction ( increasing towards the front ) . Nevertheless , tissue subdivision occurs , once the PLLP evolves into a bistable state , with WntR activity locked into the front , and FGFR activity overtaking WntR in the back . We have shown that the CXCL12a-expressing adhesive stripe provides directionality , as well as , coupled to active degradation by trailing cells , forms the guidance cue that directs PLLP migration along the lateral axis of the embryo . Given the viscous drag , it appears highly probable that all cells in the PLLP are actively motile . Simply having a “front engine” ( a few cells at the front ) pulling the entire mass would lead to migration stalling , and even fragmentation of the group , which is not observed under normal conditions . Chemotaxis is the simplest and most plausible mechanism for the active motion of trailing cells , though we do not excluded the possibility that trailing cells also actively follow leader cells through some other mechanism . The shape and speed of the collective migrating mass depends on numerous factors such as rate of cell growth ( which determines cell division rate ) , active crawling and adhesion of cells to one another and to the underlaying adhesive substrate . When these factors are increased or decreased beyond certain ranges , shape and speed are abnormal , recapitulating a variety of mutants . Wnt signalling regulates cell growth rate , and FGF signalling is implicated in cohesion . Both signalling types determine differentiation and the distinct roles of leading and trailing cells . WntR active front cells sense the CXCL12a gradient and steer the cluster . Rear cells eventually express CXCR7b chemokine receptors and degrade CXCL12a . The model demonstrates that this could form the self-propagating gradient to which WntR cells orient . FGFR active cells eventually engender the initiation of neuromast deposition that we have not yet addressed in this paper . It is interesting to compare our results with those of previous modelling papers [8–12] . In Streichan et al . [9] , the PLLP is modelled as a 1D rod , moving through a self-generated gradient of a single chemokine ( CXCL12a ) , similar in flavor to “treadmilling” solutions in [27] . The rod binds the diffusible chemokine and moves with velocity assumed to be proportional to the gradient so created . The model for chemokine profile and rod position sustains a travelling wave solution , representing the steady-state long-range PLLP migration . The authors further considered deposition , assuming that below a threshold gradient of CXCL12a , the rear cells in a discrete 1D elastic array fall off . They derive an analytic expression for the migration velocity as a function of rates of ligand degradation and diffusion , the size of the PLLP and the rate of chemokine uptake . This model provides elegant analytic insights into minimal requirements for “self-generated” gradient chemotaxis in a general setting , while simplifying the biology and neglecting key features of the PLLP . For example , the authors do not consider interactions between CXCR7b and CXCL12a ligand , which is now well established in the literature . In comparison , our model begins to address this limitations by including some of these biochemical interaction , consequently limiting many of our results to numerical experiments . Two of us ( AC , DDN ) previously crafted a Netlogo model [8 , 10] as an aide to deciphering laser ablation experimental observations . The simulation uses “turtles” attached to one another with springs: when one moves , its neighbor moves with a slight lag , disallowing overlaps . This simulation had the advantage of close integration with experiments . While remaining a fairly elementary , it was able to capture key aspects of the biology , and explain experiments . A limitation was the absence of realistic forces of cell-cell interaction and friction . Consequently , a conclusion , based partly on the assumption of pulling exerted by few leading cells , was that cells respond to the CXCL12a concentration rather than CXCL12a gradient . In our simulation here , we endow a larger number of cells with the ability to sense , orient to , and migrate up the CXCL12a gradient to correct that feature . Furthermore , as we have argued , drag forces would stall the PLLP were it not for additional chemotaxis motion of the trailing cells . Allena and Maini [11] described chemical signalling coupled to cellular mechanics in the PLLP using a reaction-diffusion finite element model . Cells are depicted by 2D Maxwell viscoelastic elements inside an elliptical PLLP . The incorporation of additional signalling ( FGF and Wnt ligands , CXCR4b and CXCR7b chemokines as “readouts” of the ligand distribution ) , coupling mechanics and biochemistry , and the successful recapitulation of mutant behaviours are all strengths of this model . As in Allena and Maini [11] , we have adopted a hybrid model to describe the PLLP dynamics in 3D . Our discrete cell model builds on limitations of [11] by including interactions between the PLLP and CXCL12a , relative motion and rearrangement of cells , and more direct representation of intercellular forces . In contrast to [11] , our model also explains the initial subdivision of the PLLP into leading and trailing zones . Another hybrid model [12] is based on cell-cell interaction , cell-substrate adhesion and cell differentiation . The model consists of discrete cells in 2D with a non-local sensing radius . The PLLP is described as a “flock” of cells accelerating via haptotaxis to CXCL12a , forces of Cucker-Smale alignment , attraction-repulsion , and chemotaxis to FGF gradient . The model captures the migration of the PLLP resulting from a graded distribution of CXCL12a . One issue with the representation of the PLLP as a “flock” is that macroscopic flocks and swarms generally operate in a high Reynold’s number regime , whereas we have chosen to neglect inertial forces and acceleration terms and to consider a low Reynold’s number regime for our cellular milieu . One interesting feature of [12] is that , by including cell-cell lateral inhibition in the trailing domain ( and foci of FGF ) , their model can depict the formation of stable rosette structures within the PLLP ( see their Figs 9 and 10 ) . In contrast , we do not yet consider the rosette formation in our model but we base our cell motion and interactions on a closer concordance with recent biological results . In particular , we include downstream effects of Wnt/FGF signalling , i . e . , formation of receptor activity domains , chemotaxis of trailing cells to FGF , and degradation of SDF1a ( CXCL12a ) in the trailing domain . All models described above , including our own , would benefit from further experiments that could help to probe hypotheses or compare with predictions , in particular , the following experiments would be of primary interest: Primarily , our suggested experiments seek to determine the roles of the Wnt and FGF ligands in sustained organization and migration of the PLLP . Does an initial signalling event start the PLLP migratory process , or , as we have assumed in our model , is a continuous source of Wnt ligand required to maintain the organization and migration of the PLLP ? In summary , we are not the first to model the PLLP migration , though ours is the first fully 3D simulation of its kind , with realistic cell-cell and cell-substrate forces . While the model is a drastic simplification of the biology , it nevertheless provides insights about the Wnt-FGF mutual inhibition , forces of propulsion and drag , and accounts for mutants . There are several opportunities for future work . First , there is the process of neuromast deposition that we consider in a future study . Second , there are many additional players in the signalling network responsible for segmenting the PLLP into leading and trailing zones , such as Dkk1 and sef , and including these players could aid in the understanding of their specific roles within the PLLP .
In these simulations we treat the system in one spatial dimension across the long axis of the PLLP . We then define the spatial variable x with values in 0 ≤ x ≤ L where x = L is the front of the primordium and x = 0 is the back . All boundaries were modelled as impermeable ( no flux ) . We used the MATLAB pdepe function to approximate solutions to the system . The sharp boundary between the leading and trailing domains makes numerical approximation difficult . For this reason , we only approximate the solution over a short time interval . To obtain the position of the boundary between signalling domains in Fig 6 , the 1D Wnt-FGF model was simulated to steady-state with 20 distinct parameter values ( F0 , F1 , W0 , pF , pW , b ) , no flux boundary conditions , and initial conditions W ( x , 0 ) = 0 . 01 , F ( x , 0 ) = 0 . 005 , WR ( x , 0 ) = 0 . 1 , FR ( x , 0 ) = 0 . 01 . From the steady-state concentration profiles , the boundary between the signalling domains was identified as the intersection between the Wnt ( WR ) and FGF receptor ( FR ) curves . For Fig 6 , we plotted the numerically calculated boundary position versus p/pbaseline , the value of a given parameter , scaled by its baseline value in Table A in Supporting Information S6 Text . Parameter Estimation and Values . For the parameter sweeps , the following ranges were used: b ∈ [0 , 0 . 05] , F0 ∈ [0 , 0 . 2] , F1 ∈ [0 , 2] , W0 ∈ [0 . 0005 , 0 . 0015] , pF ∈ [15 , 35] , pW ∈ [2 , 6] . The discrete-cell computational framework is modified from software developed by one of us ( EP ) for signalling and motility in multicellular systems [24] . ( See Supporting Information S4 Text . Discrete cell model and [24 , 28] for technical details . ) Cells are modelled as deformable ellipsoids in a three-dimensional ( 3D ) domain . Cells move relative to one another , exert and respond to forces ( chemotactic , adhesive , contact/exclusion , and random ) in a low Reynolds number regime ( neglecting inertia ) . FGF and Wnt ligand concentrations evolve in 3D based on a set of PDEs modified from Eqs . ( 19 ) in Supporting Information S4 Text . Discrete cell model to account for sources and sinks at individual cells . The ligands are depleted both through binding and internalization of the bound receptors as well as through nonspecific degradation in the domain . Cells can orient and move in the direction of a chemical gradient . In addition to chemotactic forces , the cells also adhere to other cells and to the surface on which they crawl . The strength of the cell-cell adhesion force is the same for all types of cells . However , cells adhere more strongly to the CXCL12a expressing surface , and , since CXCL12a forms a narrow stripe , this accounts for the elongated shape of the primordium . While direct data is not available , it is likely that some differences in adhesion exist between cell types . Here we arbitrarily assumed that WntR expressing cells have a 2-fold higher adhesion to the CXCL12a stripe than do other types of cells . This does not have a major qualitative impact on our results . The net force acting on each cell is calculated and the cells are moved according to the equation of motion ( 25 ) . In a given simulation , the spatial concentration of the signalling molecules is calculated and updated , and then the cell positions are recomputed and updated each timestep . CldnB:lynGFP [6] embryos were treated from 32-36hpf with the indicated concentrations of SU5402 while shaking and then fixed in 4% PFA overnight . In situ hybridization for lef1 transcript was performed as previously described [29] . Following in situ hybridization , embryos were stained with anti-GFP antibody to visualize the CldnB:lynGFP . The lef1 ratio was computed by dividing the length of the lef1 domain by the total length of the PLLP . | Collective migration of a group of cells plays an important role in the development of an organism . Here we study a specific example in the zebrafish embryo , where a group of about 100 cells ( the posterior lateral line primordium , PLLP ) , destined to form sensory structures , migrates from head to tail . We model the process from the initial polarization to the migration , with a focus on how tissue polarity could arise . Using a 3D deformable-ellipsoid cell-based simulation , we explore the effects of cell-cell , cell-substrate , and cell-chemical interactions . We discuss drag forces experienced by cells and what that implies about the inherent active motion of both leading and trailing cells . The model allows us to test how each of several biological parameters affects the shape , size , effective migration and speed of migration . A subsequent study will be aimed at understanding the formation and deposition of neuromasts . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"signal",
"transduction",
"signal",
"inhibition",
"computer",
"and",
"information",
"sciences",
"cell",
"motility",
"developmental",
"biology",
"cell",
"biology",
"simulation",
"and",
"modeling",
"network",
"analysis",
"cell",
"migration",
"chemokines",
"biology",
"and",
"life",
"sciences",
"signaling",
"networks",
"wnt",
"signaling",
"cascade",
"cell",
"signaling",
"chemotaxis",
"signaling",
"cascades",
"research",
"and",
"analysis",
"methods"
]
| 2017 | Polarization and migration in the zebrafish posterior lateral line system |
The various roles that aggregation prone regions ( APRs ) are capable of playing in proteins are investigated here via comprehensive analyses of multiple non-redundant datasets containing randomly generated amino acid sequences , monomeric proteins , intrinsically disordered proteins ( IDPs ) and catalytic residues . Results from this study indicate that the aggregation propensities of monomeric protein sequences have been minimized compared to random sequences with uniform and natural amino acid compositions , as observed by a lower average aggregation propensity and fewer APRs that are shorter in length and more often punctuated by gate-keeper residues . However , evidence for evolutionary selective pressure to disrupt these sequence regions among homologous proteins is inconsistent . APRs are less conserved than average sequence identity among closely related homologues ( ≥80% sequence identity with a parent ) but APRs are more conserved than average sequence identity among homologues that have at least 50% sequence identity with a parent . Structural analyses of APRs indicate that APRs are three times more likely to contain ordered versus disordered residues and that APRs frequently contribute more towards stabilizing proteins than equal length segments from the same protein . Catalytic residues and APRs were also found to be in structural contact significantly more often than expected by random chance . Our findings suggest that proteins have evolved by optimizing their risk of aggregation for cellular environments by both minimizing aggregation prone regions and by conserving those that are important for folding and function . In many cases , these sequence optimizations are insufficient to develop recombinant proteins into commercial products . Rational design strategies aimed at improving protein solubility for biotechnological purposes should carefully evaluate the contributions made by candidate APRs , targeted for disruption , towards protein structure and activity .
Irreversible β-strand driven protein aggregation and amyloidogenesis is a tremendous burden to biological organisms . Protein loss-of function due to aggregation causes stress to the cell and metabolic energy is lost on the expression , synthesis , and degradation of proteins which aggregate . To overcome these challenges and build cellular machineries that can sustain metabolic flux , higher organisms have developed sophisticated protein quality control mechanisms , including molecular chaperones , post-translational modifications , and degradation/clearance pathways to prevent aggregation from disrupting homeostasis [1]–[3] . When quality control mechanisms are impaired , due to aging or otherwise , protein aggregation can lead to ‘conformational diseases’ in humans and animals [1] , [3]–[5] . Despite its deleterious effects , protein aggregation remains unavoidable due to the inherent physico-chemical properties of protein sequences and the formation of non-native conformations due to sequence mutation or unfolding events in response to environmental stress . However , studies of amyloidogenic proteins have revealed that different protein sequences vary in their propensity to aggregate , which can be attributed to the presence of aggregation-nucleating short sequence stretches , capable of forming the cross-β steric zipper motif , called aggregation prone regions ( APRs ) [6]–[10] . Analyses of APRs indicate common sequence properties including a high preference for β-branched hydrophobic residues , strong β-sheet propensity , low net charge , and in the case of fibril forming patterns , position-specific charged residues [11] , [12] . Knowledge of these properties has enabled the development of phenomenological and first-principle based methods to predict APRs in any protein sequence [13]–[20] . The availability of computational APR prediction tools has facilitated large-scale investigations into the aggregation propensities of protein sequences [21]–[27] . Analyzing intrinsically disordered protein ( IDP ) sequences using APR prediction tools has revealed that the number of APRs found in IDPs is three times less than those found in sequences for ordered proteins [21] . Given the tendency for APRs to exist in ordered sequence regions , it was proposed that APRs may have a role in promoting structural order in native folds . More recent studies have extended the concept that APRs can play a role in promoting structural order based on the prevalence of APRs in protein-protein interactions sites [22] , [28] , including antibody-antigen interfaces [23] . In fact , the trend for APRs to exist in protein-protein interaction sites has led some research groups to repurpose their APR prediction methods into tools for identifying potential protein-ligand [29] and protein-protein interaction sites [28] . On the other hand , mounting evidence from analyses of large sequence datasets strongly suggests that nature is actively minimizing the occurrence/impact of aggregation prone regions in protein sequences . For example , APRs are frequently punctuated by charged or proline residues ( labeled gate-keepers ) [24] , proteins with higher aggregation propensities have shorter half-lives in the cell [25] , and mRNA expression levels in E . coli are lower for proteins with higher aggregation propensities [26] . A study linking protein evolution and aggregation discovered an overall decreasing trend for aggregation propensity among organisms with increasing complexity and longevity [30] . This implies that organisms have evolved by minimizing the aggregation propensity of their proteomes . Subsequently , it was determined that the aggregation propensity differences between prokaryotic and eukaryotic proteomes could be explained by differences in their number of IDPs , which have fewer APRs than ordered proteins do [27] . This contradictory finding , suggests that organisms may not have minimized the aggregation propensity of their proteomes during the course of evolution . In light of the above reports , there is a need to further investigate the various roles that aggregation prone regions have in protein structure and the concept that nature is actively minimizing the aggregation propensity of protein sequences . This report presents findings from comprehensive analyses on multiple non-redundant datasets of protein sequences and structures ( see Methods ) . Datasets were analyzed for aggregation propensity differences among monomeric proteins , IDPs , and randomized sequences with uniform and natural amino acid compositions . To assess if evolutionary selective pressure has minimized the aggregation propensity of protein sequences through APR disruption , average sequence identity among homologous proteins was compared to percent APR conservation among the same sequences . IDPs and monomeric proteins were used to evaluate the role that APRs have in promoting structural order and stabilizing interactions . Proximity of APRs to catalytic sites in enzyme structures was also investigated . Our findings suggest that proteins have evolved by optimizing their risk of aggregation for cellular environments through the overall minimization of aggregation prone regions and the conservation of those important for folding and function . In addition to promoting aggregation under conditions that destabilize proteins , APRs also stabilize protein structure and resist disorder , particularly , in structural areas that are important for protein function . Therefore , strategies employing site directed mutagenesis to improve protein solubility should carefully evaluate the contributions made by candidate APRs , targeted for disruption , towards protein structure and activity . The major findings from this work are summarized in Table 1 .
Tables 2–4 present data on the various statistical measures of aggregation that were used in this study . Aggregation propensities were computed by normalizing total TANGO [11] and WALTZ [12] aggregation scores for protein sequences by their lengths . The average aggregation propensities for sequences of our various datasets are summarized in Table 2 . For each dataset , the average length of protein sequences which contain at least one APR , total number of predicted APRs , average APR lengths and average proportion of APR residues in protein sequences are summarized in Table 3 . The incidence of gate-keeper residues in APR flanking regions for sequences in R10000 , N10000 , SF49500 , F495 and IDP536 datasets is presented in Table 4 . Figure 1 shows box and whisker plots of the aggregation propensities for all datasets . These plots indicate an overlap among the aggregation propensity distributions for the various datasets . Overlaps are expected because the amino acid sequences in all datasets are composed of the same twenty types of naturally occurring amino acids found in proteins . Notwithstanding the overlaps , it can be seen that there are important differences among datasets in the inter-quartile ranges for both TANGO and WALTZ predicted aggregation propensities . To determine the statistical significance of aggregation propensity distribution differences among datasets , two sample t-tests were performed ( See Methods ) and the results are summarized in Table S1 of the Supplementary Material . In the following discussion , aggregation property averages ( e . g . propensity or APR length , etc . ) are compared among the various datasets . Standard deviations are provided along with averages . Amino acid sequences in N10000 and SF49500 have a lower TANGO average aggregation propensity ( i . e . average normalized aggregation score ) than those in R10000 ( Table 2 ) . The average TANGO aggregation propensity for N10000 is 5 . 42±5 . 18 , which is ∼16% lower than the corresponding value for R10000 , 6 . 48±5 . 89 . The average TANGO aggregation propensity for SF49500 is 5 . 20±4 . 33 , almost 20% lower . Consistent with these averages , boxes representing inter-quartile ranges of TANGO aggregation propensities for N10000 and SF49500 are smaller ( Figure 1 ) and the distributions of TANGO predicted aggregation propensities for N10000 and SF49500 are significantly different from those in R10000 at the 95% confidence level ( Table S1 ) . Interestingly , the percentage of sequences with at least one TANGO predicted APR in SF49500 is 71 . 4% , which is higher than that in N10000 ( 62 . 6% ) or R10000 ( 68 . 9% ) . The number of APRs per sequence is also higher for SF49500 ( 59878 APRs/49500 sequences ) = 1 . 21 ) than that for R10000 ( 10474/10000≈1 . 05 ) or N10000 ( 8965/10000≈0 . 90 ) , ( Table 3 ) . However , the average length of APR containing sequences in SF49500 is 149±35 , which is greater than that for R10000 and N10000 ( length 100 ) , and the average proportion of APR residues ( see Methods , Equation 1 ) in SF49500 ( 10 . 3±5 . 4% ) and N10000 ( 12 . 2±6 . 3% ) is lower than those in R10000 ( 13 . 4±7 . 2% ) , ( Table 3 ) . These values clarify why the percentage of sequences with at least one TANGO predicted APR is greater in SF49500 than in R10000 but the average aggregation propensity is lower for SF49500 than in R10000 . The smaller average TANGO aggregation propensity of N10000 and SF49500 compared to R10000 is a consequence of their differences in amino acid composition . The amino acid distributions in N10000 and SF495000 are identical and significantly different from the amino acid distribution in R10000 . The χ2 value for the amino acid distributions in R10000 versus N10000 ( and SF49500 ) is 160 . 84 . For 19 parameter distributions , χ2 values >43 . 82 reject the null hypothesis at the 99 . 9% confidence level ( p-value<0 . 001 ) that the two amino acid distributions are the same [35] . N10000 and SF49500 contain both fewer aggregation-promoting residues and more gate-keeper residues than R10000 . The total proportion of aggregation-promoting β-branched nonpolar ( Ile and Val ) , aromatic ( Phe , Tyr and Trp ) and polar ( Asn and Gln ) residues is smaller in N10000 and SF49500 ( 29% ) than in R10000 ( 35% ) while the total proportion of gate-keeper charged and proline residues ( Asp , Glu , Lys , Arg and Pro ) is greater in N10000 and SF49500 ( 29 . 3% ) than in R10000 ( 25% ) . For WALTZ predicted APRs , the average aggregation propensities for N10000 and SF49500 are again smaller ( 2 . 40±2 . 75 and 2 . 34±2 . 48 , respectively ) than that for R10000 ( 3 . 39±3 . 33 ) , ( Table 2 ) . WALTZ predicted APRs also constitute 7 . 2±3 . 8% of residues in SF49500 ( 8 . 5±3 . 9% in N10000 ) as compared to 9 . 4±4 . 7% in R10000 , ( Table 3 ) . Note , only sequences that contain at least one APR were used in these calculations . WALTZ aggregation propensity box and whisker plot shows that the inter-quartile ranges for N10000 and SF49500 are smaller than the inter-quartile range for R10000 ( Figure 1 ) . The distribution of WALTZ aggregation propensities in R10000 is also significantly different from those in N10000 and SF49500 at 95% confidence level ( Table S1 ) . Other observations for N10000 and SF49500 versus R10000 show similar trends as in TANGO predictions ( Tables 2 and 3 ) . In summary , the above statistical measures have highlighted the importance of amino acid composition in explaining the average TANGO and WALTZ aggregation propensity differences between N10000 and SF49500 versus R10000 . Sequences of monomeric proteins in F495 have a lower average TANGO aggregation propensity than randomized sequences with natural amino acid composition ( N10000 and SF49500 ) , ( Table 2 ) . The amino acid sequences in N10000 and SF49500 lack the sequence patterning features of F495 due to randomization and scrambling ( see Methods ) . The average TANGO aggregation propensity for F495 ( 3 . 41±2 . 98 ) is 34% lower than that for SF49500 ( 5 . 20±4 . 33 ) and 37% lower than that for N10000 ( 5 . 42±5 . 18 ) ) , ( Table 2 ) . Box and whisker plots of TANGO aggregation propensities indicate that the inter-quartile range for F495 is smaller and the third quartile ( Q3 ) is shifted to lower values compared to SF49500 and N10000 ( Figure 1 ) . The distribution of TANGO aggregation propensities in F495 is also significantly different , at the 95% confidence level , from those in N10000 and SF49500 ( Table S1 ) . The aggregation propensity differences between F495 versus N10000 and SF49500 result from a similar , but broader , set of APR differences as those between N10000 , SF49500 and R10000 . Specifically , the percentage of sequences with at least one TANGO predicted APR decreases from 71 . 4% in SF49500 ( 62 . 6% in N10000 ) to 60 . 4% in F495 , even though , the average length of APR containing sequences in F495 ( 152±34 residues ) is similar in SF49500 ( 149±35 ) , and longer than that in N10000 ( 100 residues ) , ( Table 3 ) . The average proportion of APR residues also decreases from 10 . 3±5 . 4% in SF49500 ( 12 . 2±6 . 3% in N10000 ) to 7 . 8±4 . 1% in F495 and the average APR length in F495 decreases by one residue from 8 . 48±2 . 85 in SF49500 ( 8 . 49±2 . 88 in N10000 ) to 7 . 47±1 . 99 in F495 ( Table 3 ) . The above analysis of TANGO predicted APRs reveals the importance of sequence patterning in explaining the average aggregation propensity differences between N10000 and SF49500 versus F495 . Similar observations have been made by Chiti and coworkers via experiments on a 29 residue peptide from horse heart apomyoglobin . Sequence scrambled variants of the peptide showed substantial increases in aggregation propensity compared to the wild-type sequence due to clustering of amyloidogenic residues [36] . Aggregation propensity differences between F495 versus N10000 and SF49500 for WALTZ predicted APRs are dissimilar to those of TANGO predicted APRs ( Tables 2 and 3 , and Figure 1 ) . The inter-quartile range for WALTZ predicted aggregation propensities in F495 is similar to the inter-quartile ranges for SF49500 and N10000 ( Figure 1 ) and t-tests show that the distribution of WALTZ predicted aggregation propensities in F495 is statistically similar to SF49500 and N10000 distributions . The average WALTZ aggregation propensity for sequences in F495 ( 2 . 52±2 . 70 ) is comparable to that for sequences in N10000 ( 2 . 40±2 . 75 ) and SF49500 ( 2 . 34±2 . 48 ) . The average APR lengths for WALTZ predicted APRs in SF49500 ( 6 . 13±0 . 61 ) , N10000 ( 6 . 12±0 . 63 ) and F495 ( 6 . 13±0 . 68 ) are also the same ( Table 3 ) . The percentage of sequences that contain at least one WALTZ predicted APR in SF49500 and F495 are again similar at 63 . 6% and 63 . 2% respectively . The average proportion of APR residues in sequences of SF49500 ( 7 . 2±3 . 8% ) , N10000 ( 8 . 5±3 . 9% ) and F495 ( 7 . 7±3 . 9% ) are also similar ( Table 3 ) . From the observations above , scrambling sequences ( SF49500 ) or producing random sequences with natural amino acid compositions ( N10000 ) did not increase the number of WALTZ predicted APRs as it did for TANGO . WALTZ uses position-specific matrices that enable the program to predict APRs with fibril forming charged residues [12] . Consequently , sequence randomization may not have produced sequence patterns that WALTZ position-specific matrices have defined as aggregation prone . After dividing F495 into different SCOP subclasses , the average aggregation propensity and other statistical measures of aggregation , such as proportion of sequences with at least one APR , number of APRs per sequence and average proportion of APR residues , remain similar to F495 for all subclasses ( Tables 2 and 3 ) , except in the α/β subclass . The α/β subclass contains only 43 ( ∼9% ) proteins from F495 . The average TANGO aggregation propensity of the α/β subclass ( 4 . 95±2 . 72 ) is higher than that for F495 ( 3 . 41±2 . 98 ) , but the corresponding values for WALTZ aggregation propensities are similar ( 2 . 37±2 . 12 for α/β and 2 . 52±2 . 70 for all F495 ) , ( Table 2 ) . A greater percentage of sequences in the α/β subclass contain one or more TANGO/WALTZ predicted APRs ( TANGO , 79 . 1% for α/β subclass versus 60 . 4% for F495; WALTZ , 72 . 1% for α/β subclass versus 63 . 2% for F495 ) and the number of APRs per sequence is also higher for the α/β subclass ( TANGO , 1 . 53 for α/β subclass versus 0 . 91 for F495; WALTZ , 1 . 35 for α/β subclass versus 1 . 09 for F495 ) , ( Tables 2 and 3 ) . However , the average APR lengths in the α/β subclass are similar to that for all F495 protein sequences . Excluding the α/β subclass , the results in Tables 2 and 3 for the various SCOP classes suggest that differences in topology and secondary structure content alone do not produce average aggregation propensity changes that are similar to those observed after modifying sequence composition and patterning . Average TANGO aggregation propensities are similar between IDP536 and F495 ( 3 . 85±4 . 05 for IDP536 and 3 . 41±2 . 98 for F495 , Table 2 ) . Box and whisker plots for IDP536 and F495 also show that these datasets have similar predicted aggregation propensities ( Figure 1 ) . The χ2-test on amino acid compositions of F495 and IDP536 yields a value of 20 . 4 , thereby , accepting the null hypothesis that they have the same amino composition . The number of TANGO predicted APRs per sequence ( 2 . 46 , Table 3 ) and the average APR length ( 9 . 13±3 . 90 ) is considerably larger in IDP536 compared to F495 ( 0 . 91 and 7 . 47±1 . 99 , respectively ) . However , TANGO predicted APR differences between IDP536 and F495 are offset by longer sequence lengths in IDP536 ( 763±562 ) , which results in a similar average proportion of APR residues ( 7 . 8±4 . 1% for F495 and 6 . 9±5 . 4% for IDP536 , Table 3 ) and similar average TANGO aggregation propensities ( Table 2 ) . Longer APR lengths may result from lower sequence complexity among IDPs [33] , [37] . For WALTZ , the average aggregation propensity ( 1 . 74±1 . 60 ) and average proportion of APR residues ( 4 . 7±2 . 7% ) for IDP536 is lower than that for F495 ( 2 . 52±2 . 70 and 7 . 7±3 . 9% , respectively ) , but the average APR lengths ( 6 . 09±0 . 54 for IDP536 and 6 . 25±1 . 15 for F495 ) remain similar ( Table 3 ) . To assess if IDP536 and F495 actually have similar or different average aggregation propensities , the total proportion of ordered and disordered residues is needed for both datasets . This data is available for IDP536 , but it is not for F495 . Therefore , comparisons among ordered and disordered regions of protein sequences , instead of full length sequences from folded and intrinsically disordered proteins , are needed to better understand the relationship between aggregation and structural disorder ( see section entitled APRs have fewer disordered than ordered residues ) . An analysis of APR flanking residues among sequence datasets indicates that APRs are , on average , flanked more often by gate-keeper residues ( Asp , Glu , Lys , Arg , or Pro ) in F495 , N10000 , SF49500 and IDP536 datasets than in R10000 ( Table 4 ) . The average number of gate-keeper residues flanking TANGO predicted APRs in F495 is 2 . 6±1 . 2 . Comparable values are observed for N10000 ( 2 . 4±1 . 1 ) , SF495000 ( 2 . 5±1 . 1 ) and IDP536 ( 2 . 4±1 . 1 ) . However for R10000 , the corresponding value is lower at 2 . 2±1 . 1 . For WALTZ predicted APRs , the average number of flanking gate-keeper residues in F495 is 2 . 0±1 . 1 which is comparable to averages for N10000 ( 1 . 9±1 . 1 ) , SF49500 ( 1 . 9±1 . 1 ) , and IDP536 ( 2 . 0±1 . 2 ) . Gate-keeper residues flank WALTZ predicted APRs in R10000 less often at 1 . 7±1 . 1 ( Table 4 ) . Comparing the incidence of flanking gate-keeper residues in TANGO and WALTZ predicted APRs , gate-keeper residues flank WALTZ predicted APRs with lower frequencies ( ∼20–25% lower ) than TANGO predicted APRs ( Table 4 ) . Charged gate-keeper residues oppose aggregation by keeping APRs solvated when they become exposed due to local flexibility or protein destabilization . However , WALTZ predicted APRs are more likely to contain charged or polar residues than TANGO predicted APRs due to the parameterization of WALTZ . As a result , the need to gate-keep WALTZ predicted APRs under normal conditions may be lower . Gate-keeper residues for all SCOP classes show similar trends as for F495 , except for the α/β class which has slightly more gate-keeper residues ( 2 . 8±1 . 1 for TANGO predicted APRs and 2 . 2±1 . 1 for WALTZ predicted APRs , Table 4 ) . Taken together , these observations suggest that the aggregation propensity of protein sequences has been minimized by increasing the number of gate-keeper residues in APR flanking regions , which is consistent with earlier studies [24] , [38] . Gate-keeper residues occur with greater frequency in the flanking regions of APRs in F495 , SF49500 and N10000 than in R10000 due to amino acid composition differences . However , N10000 , SF49500 and F495 have identical amino acid compositions . Differences in the number of flanking gate-keepers residues between N10000/SF49500 versus F495 suggests that number of gate-keeper residues flanking APRs is also determined by sequence patterning as well as amino acid composition . Selective pressure to reduce the burden of protein aggregation in biological organisms has minimized the aggregation propensities of protein sequences by directing changes to amino acid composition and patterning during protein evolution . Results presented in this section indicate that changes in sequence patterning via scrambling and randomization of protein sequences increases the TANGO predicted aggregation propensity more than do changes in amino acid composition . This is reflected in a greater difference in average aggregation propensity between F495/IDP536 and N10000/SF49500 than between SF49500/N10000 and R10000 ( Tables 2 and 3 ) for TANGO , but not for WALTZ predicted APRs . To support the finding that changes to patterning impacts sequence aggregation propensity more than changes to amino acid composition , the F495 dataset was divided into two datasets ( F1 and F2 ) that retain their natural protein sequence patterning but have significantly different amino acid compositions ( see Methods ) . The average TANGO aggregation propensities for F1 ( 3 . 66±3 . 16 ) and F2 ( 3 . 21±2 . 77 ) , and average WALTZ aggregation propensities for F1 ( 2 . 84±2 . 67 ) and F2 ( 2 . 19±2 . 51 ) , are both similar to their respective values for F495 ( TANGO 3 . 41±2 . 98; WALTZ 2 . 52±2 . 70 ) , ( Table 2 ) . Furthermore , in the case of TANGO predicted APRs , the difference in average aggregation propensity between F1 and F2 is considerably less than it is between F495 and N10000/SF49500 . Thus , changes to sequence patterning ( F495 versus N10000 ) produce greater differences in average aggregation propensity than changes to amino composition do ( F1 versus F2 ) , at least for TANGO predicted APRs . Overall , the use of three different randomly generated sequence datasets , R10000 , SF49500 and N10000 , and three natural protein sequence datasets , F1 , F2 , F495 , has allowed us separate the impact evolutionary selective pressure has had on amino acid composition and sequence patterning of protein sequences . A lack of APR conservation among homologous sequences is expected if protein evolution disfavors the tendency for proteins to aggregate [27] , [30] . Here , the conservation of APRs among homologues of 9 proteins , selected from F495 , was studied at two sequence identity cut-off values ( see Methods ) . These 9 proteins contain ≥3 TANGO predicted APRs and ≥3 WALTZ predicted APRs or ≥3 Amylsegs , indicating these proteins have a high propensity to aggregate under suitable conditions . Table 5 shows APR and Amylseg conservation ( Equation 11 ) among homologues of these 9 proteins at 80% and 50% sequence identity . Note these analyses did not include remote homologues ( <30% sequence identity ) . The term ‘distant homologues’ used below refers to homologues that have at least 50% sequence identity with a parent sequence and the term ‘close homologues’ refers to homologues that have at least 80% sequence identity with a parent sequence . All homologues have the same fold as the parent protein . For the majority of these proteins , APR conservation levels are slightly lower than average sequence identity among close homologues . This observation is in agreement with that of Gromiha and coworkers [32] on protein sequence families containing highly homologous thermophilic and mesophilic proteins . However , when the sequence identity cut-off among homologues was dropped from 80% to 50% , the corresponding decrease in percent APR conservation is smaller . In other words , APRs are more conserved than average sequence identity among distant homologues that have at least 50% sequence identity to a parent sequence . Furthermore , amyloid-like fibril forming peptide segments ( Amylsegs ) in PDB entries 1KCQ ( human gelsolin domain 2 ) , 2D4F ( human β-microglobulin ) , and 2VB1 ( hen egg-white lysozyme ) , are highly conserved among their homologues at both sequence identities ( 50% and 80% ) . Overall , the data in Table 5 indicates that among close homologues , APRs are slightly less conserved than average sequence identity . A lack of APR conservation among closely related homologues suggests that evolution is continuing to disrupt aggregation prone sequence regions . On the other hand , among distantly related homologues , APRs are more conserved than average sequence identity , indicating that some of these regions may play a role in maintaining protein folds and activity . Sequences from IDPs contain both ordered and disordered regions , which facilitates an investigation into the structural ordering of residues located in predicted APRs . Of the 232 , 321 residues in IDP536 sequences , 60 , 638 ( 26 . 1% ) fall into regions annotated as intrinsically disordered by DisProt [33] , the source database for IDP536 . Amino acid residues not annotated as disordered regions were assumed to be ordered in this analysis . Therefore of the 232 , 321 residues in IDP536 , 171 , 683 ( 73 . 9% ) were assumed to be ordered . After predicting APRs in IDP536 using TANGO , it was found that 12 , 066 ( 5 . 2% ) of the 232 , 321 IDP536 residues fall within one of 1 , 316 predicted APRs . Of the 12 , 066 APR residues , 1 , 327 ( 11% of the 12 , 066 APR residues ) residues are disordered leaving the remaining 10 , 739 ( 89% of 12 , 066 ) residues as ordered . Using this data , a contingency table was prepared to categorize IDP536 residues as ordered and not part of an APR , or ordered and part of an APR ( likewise for disordered residues , see Table 6 ) . Performing a χ2 analysis on the contingency table rejects the null hypothesis ( p<0 . 001 ) that the structural classification of an amino acid residue in IDP536 , as ordered or disordered , is independent from the likelihood that the residue is part of an APR . Furthermore , the contingency table produces an odds ratio of 2 . 98∶1 indicating that an ordered residue is three times more likely to be part of an APR than a disordered one ( see the tutorial by McHugh for more information on contingency tables and odds ratios [39] ) . WALTZ predicted APRs have comparable results ( odds ratio 2 . 02∶1 , see Table 6 ) . Thus , these observations are independent of the APR prediction tool used . Similar to the results reported above , Schymkowitz and coworkers [21] have reported that ordered proteins have three times the number of APRs that IDPs do . Here , we report that APRs are three times more likely to contain ordered versus disordered residues which provides direct evidence that APRs are located in order forming regions . At a fundamental level , APRs are promoting local structural order in all conditions , irrespective of whether conditions stabilize or destabilize proteins . The observation that APRs are more likely to contain ordered versus disordered residues suggests the need for an examination of APRs within globular protein structure . To perform the analysis , protein sequences from the F495 dataset , for which atomic coordinates are available , were used to investigate the secondary structure , solvent exposure , and relative solvent isolatedness of predicted APRs . Within F495 sequences regions with atomic coordinates , TANGO predicted 409 APRs , WALTZ predicted 516 APRs , and 19 Amylsegs of length ≥6 were found . Residues in these APRs occupy all three secondary conformational states ( helix , coil , and strand ) as measured by binning STRIDE [40] output ( see Table S2 in Supplementary material ) . Overall , Amylsegs , TANGO , and WALTZ predicted APRs exist primarily in β-strands . This observation differs from the work of Doig and coworkers , who reported that amyloidogenic regions were often located in α-helical secondary structure [41] . Figure 2 compares the percent solvent exposure of TANGO and WALTZ predicted APRs and Amylsegs in F495 structures ( Equations 2 , 3 and 5 ) . The average percent solvent exposure for TANGO predicted APRs is 16 . 9±12 . 2% ( range 0–73% ) and WALTZ predicted APRs is 23 . 3±14 . 9% ( range 0–66% ) . On average , Amylsegs in F495 are more solvent exposed than TANGO or WALTZ predicted APRs ( average percent solvent exposure 32 . 1±18 . 5% and range 1–62% ) . However , Amylsegs are longer than TANGO and WALTZ predicted APRs , with an average length of 11 . 5±10 . 4 residues . In comparison , the average APR lengths are 7 . 5±2 . 0 for TANGO predicted and 6 . 1±0 . 7 for WALTZ predicted APRs . Because proteins in F495 are small globular monomers with larger surface area to volume ratios , longer sequence segments are more likely to contain solvent exposed portions than smaller segments do . The low average percent solvent exposure of TANGO and WALTZ predicted APRs implies that these sequence regions are buried in their protein structures . Similar observations have been previously reported [22] , [41] . Predicted APRs are expected to be buried in protein cores since sequence hydrophobicity is a differentiating physico-chemical property for APR prediction programs . Therefore , it is of interest to know whether predicted APRs are buried in protein cores beyond what is expected from the hydrophobicity of their constituent residues . To investigate this question , a measure called Burial Preference ( BurPref ) was devised ( see Methods ) . The BurPref of an APR is the ratio of the observed APR solvent exposure to the expected APR solvent exposure calculated from the average surface areas of its constituent amino acid residues in F495 ( Equations 2–4 and 6 ) . If the BurPref of an APR is less than one , then the APR is more buried than expected and vice versa . The average BurPref for TANGO predicted APRs in F495 is 0 . 82±0 . 56 ( range 0–4 . 15 ) and for WALTZ predicted APRs is 0 . 89±0 . 55 ( range 0–3 . 69 ) . This indicates that TANGO and WALTZ predicted APRs are , on average , more buried in the protein cores of F495 than expected . On the other hand , the average BurPref for the 19 Amylsegs is 0 . 97±0 . 47 ( range 0 . 05–1 . 65 ) , which indicates that these segments are not preferentially buried . BurPref values for TANGO and WALTZ predicted APRs suggests that sequence hydrophobicity alone cannot explain their degree of burial . Other factors are also important such as the role an APR has in providing protein stability or function . Overall , BurPref values for TANGO/WALTZ predicted APRs support the view [41] that the risk associated with aggregation prone sequence regions is minimized by their burial within protein structures . As a consequence , APR initiated aggregation may be limited to conditions in which the native structure is destabilized or in which APRs become exposed during co-translation or due to local protein flexibility . Buried APRs can form multiple interactions within their protein structure and contribute towards stabilizing native folds . To quantify the contribution an APR makes towards the stability of a protein , solvent isolatedness ( Iso ) [42] was computed . Solvent isolatedness of a protein segment ( such as an APR ) measures the fraction ( 0–1 ) of its surface buried by the rest of the protein structure ( Equations 3 , 7 , and 8 ) . A large solvent isolatedness value for a protein segment indicates that most of its surface is interacting with the rest of protein . On the other hand , a small solvent isolatedness value for a protein segment indicates that most of its surface is solvent exposed and thus contributes few stabilizing protein interactions . The average solvent isolatedness ( Iso ) values for TANGO predicted APRs ( 0 . 83±0 . 12; range 0 . 27–1 . 0 ) , WALTZ predicted APRs ( 0 . 77±0 . 15; range 0 . 34–1 . 0 ) , and Amylsegs ( 0 . 68±0 . 19; range 0 . 38–0 . 99 ) , all indicate that these sequence regions have multiple interactions within their parent protein structures and are thus mostly solvent protected . To assess if APR segments are more solvent isolated than expected , relative solvent isolatedness ( RIso ) was computed ( Equations 7 , 8 and 10 ) . This quantity provides a measure of how much more an APR contributes to the stability of a protein than expected , based on the average solvent isolatedness of equal length segments from the same protein [42] . The relative solvent isolatedness values for TANGO and WALTZ predicted APRs and Amylsegs are greater than one which indicates that they are more solvent isolated than expected ( TANGO average RIso 1 . 33±0 . 18 , range 0 . 47–1 . 73; WALTZ average RIso 1 . 20±0 . 22 , range 0 . 56–1 . 67; Amylsegs average 1 . 19±0 . 29 , range 0 . 66–1 . 64 ) , ( Figure 3 ( a ) ) . Therefore , TANGO and WALTZ predicted APRs , as well as Amylsegs , contribute towards protein stability more than expected . To assess the significance of APR solvent isolatedness values , Z-scores of solvent isolatedness were computed for predicted APRs and Amylsegs ( Equations 7 , 8 and 9 ) . Although average Z-scores for WALTZ predicted APRs ( 0 . 88±0 . 95; range −1 . 9–2 . 8 ) and Amylsegs ( 0 . 78±1 . 14; range −1 . 2–2 . 4 ) are below one , the average Z-score for TANGO predicted APRs is 1 . 44±0 . 76 ( range −1 . 9–3 . 3 ) . Therefore , the contributions made by TANGO predicted APRs towards stabilizing the protein , in which they exist , are more than one sigma greater than the average contribution made by equal length segments ( Figure 3 ( b ) ) . Others have previously proposed that APRs make stabilizing interactions in native folds [25] , [43] . However , this is the first report that TANGO/WALTZ predicted APRs and Amylsegs make more stabilizing interactions in native folds than , on average , those made by equal-length segments from the same protein . Evidence that APRs stabilize native protein folds calls for an examination into whether functional sites are located within APRs or are structurally proximal to them . In this report , co-localization of catalytic sites and TANGO or WALTZ predicted APRs in crystal structures of enzymes is investigated using the Cata dataset which contains 961 catalytic residues from 314 non-homologous protein chains ( 299 enzymes ) . Catalytic residues form a subset of active site residues in enzymes . Thus , most enzymes contain few catalytic residues and the likelihood that these residues are located either within APRs , or near APRs , by random chance is low . Incidences of catalytic residues within TANGO/WALTZ predicted APRs and their flanking regions were estimated using equation 12 ( see Methods ) . These estimated numbers were compared with the observed incidence of the catalytic residues within TANGO/WALTZ predicted APRs and their flanking regions . Ninety nine of the 961 catalytic residues ( 10 . 3% ) were estimated to fall within TANGO predicted APRs and flanking regions and 57 of them ( 5 . 9% ) were observed within these regions . Similarly , 103 out of the 961 catalytic residues ( 10 . 7% ) were estimated and 69 ( 7 . 2% ) were observed to fall within WALTZ predicted APRs and flanking regions . Therefore , the observed incidence of the catalytic residues within APRs and their flanking regions is lower than the estimates and it can be concluded that the catalytic residues are not usually found within these regions . This is expected given that most catalytic residues are charged or polar . On the other hand , catalytic residues do tend to be in structural contact with APRs ( structural contacts are inferred when a catalytic residue heavy atom is within 4 . 5 Å from an APR residue heavy atom , see Methods ) . Of the 961 catalytic residues , 373 ( 38 . 8% ) are in structural contact with at least one neighboring TANGO predicted APR residue . Similarly , 310 ( 32 . 3% ) catalytic residues are in contact with at least one neighboring residue within WALTZ predicted APRs . Although , it has previously been observed that APRs and amyloidogenic regions can be located in protein-protein interfaces [22] , [28] , including antigen-antibody interfaces [23] , this is the first report of structural proximity between APRs and catalytic residues to our knowledge . To assess the significance of the observed structural proximity between APRs and catalytic residues , statistical simulations were performed by generating one million catalytic decoy lists . Each list contained the residue coordinates of 961 randomly chosen decoy catalytic residues from the atomic coordinates of protein chains in the Cata dataset . Randomly chosen decoy catalytic residues were selected for each true catalytic residue in Cata and were limited to any residue within the same protein structure as the true catalytic residue . Thus , if there are X number of true catalytic residues from protein Y in the Cata dataset , all decoy lists contained X number of decoy catalytic residues from the same protein Y . Using decoy catalytic lists enabled us to calculate an expected number of residues in structural contact with predicted APRs . This expected number was compared to the observed number of true catalytic residues in structural contact with predicted APRs . Figure 4 ( a ) shows the distribution of the number of decoy catalytic residues which are in structural contact with TANGO predicted APR residues for all one million lists . The average number of decoy catalytic residues in contact with TANGO predicted APRs is 254±13 . This yields a Z-score of 9 . 3 for the 373 of 961 true catalytic residues in the Cata dataset that are in structural contact with TANGO predicted APRs and suggests the number of true catalytic residues in contact with TANGO predicted APRs is highly significant . These calculations were repeated by varying the distance cut-off for inferring structural contacts ( 3 . 5 Å and 6 . 0 Å ) . The observed number of true catalytic residues in contact with TANGO predicted APRs is 278 and 461 at 3 . 5 Å and 6 . 0 Å , respectively . Statistical simulations to compute the expected number of randomly chosen decoy catalytic residues in contact with TANGO predicted APRs yield Z-scores of 10 . 1 at the 3 . 5 Å cut-off distance and 11 . 7 at the 6 . 0 Å cut-off distance for the number of true catalytic residues in contact with TANGO predicted APRs . Catalytic residues in the Cata dataset are also in close structural proximity to WALTZ predicted APRs , significantly more often than expected . There are 224 , 310 and 405 catalytic residues in contact with WALTZ predicted APRs at cut-off distances of 3 . 5 Å ( Z-score , 5 . 8 ) , 4 . 5 Å ( Z-score , 5 . 7 ) and 6 . 0 Å ( Z-score , 8 . 5 ) respectively . To further probe our observation of catalytic residues in contact with APRs , an additional condition on the solvent exposure of randomly selected residues as decoy catalytic residues was imposed . Decoy catalytic residues were required to have a solvent exposure that was similar to the solvent exposure of their corresponding true catalytic residue ( ASA value of each decoy catalytic residue must be within ±10% of the ASA value for the corresponding true catalytic residue ) . At the structural contact cut-off distance of 4 . 5 Å , Z-scores for catalytic residues in contact with TANGO and WALTZ predicted APRs are 5 . 9 and 3 . 1 respectively when ±10% ASA condition was imposed . As such , Z-scores decreased but remain statistically significant . Z-scores decreased when the ASA condition was imposed because the probability that a decoy catalytic residue is in contact with an APR increases when decoy catalytic residues are limited to the same solvent exposure as their true catalytic residues . Computing expected values for the number of residues in contact with APRs , using both ASA limitations and multiple distance cut-offs , has supported our finding that catalytic residues are in close structural contact with APRs significantly more often than expected by random chance . To visualize the co-localization of APRs and catalytic residues in enzyme structures , an example from the Cata dataset is shown in Figure 4 ( b ) ) . Cholesterol Oxidase from B . Sterolicum ( PDB entry: 1I19 , UniProt entry: Q7SID ) is a 561 residue monomeric enzyme with covalently bound FAD that catalyzes oxidation and isomerization of steroids [44] . This enzyme contains three catalytic residues , namely , Glu 311 , Glu 475 and Arg 477 , and two TANGO identified APRs , 229-LTAVVW-234 and 511-VAIWLNVL-518 . Two catalytic residues , Glu 475 and Arg 477 , make several close contacts with the second APR ( Figure 4 ( b ) ) , which lies in the substrate binding domain [44] . Considering the large structural size of Cholesterol Oxidase , and the fact that is has only three catalytic residues and two TANGO predicted APRs , which are short in length , it is surprising to find its catalytic residues in structural contact with its APRs . Are there examples of catalytic residues making structural contact with experimentally validated amyloid-fibril forming peptide segments , Amylsegs ? To answer this question , the AmylSegs dataset was searched for peptide segments from enzymes . Amylsegs contain amyloid-fibril forming peptides derived from nine different enzymes . Six of these nine enzymes have been annotated in UniProtKB for catalytic residues and have at least one crystal structure deposited in the PDB ( Table 7 ) . Of these six , three enzymes ( pancreatic ribonuclease A , hen egg white lysozyme and human lysozyme ) have a catalytic residue that is in contact with at least one Amylseg residue . The criterion used here to identify a structural contact is the same as in analyses on the Cata dataset ( distance cut-off of 4 . 5 Å for a pair of heavy atoms ) . Figure 5 shows an example of catalytic residues from human lysozyme , Glu 35 and Asp 53 , which are in contact with its Amylseg regions . There are three peptides in the Amylsegs dataset from human lysozyme that have been experimentally shown to form amyloid fibrils . These are 5-RCELARTLKR-14 , 25-LANWMCLAKW-34 and 56-IFQINS-61 [45] . Catalytic residue , Glu 35 , lies immediately after the second peptide and makes contact with residues 30-CLAKW-34 as well as contact with residues 56-IFQ-58 from the third peptide . Catalytic residue Asp 53 also contacts the third peptide . Indeed , catalytic residues are making structural contact with experimentally validated amyloid-fibril forming peptide segments . Protein functional sites require optimal combinations of flexibility and stability to fulfill their biological purposes . Therefore , it is logical for catalytic residues , which require consistent and specific orientations , to be in contact with regions that promote local structural order and form stabilizing interactions . While this report has focused on catalytic residues , an interesting example of an Amylseg , 182-SFNNGDCFILD-192 , containing a Ca2+ binding residue at D187 in human gelsolin has also been identified . The D187N mutation , which disrupts the metal binding site , leads to protein instability and amyloidosis in patients with a disease called familial amyloidosis-Finnish type [46] . Co-localization of metal catalyzed oxidation ( MCO ) sites and APRs have also been observed in the structures of therapeutic proteins where metal-ion leachates contribute towards drug product degradation [47] . The results presented above show that protein sequences have evolved by optimizing their risk of aggregation for cellular environments by both minimizing aggregation prone regions and conserving those that are important for folding and function . Outside the cell , protein aggregation is commonly encountered in the laboratory and is a major hurdle in successful development of protein based biotechnology products , such as biotherapeutics . For these applications , the aggregation propensity may need to be further reduced in order to enhance protein yields from cell cultures and to improve protein solubility , especially at high concentrations . Disruption of APRs via site directed mutagenesis is an attractive protein engineering strategy to improve protein solubility [13] , [14] , [23] , [48]–[50] . Alternatively , ‘supercharging’ functional proteins with very high electrostatic surface charge has also been shown to improve protein solubility beyond levels normally observed for natural proteins [51]–[53] . Because APRs can also form part of HLA-DR binding T-cell immune epitopes [54] , [55] , disrupting APRs potentially leads to a lower risk of immunogenic reactions in patients receiving biotherapeutic drugs . Insights gained from this report caution that the contributions made by candidate APRs , targeted for disruption , towards protein stability and function should be considered when identifying sites that are suitable for rational mutagenesis . Disruption of APRs , without knowledge of their contributions , can lead to undesirable consequences , such as protein destabilization and/or loss-of-function . This work also complements our efforts to distinguish between ‘active’ and ‘inactive’ APRs in proteins [56] , [57] . A broader implication of this research is that a general strategy for identifying mutation sites for improving solubility of a candidate protein can be proposed . This strategy is presented in Figure 6 and the major steps are described below . If a protein of biotechnological interest aggregates at higher than a desired level , the following information is needed to employ the strategy: protein sequence , three dimensional structure , homologues , potential cross-β motif forming APRs and functional sites . APR prediction programs often identify several potential APRs in the sequence of a protein . For each APR , its contribution towards protein stability should be evaluated . This can be done by computing solvent isolatedness for the APR and equal length segments from the same protein ( Equation 8 ) . If the APR has a high solvent isolatedness ( low solvent exposure , buried in protein core ) then it should not be a target for disruption . If the APR has a low solvent isolatedness ( high solvent exposure , located at or near protein surface ) then it is expected to make a smaller contribution to protein stability and can be marked for disruption depending upon the outcomes of the following tests . The protein structural region around the APR should be examined for contacting functional residues and for sequence conservation among homologues . If the APR contains residues that are in structural contact with functional residues and the APR is more conserved than average sequence identity among homologues , then it should not be targeted for disruption . If the APR is not in contact with functional residues and is less conserved than the average sequence identity among its homologues , then it is a priority target for disruption . If the APR is not structurally proximal to functional residues , but is more conserved than average sequence identity among its homologues , it can still be targeted for disruption after verifying that the conservation is not due other structure-function purposes such as allostery . If the APR is structurally proximal to functional residues , but is less conserved than average sequence identity among the homologues , it can still be targeted for disruption if molecular modeling can offer clues into potential sites and mutations that can be safely substituted . In this case , care should be taken to avoid disturbing the conformations of functional and contacting residues . This can be done by choosing a residue within the APR which is not in direct contact with functional residues , but whose mutation disrupts the APR .
TANGO [11] and WALTZ [12] were used to predict APRs in all sequence datasets . Both programs have been extensively validated using independent testing sets and found to be highly accurate . The following options were used as input parameters for both programs: Temperature , 298K; pH , 7 . 0; Ionic Strength , 150 mM; Concentration 1 mM; TANGO/WALTZ aggregation , ≥10%; Minimum window size , 6; Flanking residues , 3 . The outputs from TANGO and WALTZ yield data on the total sequence aggregation score and sequence length along with position , length , sequence and flanking residues for each predicted APR . The total aggregation score for each sequence was normalized by the length of the sequence to obtain its aggregation propensity . The proportion of APR residues ( APRprop ( % ) ) in a sequence was also computed as follows: ( 1 ) APRlen ( i , j ) is the length of the ith APR in the jth sequence , nAPR ( j ) is the number of APRs in the jth sequence , seqlen ( j ) is the number of residues in the jth sequence . Averages and standard deviations were computed for both aggregation propensities and proportional APR residues . These results are reported in Tables 2 and 3 . Flanking residue positions that precede ( PB−1 , PB−2 , PB−3 ) and succeed ( PE+1 , PE+2 , PE+3 ) each APR were searched for the presence of gate-keeper residues ( Asp , Glu , Lys , Arg and Pro ) [24] , [25] and their frequencies were computed at each of these positions . These results are presented in Table 4 . Atomic coordinates from F495 proteins were submitted to STRIDE [40] to obtain secondary structure information and solvent accessible surface areas ( ASA ) for all residues within their three dimensional context . Average ASA values for each of the twenty amino acids in the F495 dataset were obtained by summing the ASAs of amino acid , k , in all proteins and dividing the sum by the number of k amino acids in F495 , where k runs from 1 to 20 . ( 2 ) AvASAResk is the average ASA for amino acid of type k . ASAReskj is the ASA of each amino acid residue of type k in protein j and NResk is the number of amino acid residues of type k in F495 dataset . ASA values for individual residues of APRs were summed to obtain observed solvent accessible surface areas ( SASAobs ) for each predicted APR . ( 3 ) APRij is the ith APR in the jth protein from F495 . SASAobsAPRij is the observed SASA of APRij . ASARes ( k ) APRij is the ASA of an individual residue , k , in APRij . The summation runs over the length of APRij , APRlen ( i , j ) . Expected ASA values for each APR ( SASAexp ) were computed by summing the average ASA values from F495 for each of the constituent amino acid residues . ( 4 ) SASAexpAPRij is the expected SASA of APRij . AvASARes ( k ) APRij is the average ASA of residue type k in APRij , computed using Equation 2 . The total surface area ( TotSA ) for an APR outside of its three-dimensional context was computed by submitting atomic coordinates , only from APR segments , to STRIDE . Note that in both calculations , APRs have identical conformations . TotSA and SASA values for each APR were used to compute percent solvent accessibility ( SolvAcc ) of the APR , burial preference ( BurPref ) for an APR , and solvent isolatedness ( Iso ) of an APR from solvent [42] as follows: ( 5 ) ( 6 ) ( 7 ) ( 8 ) SolvAccAPRij is the percent solvent accessibility of APRij . SASAobsAPRij is the solvent accessible surface area of APRij within its three-dimensional context . TotSAAPRij is the total surface area of the APR outside of the three-dimensional context of protein j . SASAexpAPRij is the expected solvent accessible surface area of APRij computed from average ASA values for its constituent amino acids in F495 . BurPrefAPRij is a ratio of the observed to expected solvent accessible surface area for APRij . It indicates the preferential burial of APRij in protein cores . If BurPrefAPRij is below one , APRij is more buried than expected from the average burial of its constituent residues . If BurPrefAPRij is above one , APRij is more solvent exposed than expected from the average solvent exposure of its constituent residues . ProtBurSAAPRij is the surface area of APRij that is buried by the rest of the protein j . IsoAPRij values can be interpreted as the contribution APRij makes towards the stability of protein j . To evaluate the significance of this contribution , IsoAPRij values were also computed for all segments the equal length as APR , i in protein , j . Each segment was obtained by sliding a window the equal length as APR , i over the structure of protein , j one residue at a time . Average ( <Isoij> ) and standard deviation ( σIsoij ) values for segments within protein j were used to compute Z-scores ( Z-score ( IsoAPRij ) ) and relative values ( RIsoAPRij ) for solvent isolatedness of APRij using the following equations: ( 9 ) ( 10 ) Nine selected monomeric proteins from F495 were used for sequence conservation analyses . These 9 proteins contain ≥3 TANGO predicted APRs and ≥3 Waltz predicted APRs or ≥3 Amylsegs , indicating these proteins have a high propensity to aggregate under suitable conditions . PDB entries for these proteins are 1FUK ( C-terminal domain of yeast initiation factor 4A ) , 1JEO ( Hypothetical protein MJ1247 from Methanococcus jannaschii ) , 1KCQ ( Human Gelsolin Domain 2 ) , 1OW1 ( SPOC domain of human transcriptional factor SHARP ) , 1SK7 ( Hypothetical protein pa-HO from Pseudomonas aeruginosa ) , 1Z77 ( Transcriptional regulator ( tetR family ) from Thermotoga maritima ) , 2D4F ( Human β-microglobulin ) , 2VB1 ( Hen Egg White Lysozyme ) and 3NR5 ( Human RNA polymerase III transcription repressor Maf1 ) . Note that Hen egg-white lysozyme , human β-microglobulin and human gelsolin are well studied amyloidogenic proteins [61]–[63] . Sequence conservation analyses for the above mentioned proteins were performed at two arbitrarily chosen levels of sequence identity , 80% and 50% . The procedure for selecting homologues at the 80% level is described below . Sequences of the nine proteins ( query sequences ) were searched for homologues in the UniProtKB database ( www . uniprot . org ) [64] , [65] using blastp and all default options . For each query , hit sequences with ≥80% sequence identity were selected , provided that homologous regions of hit sequences covered the query sequence completely . Since hits to query sequences were sometimes longer than the length of the query sequence , only portions of hit sequences that aligned with the query sequence were taken for conservation analysis . All the retrieved homologous sequences were re-aligned using ClustalW [59] . For each query sequence , the ClustalW input file included the sequence from the PDB file as the first sequence . Any sequence with a ClustalW alignment score of <80 to the first sequence was deleted from the alignment . Sequences with alignment scores of 100 to the first sequence were also removed . All the above steps were repeated to obtain homologous sequences with ≥50% sequence identity to selected PDB files . An APR was labeled as ‘conserved’ between two homologous sequences , if the APR has the same sequence in both homologues . Percent APR conservation in a multiple sequence alignment was computed using the following formula: ( 11 ) PAPRconserved is the proportion of conserved APRs in an alignment , nAPRtotal is the total number of APRs for all the sequences in the alignment and nAPRuniq is the number of unique APRs ( non-identical sequence ) over all sequences in the alignment . These calculations were performed for both TANGO and WALTZ predicted APRs . Analogous calculations were also performed for all peptide sequences from Amylsegs that were detected in the 9 proteins and their homologues . Number of the catalytic residues from the Cata dataset that fall within the TANGO/WALTZ predicted APRs and their flanking regions were estimated using the following equation: ( 12 ) NCata-APRs is the estimated number of catalytic residues that fall within TANGO/WALTZ predicted APRs and their flanking regions . Three residues preceding and three residues succeeding an APR are considered as its flanking regions . NCata is the number of catalytic residues in the Cata dataset . This number is 961 . NAPRs is the number of residues in APRs and their flanking regions predicted using TANGO and WALTZ in the sequences of the 299 enzymes ( 314 Chains ) in the Cata dataset . For 823 TANGO predicted APRs and their flanking regions , NAPRs is 11 , 414 ( 6498 residues in TANGO APRs plus 4916 residues in the flanking regions ) . NAPRs is 11 , 856 for 982 WALTZ predicted APRs and their flanking regions ( 5988 residues in the APRs and 5868 residues in the flanking regions ) . NTot is 110 , 334 , the total number of residues in the sequences of the 299 enzymes . TANGO and WALTZ predicted APRs were also mapped onto protein structures from the Cata dataset to search for structural contacts made by catalytic residues to residues in predicted APRs . A catalytic residue is considered to be in structural contact with an APR , if at least one of its heavy atoms is within 4 . 5 Å from a heavy atom in any residue that falls within an APR . The choice of a 4 . 5 Å cut-off is arbitrary but was used here because it is common in the literature [66]–[68] . Catalytic residues in structural contact with APRs residues were counted for both TANGO and WALTZ predicted APRs . To assess the significance of the observed structural proximity between APRs and catalytic residues , statistical simulations were performed by generating one million decoy catalytic lists . Each list contained the residue coordinates of 961 randomly chosen decoy catalytic residues from the atomic coordinates of protein chains in the Cata dataset . Randomly chosen decoy catalytic residues were selected for each true catalytic residue in Cata and were limited to any residue within the same protein structure as the true catalytic residue . For each of the 1 , 000 , 000 randomly generated lists , the number of residues making structural contact with APR residues ( APR contacting ) was computed again in the same way as for the Cata dataset . The number of APR contacting residues was counted for each random list to generate a distribution of expected APR contacting residues . This distribution was also used to compute Z-scores for the incidence of APR contacting residues in the Cata dataset in the same way as the Z-score for solvent isolatedness was calculated ( Eq . 9 ) . The analogous calculations were also performed using the contact distance cut-off values of 3 . 5 Å and 6 . 0 Å . To further probe our observation of catalytic residues in contact with APRs , a restriction on the solvent exposure of randomly selected residues as catalytic decoys was imposed . Decoy catalytic residues were required to have a solvent exposure that was similar to their corresponding true catalytic residue ( ASA value of each decoy must be within ±10% of ASA of true catalytic residue ) . For each of the 1 , 000 , 000 randomly generated lists , the number of residues making structural contact with APR residues ( APR contacting ) was computed again in the same way as for the Cata dataset . | Biotechnology requires the large-scale expression , yield , and storage of recombinant proteins . Each step in protein production has the potential to cause aggregation as proteins , not evolved to exist outside the cell , endure the various steps involved in commercial manufacturing processes . Mechanistic studies into protein aggregation have revealed that certain sequence regions contribute more to the aggregation propensity of a protein than other sequence regions do . Efforts to disrupt these regions have thus far indicated that rational sequence engineering is a useful technique to reduce the aggregation of biotechnologically relevant proteins . To improve our ability to rationally engineer proteins with enhanced expression , solubility , and shelf-life we conducted extensive analyses of aggregation prone regions ( APRs ) within protein sequences to characterize the various roles these regions play in proteins . Findings from this work indicate that protein sequences have evolved by minimizing their aggregation propensities . However , we also found that many APRs are conserved in protein families and are essential to maintain protein stability and function . Therefore , the contributions that APRs , targeted for disruption , make towards protein stability and function should be carefully evaluated when improving protein solubility via rational design . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
]
| []
| 2013 | On the Role of Aggregation Prone Regions in Protein Evolution, Stability, and Enzymatic Catalysis: Insights from Diverse Analyses |
An ability to sense pathogens by a number of specialized cell types including the dendritic cells plays a central role in host's defenses . Activation of these cells through the stimulation of the pathogen-recognition receptors induces the production of a number of cytokines including Type I interferons ( IFNs ) that mediate the diverse mechanisms of innate immunity . Type I IFNs interact with the Type I IFN receptor , composed of IFNAR1 and IFNAR2 chains , to mount the host defense responses . However , at the same time , Type I IFNs elicit potent anti-proliferative and pro-apoptotic effects that could be detrimental for IFN-producing cells . Here , we report that the activation of p38 kinase in response to pathogen-recognition receptors stimulation results in a series of phosphorylation events within the IFNAR1 chain of the Type I IFN receptor . This phosphorylation promotes IFNAR1 ubiquitination and accelerates the proteolytic turnover of this receptor leading to an attenuation of Type I IFN signaling and the protection of activated dendritic cells from the cytotoxic effects of autocrine or paracrine Type I IFN . In this paper we discuss a potential role of this mechanism in regulating the processes of innate immunity .
Cytokines that belong to the family of Type I interferons ( IFNs , including IFNα/β ) , play an important role in innate immunity [1] , [2] , [3] , [4] . These cytokines are produced by various cell types including dendritic cells ( DCs ) , which are equipped with a diverse set of specialized pathogen recognition receptors ( PRR ) that recognize a variety of pathogens . Activation of PRR in DCs is known to induce the production of numerous cytokines including inflammatory cytokines and IFNα/β . The latter cytokines modulate the subsequent antigen-specific adaptive immune responses and mount the overall host defenses against infection and injury ( reviewed in [5] , [6] , [7] ) . Type I IFN mediates its effects through the stimulation of the Type I IFN receptor ( which consists of IFNAR1 and IFNAR2 chains ) and the subsequent activation of the Janus kinases TYK2 and JAK1 , phosphorylation of STAT1/2 proteins , and ensuing trans-activation of a plethora of interferon-stimulated genes . The products of these genes directly suppress the spread of some pathogens ( e . g . , viruses ) and cooperate with inflammatory cytokines in DCs to promote antigen presentation linking innate and adaptive immunity ( reviewed in [8] , [9] , [10] ) . Some of these genes also elicit pronounced anti-proliferative and pro-apoptotic effects [10] . Conversely , cells that produce IFNα/β have to survive in this environment and have to be protected from autocrine/paracrine IFNα/β [11] . Accordingly , Type I IFNs play a dynamic role in homeostasis and in the function of DCs [12] . Whereas IFNα/β contribute to the maturation and activation of DCs [13] , [14] , [15] , these cytokines are also known to decrease the viability of IFNα/β-producing DCs [4] , [12] , [16] . Additional mechanisms of negative regulation are expected to prevent a hypothetical scenario where IFNα/β-stimulated maturation of DCs and subsequent production of more of IFNα/β by these DCs spirals out of control and leads to a hyperactivation of pathways induced by Type I IFN . Such hyperactivation of the IFNα/β pathways might be detrimental , not only to a population of specific IFN-producing cells , but also to the entire host because it leads to autoimmune disorders . A key role of IFNα/β in pathogenesis of such disorders , including psoriases , systemic lupus erythematosus and Type 1 diabetes mellitus , has been thoroughly documented [17] , [18] , [19] , [20] . Therefore , developing an understanding of the mechanisms , by which IFNα/β responses are kept in check , is of medical importance . All responses to IFNα/β require the presence of the Type I IFN receptor at the cell surface [21] . Upon interacting with the ligands , this receptor not only conducts the forward signaling through the JAK-STAT pathway but also undergoes rapid downregulation [22] , [23] . This downregulation is driven by the phosphorylation-dependent ubiquitination , endocytosis , and degradation of the IFNAR1 chain [24] , [25] . Ubiquitination of IFNAR1 is facilitated by the SCFβTrcp E3 ubiquitin ligase , which is recruited to the receptor upon phosphorylation on serine residues ( S535 in humans , S526 in mice ) within IFNAR1's degron [25] , [26] . Such phosphorylation can be induced by a ligand in a TYK2 activity-dependent manner [27] through the activation of protein kinase D2 [28] . This ligand-inducible pathway limits the extent of IFNα/β signaling in the cells that have already been exposed to these cytokines . We have recently discovered an alternative ( TYK2 activity-independent ) signaling pathway that leads to S535 phosphorylation and to the subsequent ubiquitination and downregulation of IFNAR1 in the absence of a ligand . As a result , IFNAR1 is being removed from the surface of those cells , which have not been yet exposed to IFNα/β [29] . Accordingly , this signaling does not require the ligand and hence can impair the ability of a naïve cell to respond to its future encounters with IFNα/β . Such a pathway can be further stimulated by the inducers of the unfolded protein response ( UPR ) . UPR is a complex of signaling pathways that are essential for protein quality control in cells [30] , [31] and can be induced by various stress stimuli such as thapsigargin ( TG ) as well as by viral infection [32] . Inducers of UPR ( including RNA-containing viruses such as vesicular stomatitis virus , VSV , and hepatitis C virus , HCV ) were shown to promote IFNAR1 degradation via a pathway that required the activation of a PKR-like ER kinase ( PERK , [33] ) . Subsequent delineation of the ligand-independent pathway revealed an important role for casein kinase 1α ( CK1α ) in the direct phosphorylation of S535/S526 within the IFNAR1 degron [34] . The ability of constitutively active CK1α to phosphorylate the IFNAR1 degron was augmented in cells treated with UPR inducers or infected with VSV via priming phosphorylation of another serine – S532 ( S523 in mouse IFNAR1 ) . The latter phosphorylation was mediated by a yet to be identified kinase that functioned downstream of PERK [35] . Here we report that the DNA-containing herpes simplex virus ( HSV ) also induced the ligand/TYK2-independent phosphorylation and downregulation of IFNAR1 . Remarkably , these effects required neither viral protein synthesis nor PERK activity but relied on the activation of the PRR signaling pathways that center around the activation of the p38 protein kinase . This kinase was required for the phosphorylation of the IFNAR1 priming site leading to an ensuing phosphorylation of the IFNAR1 degron on Ser535/526 , acceleration of IFNAR1 degradation , and attenuation of IFNα/β signaling . These events appear to be important for the protection of DCs from autocrine/paracrine Type I IFN .
RNA-containing viruses ( HCV and VSV ) can downregulate IFNAR1 in human KR-2 cells that harbor a catalytically inactive TYK2 and that are deficient in IFNα-stimulated Ser535 phosphorylation of IFNAR1 and the degradation of this receptor chain [27] , [33] . We sought to investigate whether a DNA-containing virus such as the herpes simplex virus ( HSV ) is also capable of such activity . We observed that the infection of KR-2 cells with HSV stimulates degron phosphorylation of endogenous human IFNAR1 and robustly downregulates the levels of this receptor ( Figure 1A ) . A marked decrease in cell surface levels of murine IFNAR1 in response to HSV infection was also seen in mouse embryo fibroblasts ( MEFs , Figure 1B ) . To determine whether downregulation of IFNAR1 by HSV depends on IFNAR1 degron phosphorylation , we have generated a knock-in mouse that expresses a phosphorylation-deficient IFNAR1 mutant . To this end , mouse ES cells , in which one wild type Ifnar1 allele was replaced with a mutant allele that lacks Ser526 ( a serine residue homologous to Ser535 within human IFNAR1; cells were described in [33] ) were rid of the Neo cassette . They were then used to generate knock-in mice that express the IFNAR1S526A mutant ( “SA” , Figure S1 ) . Importantly , MEFs derived from these mice were noticeably resistant to an HSV-induced downregulation of IFNAR1 levels on the cell surface ( Figure 1B ) . This result suggests that HSV downregulates IFNAR1 in a degron phosphorylation-dependent manner . Given that HSV has been known to induce the production of both IFNα and IFNβ [36] , we next sought to delineate the mechanisms by which HSV infection stimulates the phosphorylation of IFNAR1 degron . We chose to analyze endogenous IFNAR1 in human fibrosarcoma-derived cell lines harvested at the earlier time points ( 20–22 hr ) of HSV infection at low MOI ( 0 . 1 ) . Under these conditions , the total levels of IFNAR1 were not yet dramatically downregulated , enabling a better detection of Ser535 phosphorylation . HSV infection of WT-5 cells that express wild type TYK2 [37] , [38] induced a robust S535 phosphorylation of IFNAR1 ( Figure 1C ) . Comparable levels of IFNAR1 degron phosphorylation were seen in the isogenic KR-2 , which express kinase-dead TYK2 and are responsive to IFNβ [38] but not to IFNα [27] , [37] suggesting that the activity of TYK2 is not required for the effects of the virus . On the contrary , the phosphorylation of IFNAR1 Ser535 in response to treatment with recombinant IFNβ was much less evident in KR-2 cells that displayed only a rather modest STAT1 phosphorylation in response to this cytokine ( Figure 1C ) . Furthermore , in either WT-5 or KR-2 cells , infection with HSV hardly induced any STAT1 phosphorylation indicating that an increase in IFNAR1 degron phosphorylation may not rely on HSV-induced production of Type I IFN . Finally , in the isogenic fibrosarcoma U5A , cells that lack the IFNAR2 chain of the Type I IFN receptor and are insensitive to any Type I IFN [39] , infection with HSV robustly induced Ser535 phosphorylation of IFNAR1 ( Figure 1C ) . These results suggest that the effects of HSV on phosphorylation of IFNAR1 are ligand-independent . These isogenic WT-5 , KR-2 and U5A cell lines were infected with HSV for longer periods of time to determine the role of TYK2 kinase activity and endogenous Type I IFN in downregulation . A comparable decrease of IFNAR1 levels in response to infection was seen in WT-5 and KR-2 cells . Furthermore , HSV robustly decreased levels of IFNAR1 in U5A cells ( Figure 1D ) . These results indicate that neither production of the endogenous ligands nor catalytic activities of TYK2 are required for IFNAR1 downregulation in cells infected with HSV . Given the latter results we sought to determine whether , similarly to VSV , the effects of HSV infection on IFNAR1 might be mediated by the constitutively active kinase CK1α [34] whose ability to phosphorylate IFNAR1 degron is augmented by a priming phosphorylation of IFNAR1 on S532 [35] . Phosphorylation of the degron ( S535 ) in HSV-infected KR-2 cells was indeed attenuated by treating the cells with a CK1 inhibitor ( Figure 2A ) . Furthermore , HSV infection induced the priming phosphorylation of IFNAR1 on S532 ( Figures 1C and 2A–C ) and IFNAR1 ubiquitination ( Figures 2B–C ) . The IFNAR1S532A mutant lacking the priming site was noticeably more resistant to the HSV-stimulated ubiquitination and downregulation than the wild type receptor ( Figures 2C–D ) . Together , these results suggest that , similar to VSV , HSV stimulates the ligand/TYK2-independent pathway of IFNAR1 phosphorylation-dependent ubiquitination and downregulation that requires the priming phosphorylation . The experimental settings of all these experiments included infection at low doses ( MOI 0 . 1 ) that did not induce changes in the status of IFNAR1 ( data not shown ) until later periods of the infection ( 24–30 h post infection ) which were chosen for the maximal expression of viral proteins to induce UPR . Under these conditions , HCV and VSV required PERK activity to promote IFNAR1 phosphorylation and degradation [33] . Although KR-2 cells display a noticeable increase in the phosphorylation of translational regulator eIF2α ( a major PERK substrate ) in response to thapsigargin or VSV [33] , [35] , stimulation of this signaling event by HSV infection was modest at best ( Figure 2A ) . These results are consistent with previous reports that attributed low levels of eIF2α phosphorylation either to the stimulation of eIF2α de-phosphorylation by the HSV protein γ134 . 5 under conditions where PERK is activated [40] or to the suppression of PERK activation by the viral glycoprotein B [41] . Under the conditions used in our experiments in KR-2 cells to influence the phosphorylation of IFNAR1 degron , an induction of PERK phosphorylation ( indicative of its activation ) was observed in response to thapsigargin but not to infection with HSV ( Figures 3B–C ) . Both thapsigargin and HSV infection stimulated the priming phosphorylation of IFNAR1 on Ser532 in KR-2 cells . Whereas the knockdown of PERK noticeably attenuated the effects of thapsigargin , priming phosphorylation of IFNAR1 stimulated by , HSV infection was impervious to the modulations of PERK expression ( Figure 3C ) . This result suggests that PERK is dispensable for HSV-stimulated IFNAR1 phosphorylation in human cells . We further tested the requirement of PERK by using MEFs from mice harboring a conditional knockout allele of PERK ( PERKfl/fl ) where PERK is acutely excised upon transduction with a retrovirus encoding the Cre recombinase [42] . Consistent with a previous report [33] , the acute deletion of PERK prevented the downregulation of IFNAR1 upon VSV infection . However , the downregulation of IFNAR1 stimulated by HSV was not affected by the status of PERK ( Figure 3D ) . These data suggest that HSV is capable of stimulating the ligand-independent pathway in a manner that differs from viral protein synthesis-induced UPR and activation of PERK described for HCV and VSV . We next tested a possibility that a requirement for prolonged infection and ensuing HSV replication needed to observe the effects of low doses of HSV ( MOI 0 . 1 ) might be foregone if more viruses are used initially . To this end , we compared the effects of HSV ( at MOI 5 . 0 ) that was either sham treated or irradiated with UV for inactivation . The latter procedure decreased the titer of this viral preparation from 7×107 to 3 pfu . Remarkably , the treatment of KR-2 cells with a high dose of either active or inactive HSV sufficed for inducing the priming phosphorylation of IFNAR1 within 60–90 minutes ( Figure 4A ) . Furthermore , treatment with either active or inactive HSV comparably decreased the levels of IFNAR1 ( Figure 4B ) indicating that the downregulation of IFNAR1 can be stimulated by HSV in a manner that does not require virus replication . The downregulation of IFNAR1 can plausibly occur through diverse mechanisms including an increase in protein degradation and decrease in protein synthesis mediated by translational or pre-translational events ( e . g . , a decrease in mRNA levels ) . To determine the role of IFNAR1 proteolytic turnover in this process we used a standard approach of blocking the protein synthesis by treating the cells with cycloheximide . Under these conditions , the treatment of cells with inactivated HSV markedly accelerated the rate of degradation of either endogenous IFNAR1 ( Figure 4C ) or exogenously expressed Flag-tagged IFNAR1 ( Figure 4D ) . Importantly , inactivated HSV did not increase the rate of proteolytic turnover for the priming site deficient IFNAR1S532A mutant ( Figure 4D ) . Together , these results suggest that HSV rapidly initiates a specific PERK-independent signaling pathway that leads to IFNAR1 priming phosphorylation and degradation . One possibility is that such a signaling pathway could be initiated by the recognition of pathogen patterns within inactivated HSV . HSV was reported to activate Toll like receptors ( TLR ) , including TLR9 , via viral genomic DNA [43] , [44] as well as TLR2 via an unidentified molecular structure on the virion [45] . While we failed to detect an increase in priming or degron phosphorylation of IFNAR1 in KR-2 cells upon treatment with an activator of TLR2 muramyl dipeptide ( MDP , data not shown ) , such phosphorylation was readily observed when KR-2 cells were treated with HSV or with TLR9 inducer CpG ( Figure 5A ) . Whereas these results do not prove or disprove the participation of specific TLR in IFNAR1 phosphorylation mediated by HSV , they indicate a possibility that the stimulation of PRR signaling in general might lead to the same result . To test this possibility , we aimed to determine whether other known inducers of PRR signaling were capable of stimulating priming phosphorylation of IFNAR1 . To this end , we switched from KR-2 fibrosarcoma cells to the types of cells that actually function to present foreign antigens , and , accordingly , express numerous types of pathogen recognition receptors . The treatment of human monocytic U937 cells with inducers of TLR9 ( CpG ) or TLR4 ( lipopolysaccharide , LPS ) led to a robust phosphorylation of IFNAR1 on its priming site ( Figures 5B , C ) . Furthermore , an increase in Ser532 phosphorylation of IFNAR1 in U937 cells was also seen in response to other inducers of PRR such as the TLR3 ligand poly I:C and NOD2/TLR2/TLR4 ligand MDP , and in response to high doses of inactivated VSV ( Figure 5D ) . It is plausible that our previous studies , designed to test the effects of VSV using infection at low MOI and analyzed at the late time point of infection when the induction of UPR is at its maximum , [33] had missed this early effect . Signaling pathways triggered by the activation of PRR are known to induce a number of important regulatory kinases such as Jun N-terminal kinases ( JNK ) , IκB kinases ( IKK ) , stress-activated p38 protein kinases , and mitogen-activated Erk kinases ( reviewed in [6] ) . The pre-treatment of U937 cells with a pharmacologic inhibitor of p38 kinase ( SB203580 ) prevented an increase in priming phosphorylation of IFNAR1 in response to LPS . Such an effect was not observed when the JNK inhibitor SP600125 was used ( Figure 5C ) . The inhibition of p38 kinase by SB203580 decreased the phosphorylation of S532 in response to all tested inducers of PRR signaling ( Figure 5D ) . These results collectively implicate p38 protein kinase in mediating the priming phosphorylation of IFNAR1 in response to PRR signaling . To further investigate the contribution of the p38 kinase , we used an in vitro assay in which S532 phosphorylation of bacterially-produced GST-IFNAR1 by cell lysates ( as a source of kinase activity ) was assessed by immunoblotting using a phosho-S532-specific antibody ( as in [35] ) . Under these conditions , lysates from cells treated with UV-inactivated HSV exhibited a greater ability to phosphorylate GST-IFNAR1 on S532 in vitro than lysates from untreated cells . This activity could be tempered by adding p38 inhibitors ( SB203580 or VX702 ) but not by adding the JNK inhibitor SP6000125 ( Figure 6A ) . Furthermore , recombinant p38 kinase was capable of incorporating radiolabeled phosphate groups into the wild type GST-IFNAR1 protein whereas this incorporation was lower when the GST-IFNAR1S532A mutant was used as a substrate ( Figure 6B ) . Finally , the Flag-tagged p38α kinase immunopurified from KR-2 cells was capable of phosphorylating GST-IFNAR1 on S532 in an immunokinase reaction ( Figure 6C ) . This activity was increased when the kinase was purified from cells pre-treated with inactivated HSV . Importantly , no activity was observed when either the catalytically inactive p38AGF mutant was used as a source of kinase or when the phosphorylation-deficient GST-IFNAR1S532A mutant was used as a substrate ( Figure 6C ) . Given that the knock-down of endogenous p38α in U937 cells by shRNA also noticeably decreased the extent of S532 phosphorylation of endogenous IFNAR1 ( Figure 6D ) , these results collectively suggest that the p38 kinase activated by PRR signaling mediates the phosphorylation of the priming site on IFNAR1 . Whereas these data indicate that p38 kinase is capable of phosphorylating Ser532 , our results do not exclude the possibility that another kinase that associates with p38 and depends on p38 activation could function as a direct priming kinase . Consistent with the importance of priming phosphorylation in the ligand-independent pathway , the treatment of U937 cells with LPS activated p38 kinase and also stimulated the phosphorylation of S535 within the degron of IFNAR1 ( Figure 6E ) . This phosphorylation was compromised by pre-treating the cells with a p38 kinase inhibitor SB203580 . An important observation to note here is that this compound did not affect the Ser535 phosphorylation stimulated by IFNα ( Figure 6E ) . Given that IFNα is a poor inducer of priming phosphorylation and of p38 activation and is capable of stimulating the degron phosphorylation independently of the priming site ( [35] and Figure 6E ) , it appears that the role of p38 kinase in IFNAR1 phosphorylation is largely limited to the ligand-independent pathway . Furthermore , these results ( together with a rather poor induction of STAT1 phosphorylation in cells treated with LPS under conditions reported in Figure 6E ) render unlikely the possibility that phosphorylation of the IFNAR1 degron in response to PRR signaling is indirectly mediated by autocrine IFNα/β . We then investigated the effects of PRR signaling on the degradation of IFNAR1 . Consistent with previous reports in other cell lines [27] , [29] , treatment of U937 cells with IFNα markedly increased the rate of IFNAR1 turnover ( Figure 6F ) ; a similar effect observed upon the LPS treatment . Importantly , the effects of LPS ( but not of IFNα ) were blocked by pre-treating the cells with a p38 kinase inhibitor ( Figure 6F ) . These results indicate that PRR signaling specifically accelerates the degradation of IFNAR1 through the activation of p38 kinase whereas this kinase is dispensable for the stimulation of IFNAR1 proteolytic turnover by IFNα . We next sought to investigate whether the acceleration of IFNAR1 degradation by PRR signaling may attenuate the cellular responses to Type I IFN . We focused on the activation of STAT1 ( assessed by its tyrosine phosphorylation ) in mouse bone marrow macrophages that robustly responded to exogenous murine IFNβ , and where such a response could be readily abolished if neutralizing antibodies against IFNα and IFNβ were added to the reaction ( Figure 7A , lanes 1–2 and 13–14 ) . These settings were then used to examine the effect of PRR activators on the extent of IFN signaling . To this end , we pre-treated cells with various combinations of PRR activator LPS , p38 kinase inhibitor SB203580 , and IFNα/β neutralizing antibodies . We then washed off the pre-treatment agents and proceeded to treat the cells with exogenous IFNβ and detect STAT1 phosphorylation . The pre-treatment of macrophages with LPS alone noticeably increased the basal levels of STAT1 phosphorylation ( Figure 7A , lane 3 vs 1 ) . A similar result was observed when pre-treatment with LPS also included anti-IFNα/β neutralizing antibodies ( lane 9 vs . 3 ) . This outcome suggests that the effects of LPS on STAT1 phosphorylation should not be attributed to production of endogenous Type I IFN and were likely mediated by other cytokines ( for example , IFNλ that is known to be produced upon PRR stimulation [46] , [47] , [48] ) . However , the activation of STAT1 in response to the subsequent addition of exogenous IFNβ was noticeably impaired in cells pre-treated with LPS ( Figure 7A , compare lanes 2 and 4 ) . This result suggests that the activation of PRR in cells prior to their encounter with Type I IFN may temper future responses of a cell to these cytokines . Remarkably , the suppressive effect of LPS on IFNβ signaling was partially alleviated when pre-treatment with LPS was carried out in the presence of the p38 kinase inhibitor SB203580 ( Figure 7A , lane 8 vs . lane 4 ) . Given that the effects of the p38 inhibitor were largely unaffected by adding the neutralizing antibodies to the pre-treatment combination of LPS and SB203580 ( Figure 7A , lane 12 vs . lane 8 ) , it is unlikely that the inhibitor acted through stimulating an additional production of endogenous Type I IFN . Collectively , these results suggest that PRR signaling-induced activation of p38 kinase in cells may attenuate their responses to a subsequent exposure to Type I IFN . To further determine whether the suppression of IFN signaling by PRR inducers requires IFNAR1 phosphorylation , we used the knock-in mice harboring the IFNAR1S526A mutant , which is insensitive to downregulation in response to HSV infection ( Figure 1B ) . Whereas pre-treatment with LPS dramatically inhibited IFNβ-induced STAT1 phosphorylation in bone marrow macrophages from wild type mice , this inhibition was not seen in cells that express the non-degradable IFNAR1 mutant ( Figure 7B ) . These data collectively suggest that p38 kinase activity-dependent phosphorylation and the downregulation of IFNAR1 in response to PRR signaling decrease the extent of cellular responses to Type I IFN . Intriguingly , LPS was reported to promote the maturation of human DCs of monocytic origin , a process during which the downregulation of IFNAR1 and the entire Type I IFN receptor had been previously reported [4] , [49] , [50] . Autocrine/paracrine Type I IFN produced by DCs not only plays an important role in their function but also exerts pro-apoptotic effects on DCs themselves [12] , [16] . Given that suppression of cell viability elicited by some of Type I IFN species are most prominent in cells that express high levels of receptor chains [51] , it is plausible that IFNAR1 downregulation may help DCs to survive under exposure to their own IFNα/β . To test this hypothesis , we assessed the viability of mouse bone marrow-derived DCs activated by LPS . Treatment with a p38 kinase inhibitor significantly decreased the viability of these cells ( Figures 7C ) . Remarkably , this effect was less evident in DCs that were either derived from IFNAR1-null mice ( Figure 7C ) or treated with the IFNAR1-neutralizing antibody ( Figure 7D ) indicating that the activation of p38 kinase may be important for protecting DCs from detrimental effects of Type I IFN . Furthermore , whereas the percent of viable LPS-activated BMDC from the IFNAR1S526A knock-in mice was relatively low ( 10 . 7±2 . 0% ) , it could be noticeably increased by incubating these cells with the IFNAR1-neutralizing antibody ( 20 . 3±3 . 4% , p<0 . 01 ) . Similar data were obtained when Annexin V-negative CD11c-expressing cells were analyzed ( Figure S2 ) . Taken together , these data suggest that PRR-stimulated p38 kinase-dependent degradation of IFNAR1 results in protection of activated DCs from the detrimental effects of autocrine/paracrine IFNα/β .
We have previously reported that the induction of UPR activates a ligand/JAK-independent signaling pathway that leads to phosphorylation , ubiquitination , and degradation of IFNAR1 . This signaling involves stimulation of PERK-dependent priming phosphorylation of IFNAR1 followed by its degron phosphorylation by CK1α . This pathway , which can be activated in response to VSV or HCV infection , plays an important role in regulating the levels of IFNAR1 in naïve cells and in determining the sensitivity of cells to the future exposure to Type I IFN [29] , [33] , [34] , [35] . In the present study , we investigated whether HSV infection also negatively affects IFNAR1 stability and signaling . We found that ligand/TYK2-independent phosphorylation and downregulation of IFNAR1 can indeed be observed in cells infected with HSV ( Figure 1 ) . Similar to VSV [35] , HSV infection induced the priming phosphorylation of IFNAR1; and this phosphorylation was required for IFNAR1 ubiquitination and downregulation ( Figure 2 ) . However , when we next investigated the role of PERK in these processes , the differences between HSV and VSV became apparent . Unlike VSV , HSV infection caused little ( if any ) increase in the phosphorylation of the major PERK substrate , translational regulator eIF2α ( Figure 3A ) . Although available literature suggests that some of this effect could be attributed to the action of phosphatases directed by the HSV protein γ134 . 5 [40] , our studies together with another report [41] indicated a deficient activation of PERK in cells infected by HSV ( Figures 3B–C ) . Furthermore , genetic experiments using PERK knockdown in human cells and PERK knockout in mouse cells clearly demonstrated that PERK is dispensable for IFNAR1 phosphorylation and downregulation by HSV ( Figure 3C–D ) . Given that high doses of inactivated HSV also stimulated IFNAR1 phosphorylation , downregulation , and degradation ( Figure 4 ) , we proposed that there is a novel branch of the ligand-independent pathway . We hypothesized that this signaling branch could be induced by pathogen recognition receptors . Once this initial hypothesis received support from experiments that used canonical selective activators of PRR signaling ( Figure 5 ) , we changed the focus of our study to follow the effects of this signaling . Accordingly , we modified the experimental design and switched to using these canonical activators ( due to their commercial availability and better reproducibility over preparations of inactivated virus ) and to cell models that specifically reflected the function of pathogenic patterns recognition and ensuing reactions of innate immunity . Our subsequent studies demonstrated that , even in the absence of viral infection , the activation of PRR signaling robustly induces priming phosphorylation and the degradation of IFNAR1 in a manner that requires the activation of p38 kinase ( Figures 5–6 ) . Furthermore , p38 kinase-dependent phosphorylation of IFNAR1 leads to IFNAR1 degradation and , accordingly , tempers the responses of cells to their future encounters to IFNα/β ( Figures 6–7 ) . The role of the p38 kinase in these processes is intriguing . On one hand , the results of pharmacologic ( Figure 5D ) and genetic ( Figure 6D ) studies implicate p38 kinase in the PRR-induced downregulation of IFNAR1 . Furthermore , our biochemical data suggest that p38 kinase is capable of directly phosphorylating the priming site on IFNAR1 in vitro ( Figure 4A–C ) . However , given a known preference of this kinase for proline-directed Ser and Thr residues as phospho-acceptor sites [52] and the fact that the priming site on IFNAR1 does not conform to these criteria , it is plausible that the direct phosphorylation of Ser532 in cells might be carried out by a SB203580-sensitive kinase that associates with p38 kinase and depends on p38 activation . On another hand , p38 kinase can be also activated by Type I IFN in several types of cells [53] , [54] , [55] . In fact , within cells that have already encountered it , IFNα/β , p38 kinase activity is proven to contribute to the maximal extent of the IFN-induced transcriptional program [10] , [56] , [57] . Yet it appears that the ligand-induced phosphorylation and degradation of IFNAR1 does not depend on p38 kinase activity ( Figure 6E–F ) . Indeed , our recent study identified protein kinase D2 as a key TYK2-dependent IFN-inducible kinase that mediates the ligand-stimulated IFNAR1 phosphorylation , ubiquitination , endocytosis and degradation [28] . Furthermore , it appears that the activation of p38 kinase in cells that have not been yet exposed to IFNα/β may temper future sensitivity to these cytokines through an elimination of the receptor ( Figure 7E ) . Collectively , these studies describe a novel link between an activation of innate immune responses that usually govern production of Type I IFN , with modulation of the extent of cellular responses elicited by these cytokines ( Figure 7E ) . It is plausible that the temporal downregulation of IFNAR1 that precedes or coincides with the peak of IFNα/β synthesis could be important for various aspects of the host defenses . These aspects may include the maintenance of the viability of IFN-producing cells , limiting the extent of IFNα/β pathway , and affecting the sensitivity of the host to the secondary infection . Among cell types capable of producing IFNα/β , the dendritic cells ( DCs ) are distinguished with an advanced ability to recognize a variety of pathogenic patterns and , upon this activation , synthesize and secrete Type I IFNs and other cytokines that participate in shaping the immune responses [6] , [58] . Activated DCs that produce IFNα/β have to be protected from the detrimental effects of autocrine IFN [4] , [12] , [16] . Indeed , it has been shown that activated DCs are prone to apoptosis , the extent of which is decreased in cells from IFNAR1-null animals [12] , [16] . It has also been demonstrated that , upon their maturation , DCs downregulate Type I IFN receptor [49] , [50] though the mechanism underlying this downregulation or its role in DC maturation and survival remain unclear . In this study , we demonstrated that PRR-stimulated p38 kinase-dependent degradation of IFNAR1 leads to an attenuation of Type I IFN signaling and ameliorates its negative effects in DCs ( Figure 7 ) . These data suggest that the activation of the signaling pathway that involves phosphorylation and degradation of IFNAR1 may render DCs ( and perhaps other IFNα/β-producing cells ) refractory to their own Type I IFN . As indicated by our data ( presented in Figure 7 ) , such refractoriness may improve the viability of DCs . In addition , the mechanism described here may plausibly contribute to discontinuing an otherwise potentially harmful cycle of DCs being activated by Type I IFN followed by the production of more of these cytokines and a subsequent greater activation of DCs . Given the well documented role of IFNα/β in the pathogenesis of autoimmune disorders ( including systemic lupus erythematosus , psoriasis , and Type 1 diabetes mellitus [17] , [18] , [19] , [20] ) , temporal downregulation of IFNAR1 might play an important role in protecting the host from such autoimmune reactions . Whereas the latter outcomes of IFNAR1 degradation stimulated by signaling induced by pathogenic patterns may benefit the host , it could also decrease the ability of some cell types exposed to PRR inducers to mount an appropriate anti-viral response . It remains to be seen whether the suppression of Type I IFN signaling by chronic exposure to other pathogens in other cell types may play a role in the development of secondary viral infections , which have been linked to insufficient IFNα/β function [59] . Future studies in vivo aimed at a further delineation of the mechanisms of IFNAR1 degradation and its role in sensitivity to secondary infections and autoimmune disorders are consequently warranted .
Recombinant GST-p38α was purchased ( Cell Signaling ) . Vectors for bacterial expression of GST-IFNAR1 and mammalian expression of human and murine Flag-IFNAR1 were described previously [25] , [33] , [35] . The plasmids for expression of Flag-p38 ( wild type or catalytically inactive AGF ) were a generous gift from R . J . Davis . ShRNA constructs for knocking down p38α kinase were from Sigma . Constructs for knock down of PERK were previously described [33] . Recombinant human IFNα2 ( Roferon ) was from Roche . Recombinant murine IFNβ and mouse IFNα/β neutralizing antibodies were from PBL . LPS , poly IC , MDP and CpG were from InVivogen . Neutralizing antibody against mouse IFNAR1 ( MAR ) was from Leinco . P38 inhibitor SB203580 and JNK inhibitor SP600125 were from EMD Biosciences . P38 inhibitor VX-702 was from ChemTek . CK1 inhibitor D4476 was from Tocris . All other reagents were from Sigma . VSV ( Indiana serotype , a gift from R . Harty ) and HSV-1 ( KOS strain , a gift from G . Cohen and R . Eisenberg ) were propagated in HeLa cells . These cells were also used to determine the viral titers in serially diluted stocks using methylcellulose method . In experiments requiring inactivated virus , virus suspension was placed in Petri dishes and exposed either to UV-C light ( 254 nm ) for 5 min to achieve the total dose of 1500 J/m2 or sham treated for the same time period . Human monocytic U937 and HeLa cells were from ATTC . Replication-deficient lentiviral particles encoding shRNA against GFP ( shCON ) , PERK , p38 , or the empty virus control were prepared via co-transfecting 293T cells with three other helper vectors as described previously [33] , [35] . Viral supernatants were concentrated by PEG8000 precipitation and were used to infect U937 cells in the presence of polybrene ( 4 µg/mL , Sigma ) . Cells were selected and maintained in the presence of 1 µg/mL of puromycin . Human fibrosarcoma 2fTGH-derived IFNAR2-null U5A [39] cells were kindly provided by G . Stark; the isogenic derivatives of TYK2-null 11 . 1 cells including the KR-2 cells that harbor catalytically inactive TYK2 or WT-5 cells that re-express wild type TYK2 [27] were a generous gift from S . Pellegrini . 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 . The protocol was approved by the Institutional Animal Care and Use Committee ( IACUC ) of the University of Pennsylvania ( protocol # 800992 ) . Every effort was made to minimize animal suffering . IFNAR1-null mice were kindly provided by D . E . Zhang ( UCSD ) . Bone marrow-derived macrophages ( BMM ) and bone marrow-derived dendritic cells ( BMDC ) were produced as described previously [12] . These assays were carried out as described previously [34] . Monoclonal antibodies against human IFNAR1 that were used for immunoprecipitation ( EA12 ) or immunoblotting ( GB8 ) were described in detail elsewhere [60] . Antibodies against PERK [33] were kindly provided by J . A . Diehl . Antibodies against p-STAT1 , p-p38 ( Cell Signaling ) , phospho-Ser532 , phospho-Ser535 ( or phospho-Ser523 , phospho-Ser526 , respectively , in the murine receptor ) [26] , [35] , murine IFNAR1 ( LeinCo ) , STAT1 , p38 , phospho-PERK ( Santa Cruz ) , Flag , β-actin ( Sigma ) and ubiquitin ( clone FK2 , Biomol ) were used for immunoprecipitation ( IP ) and immunoblotting ( IB ) as described previously [26] , [35] . Cell viability assays were analyzed by FACS ( BD Calibur ) to calculate CD11c positive and propidium iodide negative cell population as described previously [12] , [33] . Kinase assays were carried out as described previously [29] , [34] . In brief , p38 was immunoprecipitated and the immunoprecipitates were incubated with 1 µg of substrates ( bacterially-expressed and purified wild type or S532A GST-IFNAR1 mutant ) in kinase buffer ( 25 mM Tris-HCl pH 7 . 4 , 10 mM MgCl2 , 1 mM NaF , 1 mM NaVO3 ) and ATP ( 1 mM ) at 30°C for 30 min . Samples were then separated by 10% SDS-PAGE and analyzed by immunoblotting with phospho-specific antibodies . | We have previously observed and reported that the activation of unfolded protein responses during infection with some viruses such as vesicular stomatitis virus and hepatitis C virus compromises the cellular responses to the cytokines that belong to family of Type I interferons ( IFNα/β ) . These effects apparently rely on the activation of a specific protein kinase termed PERK that transduces signals from unfolded proteins to accelerate phosphorylation-dependent degradation of the IFNα/β receptor . Here we found that the same pathway can be PERK-independent and can be stimulated by signaling induced in the specialized cells ( e . g . dendritic cells that produce IFNα/β ) , which recognize pathogens ( e . g . viruses ) . This work delineates a mechanism by which the degradation of IFNα/β receptor is accelerated . Our studies also highlight the importance of this mechanism for limiting the magnitude and the duration of IFNα/β signaling and maintaining the survival of IFN-producing dendritic cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"biology"
]
| 2011 | Pathogen Recognition Receptor Signaling Accelerates Phosphorylation-Dependent Degradation of IFNAR1 |
Circadian clocks control the timing of animal behavioral and physiological rhythms . Fruit flies anticipate daily environmental changes and exhibit two peaks of locomotor activity around dawn and dusk . microRNAs are small non-coding RNAs that play important roles in post-transcriptional regulation . Here we identify Drosophila miR-210 as a critical regulator of circadian rhythms . Under light-dark conditions , flies lacking miR-210 ( miR-210KO ) exhibit a dramatic 2 hrs phase advance of evening anticipatory behavior . However , circadian rhythms and molecular pacemaker function are intact in miR-210KO flies under constant darkness . Furthermore , we identify that miR-210 determines the evening phase of activity through repression of the cell adhesion molecule Fasciclin 2 ( Fas2 ) . Ablation of the miR-210 binding site within the 3’ UTR of Fas2 ( Fas2ΔmiR-210 ) by CRISPR-Cas9 advances the evening phase as in miR-210KO . Indeed , miR-210 genetically interacts with Fas2 . Moreover , Fas2 abundance is significantly increased in the optic lobe of miR-210KO . In addition , overexpression of Fas2 in the miR-210 expressing cells recapitulates the phase advance behavior phenotype of miR-210KO . Together , these results reveal a novel mechanism by which miR-210 regulates circadian locomotor behavior .
The circadian locomotor rhythms of flies are generated by a neuronal network , which consists of ~150 brain circadian neurons expressing core pacemaker genes [1–4] . These neurons can be further divided into 7 clusters based on their cell localization and neurotransmitters they express [4] . There are 3 groups of dorsal neurons ( DN1 , DN2 , and DN3 ) , 2 groups of ventral-lateral neurons ( large and small LNv ) , dorsal lateral neurons ( LNd ) , and lateral posterior neurons ( LPN ) [4] . The large LNvs ( lLNvs ) and 4 small LNvs ( sLNvs ) express the neuropeptide Pigment Dispersing Factor ( PDF ) , while the 5th sLNv is PDF negative . Under regular light-dark ( LD ) conditions , flies exhibit a bimodal pattern of locomotor rhythms , peaking around dawn and dusk , which are termed morning peak and evening peak , respectively . According to the dual-oscillator model , separate morning and evening oscillators track these behavior peaks separately [5] . Indeed , studies have shown that the PDF-positive sLNvs are responsible for promoting the morning peak , while the LNds , as well as the fifth sLNv , are mainly responsible for generating the evening activity [6 , 7] . In addition , the DN1s are important to integrate environmental inputs such as light and temperature , and regulate circadian rhythms and sleep behavior [8 , 9] . A recent study indicated that DN1s sense acute temperature fluctuations to modulate fly sleep [10] . The PDF positive sLNvs are the master pacemaker neurons in fly brain: they synchronize other circadian neurons to maintain robust circadian rhythms under constant darkness [11] . These sLNvs send axonal projections toward the dorsal brain region , where the DN1s and DN2s are located [12 , 13] . The dorsal projections of sLNvs exhibit circadian arborization rhythms with higher complexity of axon terminals found in the early day and lower complexity at the night [14] . The physiological relevance and molecular mechanisms underlying this structural plasticity of sLNvs remain unclear . The Rho1 GTPase was identified to modulate the sLNv plasticity and affect seasonal adaptation by phosphorylating Myosion [15] . The cell adhesion molecule Fasciclin 2 ( Fas2 ) has been found to regulate the arborization rhythms of sLNvs [16] . In addition , two matrix metalloproteinases , MMP1 and MMP2 as well as the pacemaker protein VRILLE were shown to be required for the structural remodeling control of sLNv projections [17 , 18] . Circadian rhythms are generated by an intracellular molecular clock , which is conserved across the animal kingdom [19] . The core of this molecular pacemaker is a negative transcriptional-translational feedback loop [20–22] . In Drosophila , CLOCK ( CLK ) dimerizes with CYCLE ( CYC ) to activate rhythmic transcription of hundreds of genes through the E-box region [23 , 24] . Among these clock-controlled genes , PERIOD ( PER ) and TIMELESS ( TIM ) are the key transcription repressors: they form heterodimers in the cytoplasm , and then enter into the nucleus to block their own transcription . The abundance of PER and TIM is tightly regulated by a series of post-translational modifications such as phosphorylation , glycosylation and ubiquitination [25–28] . Recent research has revealed an abundance of post-transcriptional regulation of circadian rhythms by both RNA binding proteins and miRNAs [29–31] . miRNAs are small non-coding RNAs , which repress target gene expression through mRNA degradation and/or translation inhibition , thus playing crucial roles in post-transcriptional regulation [32] . Recent studies have uncovered functions of the miRNA biogenesis pathway and specific miRNAs in regulation of different aspects of animal circadian rhythms [31 , 33–35] . For instance , mouse miR-132/212 modulates the seasonal adaptation and circadian entrainment to day length [36] . miRNAs also target the molecular clock to control period length and rhythmicity of circadian locomotor activity [37] . In Drosophila , the miRNA bantam regulates circadian locomotor period by repressing Clk , while miR-276a and let-7 inhibit the clock genes timeless and clockwork orange to modulate circadian rhythms respectively [33 , 38 , 39] . Moreover , various circadian output pathways are controlled by miRNAs . The miR959-964 cluster miRNAs regulate the circadian timing of feeding and immune response , while miR-279 modulates circadian locomotor behavior output [40 , 41] . Recently , studies have demonstrated that miR-124 specifically controls the phase of circadian locomotor rhythms under constant darkness [42 , 43] . Previously we identified that depletion of GW182 , the key protein for miRNA function , affects PDF-receptor signaling and disrupts circadian rhythms [35] . To further understand the roles of miRNA , we performed a genome-wide screen for miRNAs controlling circadian rhythm phenotypes . Here we show that a conserved miRNA in animals , miR-210 , regulates circadian rhythms in Drosophila . miR-210 determines the phase of evening anticipatory behavior through inhibition of Fas2 .
To identify miRNAs that regulate circadian rhythms , we screened the circadian locomotor rhythms of available miRNA mutants from the Bloomington stock center . Circadian behavior of 46 lines in total was observed both under constant darkness ( DD ) and light dark ( LD ) cycles . Most of the miRNA mutants we tested exhibited normal circadian rhythms under DD; though , a few miRNAs had a slightly altered period or rhythmicity ( S1 Table ) . Interestingly , we found that the miR-210 mutant ( miR-210KO ) exhibits a ~1 . 5 hour advance of the subjective evening peak in DD ( S1 Fig ) . We then closely examined for the locomotor behavior under LD . Indeed , we found that the miR-210 mutant ( miR-210KO ) clearly advances the phase of the evening anticipatory peak about 2 hours ( Fig 1A and 1B ) . The miR-210KO mutant was generated by knocking-in a GAL4 cDNA in replacement of the endogenous miR-210 sequence [44] . To rule out the possibilities of off-target or genetic background effects , we utilized the miR-210 knock-in mutant to perform a rescue . By restoring the expression of miR-210 , we successfully corrected the evening anticipation in the miR-210KO ( Fig 1A and 1B ) under LD , which confirms that the phase advance is due to lack of miR-210 . We also observed a partial but significant rescue of the phase in DD ( S1 Fig ) . To examine whether miR-210 plays a role on evening anticipatory behavior under other entrainment conditions , we entrained the flies under 29°C: 21°C cycles ( TC ) in constant darkness . Under TC , the evening anticipation of wild-type flies exhibited an advance about 2 . 5 hours ( Fig 1C and 1D ) [45] . In miR-210KO , we still observed an additional 2 hrs phase advance compared to the controls ( Fig 1C and 1D ) . Furthermore , we were able to rescue this behavior defect . After released into constant 21°C or 29°C in DD , the miR-210KO flies exhibited normal period and rhythmicity as wild-type controls ( S2 Table ) . These data confirmed that miR-210 is a clear regulator of evening phase during different entrainment conditions . The phase advance of evening anticipation in miR-210KO is reminiscent of the pdf mutant phenotype [11] . To test the genetic interaction of miR-210 and the PDF pathway , we generated miR-210KO; pdf01 double mutants . Under 12:12 LD conditions , the phase advance of the evening peak in miR-210KO was indistinguishable from the pdf01 mutants ( S2 Fig ) . No additive effect on the phase advance was observed in the miR-210KO; pdf01 double mutants comparing with miR-210KO or pdf01 ( S2A and S2B Fig ) . To observe the phenotype even more clearly , we tested these flies under long photoperiods with 16:8 LD cycle so that we avoid acute effects of the light-off transition on the shape of the evening peak . As observed in pdf01 mutant , miR-210KO advanced the phase of evening anticipation by about 2 hrs ( S2C Fig ) . Again , no additive phase advance was observed in the miR-210KO; pdf01 double mutants ( S2D Fig ) . These data suggest that miR-210 functions in the same pathway as PDF in regulation of evening phase under entrainment . To further test the role of miR-210 , we first overexpressed it in all circadian tissues using tim-GAL4 [46] . However , most of the flies with overexpression died rapidly after the LD cycles , which indicates that overexpression leads to severe health issues ( Table 1 ) . Thus , we then used the pdf-GAL4 to specifically drive miR-210 expression in PDF positive pacemaker neurons . The majority of flies ( ~63% ) with miR-210 overexpression in PDF neurons became arrhythmic under DD . However , for the 37% rhythmic flies , the circadian period was lengthened about 2 hours . Thus , while loss of miR-210 has no effect on the period and amplitude of circadian rhythms in DD , overexpression is detrimental to the proper function of the circadian oscillator in PDF neurons . Since the period and rhythmicity of circadian rhythms are unaffected in miR-210KO flies under DD , it is likely that the molecular pacemaker of miR-210KO flies is still functional . To confirm this , we examined the oscillation of the key pacemaker protein PER in circadian neurons under DD . We focused on the PDF positive sLNvs , which are the master pacemaker neurons under constant conditions . As expected , we found no obvious changes in PER oscillation or abundance in these sLNvs ( S3A and S3B Fig ) . We further examined PER levels in another two important groups of circadian neurons: LNds , and DN1s . Similar as in sLNvs , PER oscillation in these neurons were not affected either ( S3C and S3D Fig ) . Since we observed the dramatic phase advance phenotype under LD , next we examined the PER oscillation at different times of day . Consistent with the DD results , there was no significant change of PER in the PDF positive sLNvs in miR-210KO flies ( Fig 2A and 2B ) . Oscillation of PER was unaffected in LNds and DN1s either ( S4 Fig ) . Because PDF is an important circadian output molecule , we also quantified the PDF abundance in sLNvs and found no significant differences in miR-210KO flies in sLNvs ( Fig 2C ) . Since lLNvs have been identified in regulation of evening activity onset [47] , we further examined the PDF abundance of lLNvs across different time of the day . No obvious changes of PDF abundance in lLNv were observed either ( S5 Fig ) . Taken together , these data suggest that loss of miR-210 has no effects on the molecular clock . The phase advance phenotype is unlikely mediated through changes of PDF abundance . Although genetic interaction between miR-210 and the PDF pathway was identified ( S1 Fig ) , we observed no significant changes in PDF abundance in the flies missing miR-210 ( Fig 2C and S5C Fig ) . Thus , we decided to examine the PDF projections of the sLNvs . Circadian arborization rhythms of the dorsal projections of sLNvs have been previously observed [14 , 15] . To determine whether miR-210 affects the arborization rhythms of sLNv projections , we used a PDF-specific antibody to examine the termini of sLNv dorsal projections in miR-210KO flies at early day ( Zeitgeber time 2 ( ZT2 ) , ZT0 is light on and ZT12 is light off ) and early night ( ZT14 ) . We found that the wild-type flies had more PDF positive branches of axon terminals at ZT2 than at ZT14 ( Fig 3A ) , as previously observed [15] . Remarkably , this arborization rhythm was abolished in the miR-210KO flies ( Fig 3A ) . We quantified the dorsal axon arborization using the method as previously described ( see Material and Methods ) [16] and confirmed that the lack of aborization rhythm in miR-210KO is mainly due to the significant decrease of axonal crosses at ZT2 compared to control flies ( Fig 3B ) . Furthermore , restoration of miR-210 expression rescued the arborization rhythm in miR-210KO mutants ( Fig 3A and 3B ) . Since we found there were no significant changes of PDF abundance in the sLNv soma , these data indicate that the decrease of axonal branches was not due to the decrease of PDF staining ( Fig 2 and Fig 3 ) . Together , these results indicate that miR-210 is required for the circadian arborization rhythms in the sLNv dorsal projections . miRNAs play negative posttranscriptional functions through binding to the 3’ untranslated regions ( UTRs ) of target mRNAs [31] . After searching putative miR-210 targets by using an in silico prediction algorithm ( http://www . targetscan . org/fly_12/ ) , we identified Fas2 as one of the potential targets ( Fig 4A ) . The putative miR-210 binding sites within the 3’UTR of Fas2 are highly conserved across most Drosophila species ( Fig 4A ) . Interestingly , Sivachenko et al . have reported that overexpression of Fas2 in the PDF neurons caused fasciculation of sLNv dorsal projections and disrupted arborization rhythms [15] . Since miRNAs are normally negative regulators of target genes , and overexpression of Fas2 recapitulated the anatomical phenotype of miR-210KO flies , we decided to use the CRISPR-Cas9 system to delete the miR-210 binding site within the Fas2 3’UTR . As the seed region of miRNAs ( positions 2–7 ) is critical for miRNA function [48] , to minimize potential off-target effects , we managed to delete only 7 bps of the matching sequence on the 3’UTR ( Fig 4B ) . We recovered several lines , which were homozygous viable , and named the mutant as Fas2ΔmiR-210 . Next we tested the circadian locomotor behavior of the Fas2ΔmiR-210 . There was no change in period or rhythmicity under DD ( Table 1 ) . However , the Fas2ΔmiR-210 flies clearly advanced the phase of evening activity , similar to miR-210KO flies ( Fig 4C and 4D ) . To determine whether Fas2 and miR-210 function in the same pathway , we tested the potential genetic interactions using two methods . First , we tested the miR-210KO/ Fas2ΔmiR-210 transheterzygous flies . While both miR-210KO and Fas2ΔmiR-210 are recessive mutations , the transheterzygous flies had a strikingly similar phase advance phenotype as single mutants ( Fig 4C and 4D ) . Second , we examined the double mutant of miR-210KO , Fas2ΔmiR-210 flies . Under LD , the miR-210KO , Fas2ΔmiR-210 flies exhibited a similar phase advance phenotype , which is insignificant from either miR-210KO or Fas2ΔmiR-210 ( Fig 4C and 4D ) . This data indicated that miR-210 and Fas2 function in the same pathway regulating circadian phase of activity . Taken together , Fas2 genetically interacts with miR-210 , and ablation of the miR-210 binding sites within the 3’UTR of Fas2 recapitulated the behavior phenotypes of miR-210KO mutants . These results indicate that Fas2 is a functional target of miR-210 . To identify the cellular mechanism of miR-210 functions , we first characterized the miR-210 expression pattern in the fly brain by using the miR-210 knock-in Gal4 to drive GFP as a reporter . We observed GFP signals in the optic lobe , Hofbauer-Buchner eyelet ( H-B eyelet ) , as well as several other brain regions , including the mushroom bodies and antennal lobe ( Fig 5A ) . To our surprise , no GFP signal was detected in LNvs as labeled by anti-PDF antibody staining . Then we examined Fas2 expression using the commercial antibody ( 1D4 ) from DSHB . As previously reported , in wild type flies Fas2 was strongly expressed in the mushroom bodies as well as the ellipsoid bodies ( Fig 5B ) [49] . Interestingly , we detected an intensive Fas2 signal in the optic lobe of miR-210KO mutant flies , which disappeared when miR-210 expression was rescued ( Fig 5B ) . We quantified Fas2 abundance in the optic lobe and observed significant increase of Fas2 in miR-210KO mutants compared to the wild-type and rescue ( Fig 5C and 5D ) . Lastly , we wondered whether the increase of Fas2 abundance occurs at the same sites where miR-210 is expressed . While female heterozygous miR-210KO flies exhibited no detectable Fas2 in the optic lobe , dramatic increase of Fas2 signal clearly overlapped with miR-210 expression in the optic lobe ( Fig 4E ) . Although we did not detect miR-210 expression in circadian neurons with the miR-210 knock-in Gal4 , the single cell RT-PCR data from Chen and Rosbash indicated that miR-210 has an oscillating expression in the sLNvs [50] . Thus , we first tested the requirement of miR-210 in the sLNvs for normal circadian behavioral rhythms under LD , combining miR-210KO GAL4 and Pdf-GAL80 . GAL80 is a repressor of GAL4 function , and Pdf-GAL80 has been efficiently used to inhibit gene expression in the PDF neurons [7 , 51] . Surprisingly , we observed a clear rescue of evening phase when we excluded miR-210 expression in the sLNvs , which suggests that miR-210 is not necessary in PDF neurons for circadian behavior ( Fig 6A and 6B ) . Next , we tested whether overexpression of Fas2 in the miR-210-expression cells would mimic the behavior phenotype of miR-210KO mutants using the miR-210KO GAL4 . While heterozygous miR-210KO mutant has no phase advance , we observed a weaker ( ~1 . 5 hr ) but significant advance of evening peak in Fas2 overexpression ( Fig 6C and 6D ) . Consistent with the miR-210 data , Fas2 is not required in the PDF positive neurons ( Fig 6C and 6D ) . Since increase of Fas2 in the optic lobe was observed in the miR-210KO mutants , we tested whether up-regulation of Fas2 is responsible for the phase advance phenotype . When we used GMR-GAL80 to block Fas2 expression in the photoreceptors , the advance of evening phase was rescued ( Fig 6C and 6D ) . Together these data suggest that inhibition of Fas2 by miR-210 in the optic lobe is critical for the timing of evening activity .
Emerging roles of miRNA in the control of different aspects of circadian rhythms have been recently uncovered [33] . Here , we screen Drosophila miRNA mutants and demonstrate that miR-210 is critical for circadian locomotor rhythms . miR-210 determines the proper phase of evening activity peak under entrainment . miR-210 plays its role in circadian rhythms and axonal arborization via repression of Fas2 . Drosophila gradually increases its locomotor behavior and reaches its peak of evening activity at light off . The molecular mechanism underlying the phase control of evening peak is still largely unknown . Loss of miR-210 advanced the evening phase to the same extent as pdf01 mutants . Beside PDF , other neuropeptides such as ITP and NPF also play roles in regulation of evening peak . Downregulation of ITP in circadian neurons has weak effects on evening phase and also prolongs circadian period under constant darkness [52] . While ablation of NPF positive circadian neurons advances the phasing of evening activity , it also affects period during free running [53] . These phenotypes are inconsistent with what we observed in miR-210KO mutants ( Table 1 ) . In fact , our genetic interaction data suggests that miR-210 functions in the same signaling pathway as PDF . So we decided to focus on PDF pathway . However , unlike the pdf01 mutants [11] , loss of miR-210 has little effect on period and circadian rhythmicity . These results indicate that miR-210 may specifically regulate one aspect of PDF signaling functions . Circadian structural plasticity of sLNv dorsal projections has been recently identified [54 , 55] . The biological function and molecular mechanisms underlying the structural plasticity however remain elusive . Here we identified that miR-210KO mutants disrupted the axonal arborization rhythms of sLNv , which is likely through inhibition of the neural cell adhesion molecule Fas2 . Deletion of the 7-bp seed region of miR-210 binding sites within the 3’UTR of Fas2 mimics the effect of miR-210KO mutants on the sLNv arborization rhythm . However it seems unlikely that the phase advanced phenotype of miR-210KO mutants is because of the defects in sLNv arborization rhythms . First , as far as we know , previous genetic manipulations that cause severe defects on sLNv structural plasticity have no effect on evening phase [54 , 55] . Second , although overexpression or down regulation of Fas2 in the sLNv abolished the arborization rhythms , it does not affect the evening phase either [16] . So we conclude that these effects of miR-210 on evening phase control and axonal structural plasticity may not be functionally related . Our evidence that Fas2 is a key target for miR-210 in regulation of circadian rhythms is particularly strong . First , miR-210KO mutants elevated the Fas2 abundance in the optic lobe , where miR-210 is highly expressed . This is consistent with the general role of miRNAs as negative regulators of target proteins . Second , Fas2ΔmiR-210 recapitulates the arborization and circadian behavior phenotypes of loss of miR-210 . Furthermore , our genetic interaction data also indicates that Fas2 and miR-210 function in the same pathway . In which cells are miR-210 and Fas2 required for the evening phase of activity ? Transcriptome profiling from the Rosbash lab show that expressions of both miR-210 and Fas2 are enriched and oscillating in the PDF positive neurons [50] . Overexpressing miR-210 in the PDF neurons causes flies to become arrhythmic or rhythmic with a prolonged circadian period ( Table 1 ) . However , using miR-210 knock-in GAL4 or Fas2 antibodies , we were not able to detect obvious signals for miR-210 or Fas2 in the PDF neurons . We cannot exclude the possibility that there is weak amount of miR-210 or Fas2 protein expression , which is not detectable due to sensitivity of the techniques we used . We sought to answer this question by utilizing the well-used Pdf-GAL80 repressor . Indeed , restoring miR-210 in its endogenous expressing cells except PDF neurons rescued the circadian behavior completely . This data indicates that expression of miR-210 in the PDF neurons is not necessary for its circadian function . Consistent with this notion , overexpression of Fas2 in PDF negative but miR-210 expressing cells causes a phase advance of evening peak . Since the miR-210 knock-in driver has no detectable expression in the PDF neurons , leaky expression due to potential weak suppression of Pdf-GAL80 should not be an alternative explanation . The H-B eyelets interact with lLNvs and control the phase of evening peak activity [47 , 56] . Interestingly , miR-210 exhibits strong signals of expression in the H-B eyelets ( Fig 5A ) . We labeled H-B eyelets in miR-210KO or Fas2ΔmiR-210 by expressing GFP in the Rh6-Gal4 ( S6 Fig ) . In wild-type flies , H-B eyelet axons projected close to the LNvs as previously described [47] . We did not observe obvious defects of H-B eyelet in these two mutants ( S6 Fig ) . However , we cannot exclude the possibility that miR-210KO or Fas2ΔmiR-210 mutants disrupt the communication between H-B eyelets and lLNv , thus affect the evening activity . Our data suggest that inhibition of Fas2 by miR-210 in the optic lobe is critical for the proper evening phase . How does miR-210-Fas2 expression in the optic lobe control the phase ? One possibility is that Fas2 expression in the optic lobe may affect the phase through PDF signaling in the lLNv . If that’s the case , Fas2 abundance in the optic lobe may show differences at different time points . Thus we examined the Fas2 abundance across different time of the day in the optic lobe ( S7A Fig ) . High amount of Fas2 was observed in the miR-210KO compared to the wild-type , however , no oscillation was identified in either genotype ( S7B Fig ) . This result is consistent with the fact that we did not observe changes of PDF in lLNv ( S6 Fig ) . So it is unlikely that miR-210-Fas2 affect the evening phase through changes of PDF level in the ILNv . However , we cannot exclude the possibility that it may modulate PDF-receptor signaling or other pathways . It is worth to mention that during the preparation of this paper , another independent study also characterizes the role of miR-210 in phase control of circadian behavior and reached similar conclusion with us on the circadian behavior and dorsal aborization phenotype [57] . Even though no specific target was functionally validated , it is possible that some of the pathways identified in that paper may also contribute to the phase control phenotypes in the miR-210KO . Here we identify that miR-210 regulates circadian locomotor rhythms and axonal structural plasticity of pacemaker neurons in Drosophila . A similar mechanism may exist in mammals . Interestingly , the glutamatergic synapses on the VIP neurons of mammalian master circadian pacemaker suprachiasmatic nucleus ( SCN ) also show circadian structural plasticity , which maybe important for light entrainment in mice . miR-210 is a highly conserved miRNA from worms , flies , to humans . In silico miRNA target prediction with targetscan identified a highly conserved binding site of human miR-210 in the 3’UTR of vertebrates brain-derived neurotrophic factor ( BDNF ) . BDNF has well-established functions in axonal branching and synaptic plasticity [58 , 59] . Thus , it is worthwhile to test that whether miR-210 plays conserved functions in structural plasticity in mammals .
All the flies were raised on standard cornmeal/agar medium at 25°C under 12:12 hour LD cycle . The following strains were used in this study: w1118 , yw , pdf01 , y w; tim-GAL4/CyO [46] , PDF-Gal80 , y w; Pdf-GAL4/CyO [7] , UAS-miR-210 , UAS-Fas2 , miR-210KO [44] , UAS-CD8-GFP ( BL5137 ) , GMR-Gal80 [60] , Rh6-GAL4 ( BL7464 ) , Fas2ΔmiR-210 ( generated by CRISPR-Cas9 in our lab ) . Most of the times , adult male flies ( 2–5 days old ) were used to test locomotor activity rhythms . Since both miR-210 and Fas2 are both on the x chromosome , adult female flies were also used for locomotor rhythms , which were mentioned in the relevant figures . Flies were entrained for 4 days LD cycle at 25°C and 60% humidity , and released into constant darkness ( DD ) at 25°C for at least 5 days . Temperature cycles were performed with 12h:12h 29C:21C cycling for 6d in complete darkness . Trikinetics Drosophila activity monitors were used to record locomotor activity in I36-LL Percival incubators . Activity rhythms were analyzed with Faasx software protocol [61] . Eduction profiles were generated with 3d of LD activity . FAAS-X software was used to analyze behavioral data [60] . Actograms were generated with a signal-processing toolbox for MATLAB . The evening anticipation amplitude was determined by assaying for the locomotor activity as described . Immunohistochemistry was performed with whole Drosophila brains . For staining involving ZT timepoints , Flies were entrained to LD for 3 d and dissected at Zeitgeber time ( ZT ) 0 , 2 , or 14 . For CT staining , flies were entrained to LD , released to DD , and dissected on the second day of DD at 6 CT intervals . Mouse anti-Fas2 ( 1:100 ) , mouse anti-PDF ( 1:400 ) , Rabbit anti-GFP 1:200 , Rabbit anti-PER ( 1:1500 ) was used . All brains were imaged with a leica confocal microscope . Microscope laser settings were held constant for each experiment . ImageJ was used for quantification of PER , PDF and FAS 2 luminance . At least 5 brains were used for quantification . ImageJ was used in the analysis of axonal morphology ( fasciculation ) of sLNv dorsal termini by a modified Sholl's analysis [56] . Fas2ΔmiR-210 flies were generated by CRISPR mediated deletion by Rainbow Transgenic Flies . The guide RNA target sequence for CRISPR was GCAATGTGCACAAAACGATGAGG . The primers used for genomic miR-210 f3: GCCAACAGGCAGCATCAAAC , and miR-210 r3: CCAACTTAGTGTGCCAATCGATC . PCR amplification of the Fas2 3’ UTR region containing the miR-210 motif was performed by using miR-210 f3 and miR-210 r3 primers . Sequencing of the PCR products was performed by using the miR-210 f3 primer . Identification of homozygous Fas2ΔmiR-210 flies was confirmed by observing a 7bp deletion at the miR-210 binding motif in the sequence chromatograph . | Circadian clocks control the timing of animal physiology . Drosophila has been a powerful model in understanding the mechanisms of circadian regulation . Fruit flies anticipate daily environmental changes and exhibit two peaks of locomotor activity around dawn and dusk . Here we identify miR-210 as a critical regulator of evening anticipatory behavior . Depletion of miR-210 in flies advances evening anticipation . Furthermore , we identify the cell adhesion molecule Fas2 as miR-210’s target in circadian regulation . Fas2 abundance is increased in fly brain lacking of miR-210 . Using CRISPR-Cas9 genome editing method , we deleted the miR-210 binding site on the 3’ untranslated region of Fas2 and observed similar phenotype as miR-210 mutants . Altogether , our results indicate a novel mechanism of miR-210 in regulation of circadian anticipatory behavior through inhibition of Fas2 . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
]
| [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"natural",
"antisense",
"transcripts",
"gene",
"regulation",
"neuroscience",
"biological",
"locomotion",
"animals",
"circadian",
"oscillators",
"animal",
"models",
"micrornas",
"model",
"organisms",
"drosophila",
"melanogaster",
"experimental",
"organism",
"systems",
"chronobiology",
"drosophila",
"research",
"and",
"analysis",
"methods",
"optic",
"lobes",
"animal",
"cells",
"animal",
"studies",
"gene",
"expression",
"insects",
"arthropoda",
"circadian",
"rhythms",
"biochemistry",
"rna",
"cellular",
"neuroscience",
"eukaryota",
"cell",
"biology",
"nucleic",
"acids",
"anatomy",
"physiology",
"neurons",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"ocular",
"system",
"non-coding",
"rna",
"organisms"
]
| 2019 | miR-210 controls the evening phase of circadian locomotor rhythms through repression of Fasciclin 2 |
Expanded polyglutamine ( polyQ ) proteins are known to be the causative agents of a number of human neurodegenerative diseases but the molecular basis of their cytoxicity is still poorly understood . PolyQ tracts may impede the activity of the proteasome , and evidence from single cell imaging suggests that the sequestration of polyQ into inclusion bodies can reduce the proteasomal burden and promote cell survival , at least in the short term . The presence of misfolded protein also leads to activation of stress kinases such as p38MAPK , which can be cytotoxic . The relationships of these systems are not well understood . We have used fluorescent reporter systems imaged in living cells , and stochastic computer modeling to explore the relationships of polyQ , p38MAPK activation , generation of reactive oxygen species ( ROS ) , proteasome inhibition , and inclusion body formation . In cells expressing a polyQ protein inclusion , body formation was preceded by proteasome inhibition but cytotoxicity was greatly reduced by administration of a p38MAPK inhibitor . Computer simulations suggested that without the generation of ROS , the proteasome inhibition and activation of p38MAPK would have significantly reduced toxicity . Our data suggest a vicious cycle of stress kinase activation and proteasome inhibition that is ultimately lethal to cells . There was close agreement between experimental data and the predictions of a stochastic computer model , supporting a central role for proteasome inhibition and p38MAPK activation in inclusion body formation and ROS-mediated cell death .
A hallmark feature of human neurodegenerative diseases is the accumulation of misfolded or otherwise abnormal proteins which become concentrated into large aggregates . Inclusion bodies are large nuclear or cytoplasmic protein aggregates whose predominant constituents may be characteristic of particular diseases . In many cases inclusion bodies ( IB ) are immunoreactive for ubiquitin and proteasome components [1] , indicative of abortive or incomplete proteolysis . The sustained expression of mutant protein with the propensity to misfold may ultimately overwhelm the ubiquitin/proteasome system ( UPS ) and promote the formation of inclusions . This process may be accelerated by an age-related decline in UPS efficiency ( discussed in [2] ) , which may explain why genetically transmitted neurodegenerative disorders typically affect older individuals . Consistent with the proteasome impairment hypothesis , IB form in the neurons of mice in which proteasome function has been genetically compromised [3] . Because misfolded , damaged , or genetically abnormal proteins are aggregation-prone their sequestration into inclusion bodies may actually alleviate the load on the UPS and promote neuronal survival , at least in the short term . Time lapse microscopy of a fluorescent proteasome reporter in cultured neurons has indicated that the UPS load is partially alleviated upon IB formation [4] , and there is evidence that cultured cells forming such inclusions have a survival benefit [5] over the course of the experiment . In the longer term , however , it is possible that deleterious effects from IB formation would become pronounced . Apart from potential physical perturbations imposed by large proteinaceous inclusions ( in axons , for example ) these entities may wreak havoc by depleting essential cellular components ( reviewed in [6] ) or by biochemical means . In Huntington's disease , IB form when a polyglutamine tract in the N-terminal region of the huntingtin protein exceeds the threshold length of approximately forty glutamine residues; early onset and severe disease are correlated with very long tracts , whereas huntingtin proteins with polyglutamine tracts shorter than the threshold do not form IB and are not pathogenic [7] . The nuclear IB formed by the mutant huntingtin protein are generators of reactive oxygen species [8] , and expression of such an expanded polyglutamine protein results in sustained and ultimately cytotoxic activation of p38MAPK [9] . It is likely that proteasome inhibition , ROS generation , and p38MAPK activation all feature in the death of cells containing IB , but their relative importance and potentially complex interdependencies are poorly understood . We have combined live cell imaging with mathematical modeling to explore such relationships . Our data point to a positive feedback loop between IB formation and p38MAPK activation that likely involves ROS . The existence of this loop is supported by the close agreement of laboratory data and simulations generated by a stochastic computer model .
We have previously demonstrated sustained activation of p38MAPK in cultured mammalian cells expressing expanded polyglutamine proteins and in a transgenic mouse model of expanded-polyglutamine disease [9] . This activation could be abrogated by treatment with the specific p38MAPK inhibitor SKF86002 . To confirm the activation of p38MAPK and its inhibition by SKF86002 we performed western blot analysis of U87MG cells expressing HttQ103 ( an amino terminal fragment of the human huntingtin protein containing a 103 glutamine tract ) . Expression of the expanded polyglutamine protein resulted in extensive phosphorylation of HSP-27 , a downstream target of p38MAPK ( Figure 1A ) . Reduced phosphorylation of HSP-27 was detected in cells expressing HttQ25 and HSP-27 phosphorylation was undetectable in untransfected control cells . Although the HttQ25 protein contains a polyglutamine tract below the threshold length for pathogenicity , its overexpression upon transfection may be sufficient to activate p38MAPK at a low and nontoxic level as we have documented previously [9] . The phosphorylation of HSP-27 was precluded by pre-treatment of transfected cells with a pharmacological inhibitor of p38MAPK ( Figure 1A ) or by co-expression with a dominant-negative ( kinase dead ) variant of p38MAPK ( Figure S1 ) confirming that the phosphorylation of HSP-27 was due to the activation of the p38MAPK pathway by Htt103 . To determine if blockade of p38MAPK activity affects cell survival , we transfected U87MG cells with either HttQ25 , HttQ103 or a GFP control plasmid . At 30 hours post-transfection , flow cytometry was performed using propidium iodide exclusion , and revealed that HttQ103 expressing cells exhibited the highest levels of death ( 25% , Figure 1B ) whereas the death associated with Htt25 expression was similar to that of GFP ( 10–12% ) . Pre-treatment of HttQ25 and HttQ103cells with SKF86002 resulted in a decrease in cell death that was most pronounced in cells expressing HttQ103 ( Figure 1B ) . The HttQ103 protein is known to inhibit proteasome activity in a cell-based assay [10]; to determine if the cytotoxicity of the expanded-polyglutamine proteins could be enhanced by further proteasome inhibition a pharmacological proteasome inhibitor ( PI ) was added to HttQ25 and HttQ103 transfected cells . Cells were pre-treated with PI for 6 hours prior to assessing cellular death by flow cytometry . PI-treated U87MG cells expressing HttQ103 exhibited a significant increase in cell death when compared to their untreated counterparts ( Figure 1B ) , and was 15% greater than PI treated HttQ25 expressing cells . Under the same experimental conditions the amount of cell death induced by PI in untransfected cells was approximately 5% ( not shown ) , roughly equivalent to the increase in cell death mediated by PI in cells expressing HttQ25 and HttQ103 . Under our conditions proteasomes cannot therefore be fully inhibited by expression of the polyQ proteins alone . The pharmacological data support the argument that the cytotoxicity of misfolded proteins is mediated by proteasome inhibition and p38MAPK activation , but do not reveal whether these activities are independent . To simultaneously assess functioning of the UPS and inclusion body ( IB ) formation at a single cell level we created a bicistronic construct that encodes an expanded polyglutamine protein and fluorescent proteasome substrate on the same transcript ( see schematic diagram in Figure 2A ) . This construct was designed to investigate the temporal order of events leading to HttQ103 induced cellular and proteasome toxicities and to dissect the role of p38MAPK in these events . U87MG cells were transfected with HttQ103YFP-pIRES-mRFPu and were imaged at 10 minute intervals from 24 to 48 hours post-transfection . The time lapse images revealed formation of IB at 36 hours post-transfection in cells expressing HttQ103 ( Figure 2B , first panel; time lapse videos are provided as Videos S1 to S8 ) . Values of mRFPu intensity were graphed as a function of time revealing an increase in mRFPu fluorescence prior to the formation of an IB , followed by a period of constant mRFPu intensity . The single cell analysis revealed that HttQ103-induced cellular death is preceded by gradual UPS impairment . Once this impairment reaches a threshold level , IBs begin to form and their formation correlates with a momentary recovery of UPS efficiency as measured by mRFPu intensity . These results are consistent with previously published findings in primary neuron cultures [4] . To examine the extent to which IB formation was dependent on UPS dysfunction U87MG cells expressing HttQ103YFP-pIRES-mRFPu were treated with proteasome inhibitor . As expected PI treatment resulted in a rapid and persistent increase in mRFPu intensity . Under these conditions IB formation was accelerated relative to untreated controls ( Figure 2B , second panel ) . These data suggest an immediate relationship between proteasomal inhibition and IB formation . Damaged proteins are normally eliminated by the UPS , and we speculated that an increase in cellular ROS levels would lead to oxidative damage to proteins and inflict an additional burden on proteasomes that may affect the kinetics of IB formation . To test this hypothesis , we depleted reduced glutathione levels in HttQ103YFP-pIRES-mRFPu tranfected cells by treating cells with buthionine sulphoximine ( BSO ) at 24 hours post-transfection . In BSO treated cells we observed a constant increase in mRFPu intensity ( Figure 2B , third panel ) consistent with a cumulative UPS burden . Having previously established that the activation of p38MAPK in HttQ103 expressing cells contributes to cytotoxicity , we sought to determine what effects inhibition of p38MAPK signalling pathway would have on IB formation and UPS dysfunction . U87MG cells expressing HttQ103YFP-pIRES-mRFPu were therefore treated with SKF86002 . These cells exhibited a low level of mRFPu fluorescence along with a delay in IB formation ( Figure 2B , fourth panel ) . The mRFPu fluorescence remained low and did not feature a rapid increase as seen in the untreated counterparts . These data suggest that inhibition of the p38MAPK pathway decouples the proteasome inhibition from HttQ103 protein expression , resulting in delayed formation of IB . We quantified IB formation by recording the number of cells with IB at 6 hour intervals starting at 24 hours post-transfection . The percentage of cells with IB was graphed as a function of time ( Figure 3A ) . PI treatment generated the greatest number of IB compared to untreated cells while SKF86002-treated cells were found to have the fewest IB . Comparative single-cell analysis of IB formation corresponded to time-points in which 40% of transfected cells had formed IB . Similarly , we quantified average mRFPu fluorescence from many cells and graphed the fluorescence intensities as a function of time . Treatment with PI and BSO generated the highest amounts of mRFPu fluorescence , while SKF86002 treatment resulted in the lowest levels of mRFPu in comparison to untreated cells ( Figure 3B ) . Findings from multiple cells were consistent with the mRFPu fluorescence levels observed in the single cell analysis . We have previously used stochastic computer modeling to study the age-related decline of proteolysis [11] , and have adapted this model to incorporate p38MAPK in an effort to better understand polyglutamine-mediated cell death . Our objective was to determine if a relatively simple mathematical model incorporating the components thought to be critical for polyQ-mediated cytotoxicity could recapitulate our laboratory findings; if not some critical component must have been overlooked or one or more of the starting assumptions must be invalid . Conversely , a good fit would suggest that the assumptions are valid and no critical elements have been overlooked . The stochastic computer model is represented schematically in Figure 4 . The model was constructed using the Systems Biology Markup Language as described in the Materials and Methods section; details of molecular species and reactions are given in Text S1 . The model predicted that treatment with PI would lead to reduced cell death at 30 hours ( Figure 5A ) , a short term benefit from reduced levels of small aggregates binding to proteasomes ( the consequence of which would be reduced by concentration of aggregates into IB ) . The experimental data , however , indicated a slight increase in the proportion of cell deaths under PI suggesting that the situation is more complex than is currently accounted for in the model . On the other hand , the model predicted that inhibition of p38MAPK activity should lead to much lower cell death , in close agreement with the experimental data . Also in agreement with the experimental findings described above , the model predicted that proteasome inhibition should lead to an increase in the rate of IB formation compared to untreated cells ( Figure 5B ) . When p38MAPK activity is inhibited the model predicts a lower rate of IB formation at early time-points ( Figure 5B ) although from 36–48h , the rate of increase is similar to untreated cells ( note lines are parallel for polyglutamine and p38MAPK inhibition during this time interval ) . The later increase in inclusion formation when p38MAPK activity is inhibited is probably due to ROS generation via the aggregated protein . The computer model predicts that much higher levels of mRFPu will be observed in PI treated cells than in untreated cells , as expected for a proteasome substrate ( Figure 5C ) . Critically , inhibition of p38MAPK activity reduced the accumulation of mRFPu in the computer model . Overall the simulations and experimental data are in good agreement , indicating that the model describes the important molecular relationships of the system as portrayed in Figure 4 . Based on our single-cell analysis data , we speculated that the activation of p38MAPK and UPS impairment were contributing to the production of reactive oxygen species ( ROS ) and that SKF86002 may be counteracting this cellular response . A genetic approach was adopted to test this hypothesis , utilizing transfection of cells with expression vectors encoding wild type or kinase dead versions of p38MAPK . Whereas pharmacological inhibition may affect multiple p38MAPK isoforms any modulation of cellular response observed with the genetic approach would be attributable to the alpha isoform of p38MAPK exclusively . We first confirmed that in our U87MG cell system the expression vectors were capable of modulating p38MAPK activity using phosphorylation of HSP27 as a proxy marker ( Figure S1 ) . Cells were then transfected with the p38MAPK expression vectors or with an empty vector control construct and lysates were assayed for reduced glutathione ( GSH ) content , a marker of oxidative status within cells . We found that overexpression of wild type p38MAPK resulted in a significant decrease in GSH levels whereas the GSH content in cells expressing kinase dead p38MAPK was less affected in comparison to controls ( Figure 6A ) . To determine if overexpression of wild type p38MAPK resulted in dysfunction of the UPS , we transfected the p38MAPK expression constructs into a cell line stably expressing GFPu , a well characterized proteasome sensor [10] . For these experiments we made use of a stable NIH 3T3 cell line we had previously generated; the GFPu reporter is expressed at a lower level in the stable line and does not accumulate as an artefact of transfection-mediated overexpression . By flow cytometric analysis , we found that cells overexpressing wild type p38MAPK exhibited the highest levels of GFPu intensity whereas the levels of GFPu in cells expressing the kinase dead p38MAPK or the empty vector were similar ( Figure 6B ) . These data suggested that the activation of p38MAPK was negatively affecting UPS function . To test whether this inhibition of the proteasome was a direct affect of p38MAPK activation , we transfected cells with the same set of plasmids and assayed their proteasome activity using fluorogenic substrates . No significant differences were noted ( Figure S1 ) . Taken together , the data indicate that the expression of p38MAPK does affect the oxidative status of cells but does not directly inhibit the proteasome . Since aggregated protein leads to increased levels of ROS , inhibition of p38MAPK may simply delay cell death and thereby allow more aggregation to take place . To determine the predicted outcome should p38MAPK not be involved in generating more ROS we removed the reaction for ROS generation via p38MAPK from the model and repeated the computer simulations . Without p38MAPK-generated ROS less cell death was predicted under all conditions ( Figure 7A ) . With the feedback loop broken in this way the model also predicted that there would be no significant difference in the numbers of inclusions at each time point between untreated cells and cells treated with a p38MAPK inhibitor ( Figure 7B ) . We also refitted the model for HttQ103 without treatments and no feedback loop using the experimental data for cell death and inclusion formation ( Figure 7C–D ) . Since the model predicted less cell deaths and lower levels of inclusions than the data , we increased the parameters for inclusion formation ( kaggPolyQ ) and cell death ( kp38death and kPIdeath ) . We also increased the parameter for ROS generation via p38MAPK ( kgenROSp38 ) since this had the effect of increasing both the levels of inclusions and the number of cell deaths . We then ran the model with the treatment for proteasome inhibition and p38MAPK inhibition . The model was not able to reproduce the decline in cell death or the lower levels of inclusions when p38MAPK is inhibited ( evident from the comparison of Figure 7C–D with Figure 5A–B ) . Therefore the model indicates that a feedback loop from p38MAPK to ROS is required to explain the experimental data .
Based on the experimental data and the computer simulations reported herein we propose a vicious cycle mechanism of polyglutamine-mediated cytotoxicity ( Figure 8 ) . In the proposed mechanism the initiating event is the inhibition of the proteasome by small aggregates of misfolded protein . It has been previously shown that proteasome inhibition leads to the generation of ROS ( reviewed in [12] ) and the activation of MKK3 and MKK6 kinases upstream of the stress kinase p38MAPK [13] , [14] . Activation of p38MAPK by its upstream regulators may exacerbate the problem by promoting downstream ROS production ( through a mechanism discussed below ) . By damaging other cellular proteins the reactive oxygen would provide an additional burden to the UPS , which is normally charged with the proteolytic degradation of damaged or abnormal proteins . Increasing proteasome inhibition would lead to further accumulation of misfolded proteins , ultimately coalescing into inclusion bodies . By reducing local concentrations of the small aggregates that are thought to be most inhibitory to the proteasome the IBs may temporarily alleviate proteasome inhibition , but by concentrating iron they may promote the further generation of ROS , ensuring yet more damage and the conditions that will sustain p38MAPK activation . Whether or not there is self-amplification of the vicious cycle as a consequence of increasing ROS production we postulate that it is sustained p38MAPK activity and proteasome inhibition that will ultimately lead to cell death . The mathematical model was based on our previous model of the ubiquitin/proteasome system which we modified and extended to include turnover and aggregation of polyglutamine proteins . The model predictions were in close agreement with the experimental data indicating that the proposed network in Figure 4 captures the important components in the system . However , we found that there were some discrepancies between the model predictions and experimental data regarding cell death at early time-points under conditions of proteasome inhibition . This suggests that there is something missing from the model . Since proteasome inhibition affects all protein turnover , the missing part could be a pro-apoptotic protein . One such candidate is the transcription factor p53 which is normally rapidly turned over by the proteasome . A future extension of our current model to include a pro-apoptotic protein would be fairly straightforward as we have previously modelled the p53 system [15] . An advantage of modelling over laboratory experiments is that it is easy to manipulate the system on a computer , whereas the same experiments in the laboratory may be impossible or very costly to do . For example , although it would be difficult to prove experimentally that ROS generation by p38MAPK is required to explain the experimental data , the hypothesis could easily be tested in the mathematical model by simply removing the reaction of p38MAPK-dependent ROS generation and repeating the simulations . The model confirmed that p38MAPK is involved in generating more ROS since with the feedback loop broken the model output no longer fitted the experimental data . We have previously shown that pharmacological blockade of p38MAPK can protect cells from polyglutamine-mediated cell death [9] , and the current experimental data and mathematical modeling provide an explanation for the efficacy of this intervention: by breaking the vicious cycle the p38MAPK inhibitor precludes further damage and proteasome inhibition . The data we present herein support a central role for ROS in the proposed cycle of p38MAPK activation , proteasome inhibition , and protein aggregation that ultimately leads to cell death . The importance of ROS in protein misfolding disorders is not a new concept; indeed the cytotoxicity of elevated ROS generated by mutant huntingtin has been convincingly demonstrated by the laboratory of Rubinsztein [16] . Intriguingly , the proteasome is itself an important regulator of oxidative damage in neurons and proteasome inhibition is known to induce mitochondrial dysfunction and promote oxidative damage to DNA and protein ( reviewed in [12] ) . Once some threshold of polyQ-mediated proteasome inhibition is reached it seems entirely plausible that a self-perpetuating loop of ROS generation and p38MAPK activation could lock the cell into a dysfunctional state . Indeed , a directly analogous positive feedback loop involving ROS and p38MAPK activation was recently proposed for the induction of senescence in mammalian cells . By combining bioinformatics , stochastic computer modeling , and direct experimental interventions Passos et al . have provided convincing evidence that sustained activation of p38MAPK is required for generation of mitochondrial ROS , which by generating DNA damage ensures the continued activation of p38MAPK in senescent cells [17] . The ROS generated as a consequence of proteasome inhibition also appears to be of mitochondrial origin [18] , so the cascade of events in cells expressing misfolded protein may be very similar to that in cells in which the initiating event is exposure to ionizing radiation ( as was the case in the Passos paper ) . If a similar loop were operating in cells expressing expanded polyglutamine proteins one might expect to find ROS-mediated DNA damage leading to activation of ATM and phosphorylation of histone H2AX , as has indeed been documented in cells from Huntington's disease and SCA-2 patients [19] . The activation of ATM and formation of H2AX repair foci appears to precede the formation of IB [20] , but may be coincident with p38MAPK activation and the generation of ROS . If our model is correct pretreatment of cells with the p38MAPK inhibitor should reduce the number of H2AX foci in cells expressing HttQ103 . We are currently testing this hypothesis . Enhanced levels of autophagy may also contribute to the protective effects of the p38MAPK inhibitor we have observed . Although the effects may be dependent on cell type , there is evidence that the alpha isoform of p38MAPK inhibits autophagy [21] [note that this is the same isoform utilized in our genetic experiments , and may be the target of primary importance to all of the interventions described herein] . Low level inhibition of the proteasome is known to promote autophagy [22] , but as proteasome inhibition increases the activation of p38MAPK may limit autophagic clearance of protein aggregates . Enhancement of autophagy has been proposed as a therapeutic strategy for the treatment of polyglutamine disorders such as Huntington's disease [23] , and pharmacological blockade of p38MAPK may provide benefit by multiple mechanisms . It may break the vicious cycle of ROS generation while promoting autophagy-mediated clearance of aggregates in cells already compromised for proteasome function . We have performed the current set of experiments in U87MG cells , a human glioblastoma cell line . These cells were chosen for their ease of handling , including their high transfection efficiency . Because the critical components of the proposed vicious cycle are universally present in mammalian cells ( including p38MAPK , proteasomes , and ROS from mitochondria ) it is likely that a self-perpetuating loop could be triggered by expanded polyglutamine proteins in any cell type , and we have previously demonstrated the protective effect of a p38MAPK inhibitor in normal and transformed cells of both human and mouse origin [9] . We nevertheless recognize that the kinetics of the events we have described may be markedly different for neurons in situ . Although the short term benefit of IB formation ( through improved proteasome function ) has been demonstrated in cultured neurons [4] , [5] the longer term implications of IB in neurodegenerative diseases must be considered . A transient alleviation by IB of the polyglutamine-mediated proteasome burden was recently demonstrated in vivo in an inducible model of Huntington's disease [24] , but no long term impairment of proteasome function was observed in the brains of these mice . If the IB concentrates iron and promotes ROS generation through Fenton chemistry [8] the IB , once formed , might ensure the perpetuation of the vicious cycle proposed in Figure 3C . The failure to detect this effect in older mice expressing a polyQ protein may relate to the inability of the reporter system to detect less than a 40% decrease in proteasome inhibition [24] or may relate to differences between the mouse model and the human disease . In human polyglutamine disorders the disease may progress over years if not decades , and the later consequences of IB formation might therefore be more severe . Clinical evidence supports a deleterious role for IB in human disease . In polyglutamine repeat diseases such as HD , frontotemporal lobar degeneration , and RNA-mediated diseases such as myotonic dystrophy , inclusion-mediated titration of transcription factors ( like CBP ) , tar DNA-binding protein 43 ( TDP-43 ) , and RNA splicing factors ( for example muscleblind ) , respectively , may represent an important molecular mechanism of disease [6] . Added to these effects would be the deleterious effects we ascribe to the ROS-mediated vicious cycle .
The pEGFP-N1 expression construct was purchased from Clontech ( Palo Alto , California , USA ) . Htt-Q25 and Htt-Q103 expression constructs containing a synthetic insert encoding exon 1 or human Huntingtin containing a polyglutamine tract of either 25Q or 103Q fused to the yellow fluorescent reporter protein ( YFP ) were generous gifts from Dr . Ron Kopito ( Stanford University ) . These are designated HttQ25YFP or HttQ103YFP . A red fluorescent proteasome reporter was generated by PCR-mediated transfer of the degron sequence from the GFPU reporter ( Bence et al . ; the gift of Dr . Ron Kopito ) to the C terminus of the monomeric red fluorescent protein ( the gift of Dr . Robert Campbell , University of Alberta ) . Under normal conditions , mRFPu is quickly degraded by the 26S proteasome , but during conditions of proteasomal impairment , turnover of mRFPu is reduced , leading to an accumulation of mRFPu that is visible by fluorescent microscopy . To simultaneously express the expanded YFP-tagged polyglutamine proteins and the red fluorescent proteasome reporter the former was inserted into NheI site upstream of the internal ribosome entry site ( IRES ) in the vector pIRES ( Clontech , Palo Alto , California , USA ) and the latter was inserted between the Xba I and Sal I sites downstream of the IRES element . The wild type and kinase dead p38MAPK variants were generous gifts from Dr . J . Han ( The Scripps Research Institute , La Jolla , CA ) . The hyper-active p38MAPK construct was a gift from Dr . Oded Livnah ( The Hebrew University of Jerusalem ) . The human U87MG glioblastoma cells ( a gift from Dr . I . Lorimer at the Ottawa Hospital Research Institute ) were maintained in Dulbecco's modified Eagle's medium ( DMEM ) and supplemented with 10% FBS and maintained in a 37°C incubator with 5% CO2 . For transient transfections , cells were plated in either 96- or 6 well dishes 24hours prior to transfections . Subsequently , they were transfected using GeneJuice Transfection Reagent ( Novagen , Madison , WI , USA ) as per the supplier's protocol . 0 . 5µg or 3 . 0µg of plasmid DNA was used in each well of a 96 or 6 well dish . For p38MAPK inhibition experiments , cells were pre-treated for 2h with 20µM SKF86002 ( Calbiochem ) prior to transfection with various expression constructs . U87MG cells were harvested in protein lysis buffer consisting of 100mM Tris pH 6 . 8 , 20mM DTT , 4% SDS , 5% glycerol . Protein concentrations were determined using the Bradford assay reagents ( Bio-Rad , Hercules , CA , USA ) . Reduced proteins were resolved on a 10% SDS-polyacrylamide gel and electro-blotted onto a Hybond C nitrocellulose membrane ( Amersham Bioscience Corp , Baie d'Urfé , QC ) . The membranes were stained with Ponceau S prior to immunoblotting with phospho-HSP-27 ( polyclonal rabbit ) , total HSP-27 ( rabbit polyclonal ) ( Cell Signaling , Danvers , MA ) , or Actin ( Sigma-Aldrich ) . Proteins were detected using the HRP method and SuperSignal West Pico Chemiluminescent Substrate reagents ( Pierce , Rockford , IL , USA ) . Proteins were visualized using the GeneGnome ( Syngene , Frederick , MD , USA ) . Cell viability was assessed by flow cytometry using propidium iodide exclusion . Adherent and non-adherent cells were transfected with various constructs for 30 hours , harvested and stained with Propidium Iodide . For p38MAPK inhibition experiments , cells were pre-treated for 2h with 20µM SKF86002 ( Calbiochem ) prior to transfection with various expression constructs . For proteasome inhibition experiments , cells were treated with Proteasome Inhibitor I ( Calbiochem ) 24h post-transfection at a final concentration of 25µM . Fluorescent detection was analyzed by flow cytometry using a Beckman Coulter Quanta SC MPL . Data and analysis were done using Quanta Analysis software ( Beckman Coulter , Inc . , Brea , CA , USA ) . 75 000 U87MG glioblastoma cells were seeded onto a Delta T4 culture dish system ( Bioptechs , Butler , PA ) and maintained in a 37°C incubator with 5% CO2 for 24h hours . Cells were transfected with 2ug of plasmid DNA encoding HttQ103YFP-pIRES-mRFPu for 24 hours before being transferred onto a heated stage maintained at 37°C and at 5% CO2 using a Delta T4 culture dish temperature controller and cell perfusion system ( Bioptechs , Butler , PA ) . For p38MAPK inhibition experiments , cells were pre-treated for 2h with SKF86002 for a final concentration of 20µM to preclude kinase activation upon transfection . For proteasome inhibition experiments , cells were treated with Proteasome Inhibitor I ( Calbiochem ) 24h post-transfection at a final concentration of 25µM ( treatment with PI prior to transfection is not possible due to its immediate toxicity ) . For buthionine sulphoximide ( BSO ) -induced depletion of glutathione experiments , cells were treated 24h post-transfection with BSO ( Sigma ) 24h post-transfection at a final concentration of 5mM . Microscopy was performed 24 hours post-transfection on a Zeiss Axiovert 200M inverted fluorescent microscope for a total of 24 hours . Fully automated multidimensional acquisition was controlled using Axiovision 4 . 8 software . Images were acquired using a 10× objective ( EC Plan-Neofluar ) with a side-mounted AxiocamHRm camera . Yellow fluorescent protein or red fluorescent protein was excited using the Zeiss Colibri LED illumination system ( LEDmodule 505nm or LED module 590nm ) and detected using the appropriate filters ( 46HEYFP or 61HEGFP/HcRED , respectively ) . Fixed exposure times were as follows: Brightfield phase contrast 1ms; YFP 100ms; RFP 188ms . Images were taken at 10 minute intervals for 48 hours and compiled into video files using Axiovision 4 . 8 software ( Carl Zeiss , Thornwood , NY ) . Cell lysates from U87MG cells over-expressing wild type p38MAPK or kinase dead p38MAPK were analyzed for reduced GSH content using a luciferase kit ( GSH-Glo ) from Promega ( Madison , WI ) . 10 , 000 cells were seeded in 96-well plates and transfected with 0 . 5µg of DNA for 48 hours . Cells were collected and analyzed for GSH following the manufacture's protocol . Mouse NIH 3T3 cells were co-transfected with GFPu and a PGK-driven puromycin resistance gene ( gift of Dr . M . McBurney , Ottawa Hospital Research Institute ) . Cells stably expressing the proteasome reporter GFPu were selected over 2 weeks in a final concentration of 2 . 0µg/ml . For proteasome assays , wild type p38MAPK , kinase dead p38MAPK , or pcDNA were transfected into 3T3-GFPu cells for 48 hours . Cells were collected and analyzed for GFPu expression by flow cytometry ( Beckman Coulter Quanta SC MPL ) . Mean GFP intensity was analyzed using the Quanta Analysis software and subsequently graphed using Excel ( Microsoft ) . Statistical significance was determined by a two tailed Student's t-test . Unless otherwise indicated , values were considered significant when p<0 . 05 . The model was developed to mimic the experimental system so that simulations could be performed to see which parameters affected the different cellular outcomes . The model was initially fitted to experimental data where HttQ103 had been added to cells but without any inhibitors . The model was then used to mimic the experimental treatments and the model predictions were compared to the experimental results . If there were discrepancies between the model predictions and experimental results , the model was modified ( either by changing parameter values or by the addition of further reactions ) . The model was then re-run for the experiment without any treatments to check that it still fitted the experimental data . If the model did not fit , further adjustments were made and the procedure repeated . We originally started with a model that did not contain a feedback loop from p38MAPK to ROS but found that it was necessary to include this loop in order to get the model to fit both the data for the treatment with p38MAPK inhibitor and the data without the treatment . Further details are given in Text S1 . The model was encoded in the Systems Biology Markup Language ( SBML ) as this standard allows models to be easily shared , modified and extended [25] . SBML is a way of representing a network of interactions so that it can be simulated on a computer and the evolution of the system over time can be followed . SBML shorthand was used to create the SBML code which was then converted into full SBML [26] . The network diagram is given in Figure 4 , and Tables S1 and S2 in Text S1 give details of the species and reactions respectively . Text S1 also contains more detail of events and parameter values ( Tables S3 and S4 in Text S1 ) . Since there is large variability in cellular outcomes in terms of both inclusion formation and cell death , we used stochastic simulation . This was based on the Gillespie algorithm [27] which assumes that collisions of molecules occur within a reaction vessel and that at most only two molecules can collide . We chose this method of simulation as there are low copy numbers of many of the species and random effects play a major role in this model , as can be seen by the cell to cell variability of the model output . It should be noted that we have a few reactions which have more than two molecules in the list of reactants . These are the reactions for the aggregation of polyQ where we assume that ROS affects the reaction kinetics although ROS itself is not consumed by the reaction ( and so also appears in the list of products ) . We use a function of ROS in the kinetic law so that we have a pseudo second order reaction rather than a third order reaction . Although this may seem to be a violation of the assumptions of the Gillespie algorithm , this provides a simple way to allow for the effects of ROS on the aggregation process rather than adding many more reactions and parameters . The model is available from the Biology of Ageing e-Science Simulation and Integration ( BASIS ) system ( [28] , [29] ) and the Biomodels database ( ID:MODEL1002250000 ) [30] . We assume that the addition of the polyQ gene to the cell resulted in continuous synthesis of the polyQ protein . It is also degraded by the proteasome so that total levels remain fairly constant with a half-life of about 20 hours [31] . We set the levels of polyQ sythesis and degradation so that the half-life would be 20 hours if the proteasome did not become inhibited by aggregates . Two molecules of polyQ interact to form a small aggregate ( AggPolyQ1 ) . We assume that ROS affects the aggregation kinetics if it rises above basal levels . The aggregate can grow in size by the addition of further polyQ proteins in a reversible manner . However , when the aggregate reaches a certain threshold size , we assume that disaggregation can no longer take place and that instead an inclusion forms ( SeqAggP ) . This threshold represents the seed and is assumed to be of size six based on data for amyloid fibril polymerization [32] . Since mutant huntingtin forms amyloid-like filaments , it is reasonable to assume that it has similar aggregation kinetics [33] , [34] . A very recent study shows that mutant huntingtin forms three major pools: monomers , oligomers and inclusion bodies [35] . Interestingly , the study showed that the pool of oligomers as a proportion of total huntingtin did not change over a time period of 3 days despite continued conversion of monomer to inclusion bodies . We also compared the levels of oligomers ( represented by the species AggPolyQ[i] , where i = 1–5 ) in our simulation output in cells which formed inclusion bodies ( Figure S2 ) and discuss the results in the Text S1 section . It has been shown that small aggregates bind to proteasomes and inhibit proteasomal function [36] . Therefore , we assume that AggPolyQ can bind to the proteasome and so reduce the pool of available proteasomes . However we assume that inclusions do not interfere with the degradation machinery . We also include a species to represent mRFPu and assume that this is turned over with a half-life of about 30 minutes [36] . We assume that ROS is continuously generated and removed with a half-life of 1 hour [37] but that basal levels are low ( about 10 molecules ) . We assume that small aggregates may generate ROS , so that the level of ROS is dependent on the amount of small aggregates ( either bound to the proteasome or free pools ) . We also assume that the presence of inclusions will increase levels of ROS but with a much smaller effect than small aggregates . We represent p38MAPK in two forms: unphosphorylated ( p38 ) and phosphorylated ( p38-P ) with p38-P being the active state . We assume that high levels of ROS activate p38MAPK and that high levels of p38-P initiate a signalling cascade that results in cell death . We set the rate of this reaction so that it is unlikely to occur when p38-P levels are low and the probability of the reaction occurring increases with increasing levels of p38-P . However , since the model is stochastic , it is possible that even low levels of p38-P will occasionally signal for cell death . We also assume that if the level of proteasomes bound by aggregates increases above a threshold of about 50% , then another signalling pathway leads to cell death due to the accumulation of the pro-apoptotic protein p53 . As in the p38 death pathway , cell death due to aggregates inhibiting the proteasome may occur even when levels of AggP Proteasome are fairly low . The reactions for the cell death pathways are shown in Table S2 in Text S1 . After cell death occurs , a dummy parameter kalive is set to zero to prevent further reactions occurring , and a dummy species to record the cause of cell death is set to 1 . This makes it possible to plot the time of cell death , the cause of death and to count the number of cell deaths of each type in multiple simulations . We also assume that proteasomes bound by AggP , polyQ or mRFPu may be sequestered into inclusions if degradation does not take place . If AggP is sequestered into inclusions , then this will help alleviate the increase in ROS due to protein aggregation , since we assume that small aggregates lead to greater ROS generation than inclusions . We also include a generic pool of protein ( NatP ) which can misfold to become ( MisP ) . We assume that misfolded protein can be either refolded , ubiquitinated and degraded or at high concentrations it may start to aggregate . Once an inclusion forms , misfolded protein may be sequestered into the inclusion body , including MisP bound to proteasomes . | Neurodegenerative diseases feature concentration of misfolded or damaged proteins into inclusion bodies . There is controversy over whether these entities are protective , detrimental , or relatively benign . The formation of inclusion bodies may be accelerated by inefficient protein degradation and may promote activation of stress signalling pathways . Each of these events may promote the generation of reactive oxygen species which may exacerbate the problem by damaging more proteins , possibly damaging components of the UPS itself , but in either case further impeding the function of cellular proteolytic systems . To determine how these events are related and which are critical , we generated a live cell imaging system in which inclusion formation and proteolytic efficiency can be evaluated , and created a stochastic computer model incorporating the same components . Laboratory data and computer simulations were found to be in close agreement , supporting a mechanism wherein misfolded protein induced a vicious cycle of stress kinase activation , ROS generation , and proteasome inhibition which was ultimately cytotoxic . Inclusion body formation partially alleviated the burden on the proteolytic system , but may not provide long term benefit . Pharmacological blockade of a stress-activated kinase was effective in breaking the vicious cycle , as predicted by the computer model and confirmed experimentally . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"molecular",
"biology/post-translational",
"regulation",
"of",
"gene",
"expression",
"cell",
"biology/cellular",
"death",
"and",
"stress",
"responses",
"neurological",
"disorders",
"biochemistry/cell",
"signaling",
"and",
"trafficking",
"structures",
"biochemistry/macromolecular",
"assemblies",
"and",
"machines",
"computational",
"biology/systems",
"biology"
]
| 2010 | Experimental and Computational Analysis of Polyglutamine-Mediated Cytotoxicity |
The inhibitory receptor programmed death-1 ( PD-1 ) has the capacity to maintain peripheral tolerance and limit immunopathological damage; however , its precise role in fulminant viral hepatitis ( FH ) has yet to be described . Here , we investigated the functional mechanisms of PD-1 as related to FH pathogenesis induced by the murine hepatitis virus strain-3 ( MHV-3 ) . High levels of PD-1-positive CD4+ , CD8+ T cells , NK cells and macrophages were observed in liver , spleen , lymph node and thymus tissues following MHV-3 infection . PD-1-deficient mice exhibited significantly higher expression of the effector molecule which initiates fibrinogen deposition , fibrinogen-like protein 2 ( FGL2 ) , than did their wild-type ( WT ) littermates . As a result , more severe tissue damage was produced and mortality rates were higher . Fluorescence double-staining revealed that FGL2 and PD-1 were not co-expressed on the same cells , while quantitative RT-PCR demonstrated that higher levels of IFN-γ and TNF-α mRNA transcription occurred in PD-1-deficient mice in response to MHV-3 infection . Conversely , in vivo blockade of IFN-γ and TNF-α led to efficient inhibition of FGL2 expression , greatly attenuated the development of tissue lesions , and ultimately reduced mortality . Thus , the up-regulation of FGL2 in PD-1-deficient mice was determined to be mediated by IFN-γ and TNF-α . Taken together , our results suggest that PD-1 signaling plays an essential role in decreasing the immunopathological damage induced by MHV-3 and that manipulation of this signal might be a useful strategy for FH immunotherapy .
Although liver transplantation has emerged as an effective therapeutic approach for treating fulminant virus hepatitis ( FH ) , mortality rates associated with FH remain very high worldwide [1] . The recent development of a mouse FH model , based upon infection with the murine hepatitis virus strain-3 ( MHV-3 ) , has provided insights into mechanisms underlying the disease pathogenesis and resulted in some novel treatment strategies [2] . MHV-3 is a single-stranded , positive-sense RNA virus that belongs to the Coronaviridae family . In inbred laboratory mice , the virus produces strain-dependent disease profiles that depend on the infection route , age , genetic background , and immune status of the host . For example , Balb/c , C57BL/6 and DBA/2 mice develop acute fulminant hepatitis , while C3H mice develop a mild chronic disease and mice of the A strain exhibit no evidence of hepatitis [3] , [4] . In contrast to the resistant A strain mice , FH induced by MHV-3 in susceptible mice is characterized by the presence of sinusoidal thrombosis and hepatocellular necrosis [2] , [3] . These pathological findings occur concomitantly with expression of fibrinogen-like protein 2 ( FGL2 ) , a virus-induced procoagulant molecule in the sinusoidal lining cells in the liver . FGL2 has the capacity to promote fibrinogen deposition and subsequently directly induce the coagulation cascades by the expression of procoagulant activity ( PCA ) [5] . Thus , up-regulation of FGL2 is an essential component of the lethal effects of MHV-3-induced FH . Programmed death ( PD ) -1 is an inhibitory receptor expressed on activated T cells , B cells and myeloid cells . PD-1-deficient mice ( Pdcd1−/− ) develop various spontaneous autoimmune diseases , including glomerulonephritis and dilated cardiomyopathy , indicating that this receptor plays a critical role in maintenance of peripheral tolerance [6] . PD-L1 ( B7-H1 ) and PD-L2 ( B7-DC ) , two immunoregulatory molecules belonging to the B7 superfamily , were identified as ligands for PD-1 , engagement of PD-1 with its ligands mediates negative signaling events via recruitment of phosphatases , such as SHP-2 , and dephosphorylation of some effector molecules involved in downstream T cell receptor ( TCR ) signaling [7] , [8] . PD-1 signaling has also been shown to modulate the balance between antimicrobial immune defense and immune-mediated tissue damage . For example , PD-1-deficient mice develop more severe hepatocellular injury than their wild-type ( WT ) littermates in response to adenovirus infection [9] . In a herpes simplex virus ( HSV ) stromal keratitis mouse model , blockade of PD-1 signaling led to increased HSV-1-specific effector CD4+ T cell expansion , IFN-γ production , and exacerbated keratitis [10] . A functionally-significant high level of PD-1 expression has been found to be maintained by exhausted CD8+ T cells in mice chronically infected with lymphocytic choriomeningitis virus ( LCMV ) , in primates exposed to simian immunodeficiency virus ( SIV ) , and in humans suffering from infection with human immunodeficiency virus ( HIV ) , hepatitis B or C virus ( HBV or HCV ) , or human T-lymphotropic virus ( HTLV ) . However , blockade of the PD-1/PD-Ls pathway efficiently restored the virus-specific effector functions of the exhausted T cells , and lead to substantially reduced in viral load [11] , [12] , [13] , [14] , [15] . The PD-1 signal is also known to play a key role in the chronicity of infections with bacteria ( Helicobacter pylori and Schistosoma mansoni ) [16] , [17] , pathogenic fungus ( Histoplasma capsulatum ) [18] , and parasitic worms ( Taenia crassiceps ) [19] . It appears that a number of pathogenic microorganisms exploit the PD-1 signal in order to evade host immune responses and to establish persistent infection . Although the influence of PD-1 signal activity has been studied in several infection models , there are no data available concerning the role of this pathway in FH . To this end , we used the MHV-3-induced mouse FH model to demonstrate that PD-1 signaling acts to limit the immunopathological damage during disease progression . Furthermore , our findings suggested that enhanced PD-1 signaling might represent a useful immunotherapeutic strategy for treating FH .
PD-1 expression has been previously described as being induced on specific cell subsets in response to viral or bacterial infection [20] . Thus , we first determined the status of PD-1 expression at 72 h after MHV-3 infection ( 10 PFU ) by immunohistochemical techniques . PD-1-positive cells were observed in tissues from the thymus , spleen , lymph nodes and liver . Cellular expression was localized to the cell membrane and in the cytoplasma while was completely absent from the nuclear compartment . PD-1-positive cells were distributed throughout the medulla and cortex of the thymus and lymph nodes . In the spleen , PD-1-positive cells were restricted to the germinal center under normal conditions , but extended to the red pulp after infection . In infected liver , more PD-1-positive cells were present in the portal and parenchymal areas , as opposed to the relatively low presence of PD-1-positive cells in only the portal area in phosphate-buffered saline ( PBS ) treated-mice ( Fig . 1A ) . The amount of PD-1-positive cells in the different organs of infected and control mice were counted and compared , results showed that the number of positive cells was significantly higher in infected mice ( Fig . 1B ) . Furthermore , FACS analysis revealed that PD-1 expression was enhanced on multiple subsets of immune cells , including the CD4+ and CD8+ T cells , NK1 . 1+ NK cells and CD68+ macrophages ( Fig . 1C ) . PD-1-positive cells were also observed in the lung , heart and kidney , however , the numbers of PD-1 positive cells in these tissues did not significantly increase in response to MHV-3 infection ( Fig . S1 ) . To investigate the potential role of PD-1 signaling in regulating FH tissue pathology , organs from MHV-3 infected PD-1-deficient ( PD-1 KO ) and WT mice were assessed for morphological differences . Small and discrete foci of necrosis with sparse polymorphonuclear leukocyte infiltration were observed in liver tissues from PD-1-deficient mice after 24 h of infection . In contrast , WT mice exhibited normal liver architecture at this time point . Slight liver damage became apparent in WT mice after 48 h of infection , meanwhile , the damaged areas of PD-1-deficient mice had enlarged and confluent necrosis had become evident . By 72 h of infection , the damaged region in PD-1-deficient mice had extended throughout the entire liver , while WT mice suffered much less damage and up to 60% of their liver tissue remained normal at this time point ( Fig . 2A ) . Likewise , higher levels of alanine aminotransferase ( ALT ) and aspartate aminotransferase ( AST ) were observed in serum from PD-1-deficient mice after 72 h of infection ( Fig . 2B ) . More interestingly , PD-1-deficient thymus , spleen and lymph node tissues infected with MHV-3 for 72 h exhibited severely disrupted architecture , loss of cellularity , and the presence of substantial amounts of karyorrhectic/apoptotic cell bodies . The histology of these organs from infected WT mice at 72 h was relatively normal ( Fig . 2C ) . In conjunction with the apparent tissue necrosis , higher levels of cell apoptosis were also evidenced in the organs from PD-1-deficient mice by TUNEL staining ( Fig . 2D ) . The architecture of other organs , including the heart , kidney and lung was relatively normal and only rare apoptosis events were observed in these tissues after infection ( Fig . S2 ) . In all , these results demonstrated that PD-1 deficiency led to enhanced pathological damage by MHV-3 in the liver , spleen , lymph node and thymus , where higher levels of PD-1-positive cells were found after infection . The earlier and increased organ damage suffered by PD-1-deficient mice infected with MHV-3 instigated our monitoring of the mortality rates of PD-1-deficient mice and their similarly-infected ( 10 PFU ) WT littermates . As shown in Fig . 3 , all of the PD-1-deficient mice died within four days after infection , while 38% of the WT mice survived up to the end of the 15-day survey period ( p = 0 . 007 ) . These data indicated that PD-1 is likely a critical factor that controls MHV-3-mediated tissue damage and mortality . To understand the mechanisms of PD-1 deficiency-mediated tissue damage and mortality , we performed a comparative genome-wide microarray analysis ( NimbleGen ) of genes expressed in liver tissues of PD-1-deficient and WT mice after 72 h of MHV-3 infection . The most notable finding was pronounced up-regulation ( 3 . 75-fold ) in the liver of PD-1-deficient mice of the fgl2 transcripts ( Fig . 4A ) , the protein product of which has been demonstrated to induce lethality of MHV-3-induced FH [5] . In addition , the enhanced fgl2 expression was confirmed by quantitative ( q ) PCR , results revealed a 2 . 84-fold and 5 . 72-fold higher level was present in liver from WT and PD-1-deficient mice , respectively , after 72 h of MHV-3 infection , as compared to their uninfected controls . Moreover , its level in PD-1-deficient liver was 2 . 01-fold higher than that in the WT group at this time point ( Fig . 4B ) . Immunohistochemistry was used to show that FGL2-positive cells were present in necrotic liver tissues in PD-1-deficient mice at 24 h after MHV-3 injection . The protein expression was found to be enhanced rapidly upon infection , and the highest level occurred at 72 h post-infection . However , occasional FGL2-positive cells were detected in the livers of WT mice at 24 h post-MHV-3 infection and these cells were also present , and slightly enhanced in number , at both the 48 h and 72 h time point ( Fig . 4C ) . Western-blot was used to verify the higher FGL2 protein level in the livers of PD-1-deficient mice , as compared to WT littermates after 72 h of infection ( Fig . 4D ) . FGL2 has the capacity to induce fibrinogen deposition , which then activates the coagulation cascades and finally induces procoagulant activity . Therefore , the expression of FGL2 and fibrinogen deposition in damaged liver tissues was measured . Dual fluorescent staining evidenced that substantial fibrinogen deposition occurred in the FGL2-positive damaged liver tissue ( Fig . 4E ) . Likewise , the level of fibrinogen deposition was more robust in livers from PD-1-deficient mice that in livers from WT littermates , at both the 48 h and 72 h time points ( Fig . 4F ) . To determine whether FGL2-mediated PCA activity was also involved in inducing damage in the other organs of PD-1-deficient mice , the expression of FGL2 was analyzed in the thymus , spleen and lymph nodes . Immunohistochemistry evidenced that FGL2-positive cells were also present in these organs . In thymus and lymph nodes , FGL2-positive cells were detected in both the medulla and cortex . In spleen , however , the positive cells were only found in the red pulp . Again , the expression of FGL2 appeared to be restricted to the cell membrane and cytoplasma . The distribution of FGL2-positive cells in PD-1-deficient mice had not changed after 72 h of MHV-3 infection , but the number of positive cells in the examined organs was enhanced significantly and the levels of expression were much stronger ( Fig . 5A ) . In addition , the transcription of fgl2 in the spleen of PD-1-deficient mice was also significantly increased in response to infection ( Fig . 5B ) . Meanwhile , higher levels of fibrinogen deposition were found in the spleen and lymph node tissues of PD-1-deficient mice ( Fig . 5C ) . Moreover , the level of FGL2 present in serum , as measured by ELISA , was found to increase rapidly after infection , and the level in PD-1-deficient mice was significantly higher than that in WT littermates ( Fig . 5D ) . To clarify the source of FGL2 , fluorescent dual staining was performed on spleen tissues and results demonstrated that FGL2 was principally associated with CD11c-positive dendritic cells ( DCs ) , CD68-positive macrophages and CD31-positive endothelial cells ( Fig . 5E ) . All of these results indicated that the absence of PD-1 signaling can result in enhanced FGL2 expression , consequently inducing stronger fibrinogen deposition and more severe tissue necrosis in PD-1-deficient mice following MHV-3 infection . We further examined whether FGL2 secretion was regulated by PD-1 directly or indirectly . FGL2/PD-1 dual fluorescent staining was performed and results indicated that FGL2 and PD-1 were not co-expressed on the same cells in the liver , thymus , spleen or lymph nodes ( Fig . 6A ) . Previous studies have shown that the secretion of FGL2 can be triggered by the pro-inflammatory factors IFN-γ and TNF-α [21] , [22] . On the other hand , the production of IFN-γ and TNF-α by activated T cells , NK cells and macrophages can be inhibited by the PD-1 signal [6] . Therefore , we compared the status of IFN-γ and TNF-α in PD-1-deficient and WT mice in response to MHV-3 infection . qPCR revealed that the transcription of the IFN-γ gene in liver was significantly higher in PD-1-deficient mice than in WT mice at 72 h post-MHV-3 infection ( Fig . 6B ) . In PD-1-deficient spleen , the transcription of both IFN-γ and TNF-α was found to be rapidly enhanced upon MHV-3 exposure ( Fig . 6C ) . FACs analysis indicated that IFN-γ secretion from NK cells , but not from CD3+T cells , in the liver was much higher in PD-1-deficient mice at 72 h after MHV-3 infection ( Fig . 6D ) . The fact that IFN-γ and TNF-α both have the capacity to initiate FGL2 expression may explain why higher FGL2 expression was observed in the PD-1-deficient mice . To further demonstrate that IFN-γ and TNF-α were responsible for the observed FGL2 up-regulation in MHV-3 infected PD-1-deficient mice , PD-1-deficient mice were infected with MHV-3 and simultaneously treated with a combination injection of anti-IFN-γ and anti-TNF-α blocking mAbs . The expression of fgl2 was measured by qPCR and protein detected by immunohistochemistry . The transcription of fgl2 mRNA ( Fig . 7A ) and its protein levels ( Fig . 7B ) were completely inhibited by 48 h after injection of anti-IFN-γ and anti-TNF-α mAbs , as compared to the control rat IgG1 isotype antibodies-treated group . Moreover , the tissue necrosis ( Fig . 7C ) and liver damage ( as indicated by ALT and AST levels ) ( Fig . 7D ) in PD-1-deficient mice were also significantly reduced , thus the MHV-3-mediated mortality rates were decreased as well ( Fig . 7E ) .
The PD-1 signaling is best known for its ability to inhibit or dampen the immune response . Most of the evidence for this function , however , comes from models of tolerance or chronic infections [11 , 12 , 13 , 14 15] . Although some studies have indicated that this signal might also participate in regulating acute infections [23] , [24] , [25] , [26] , its functions in this disease condition are much less clear . Here , we used a mouse FH model mediated by MHV-3 infection to describe the effects of PD-1 in this disease process . Firstly , PD-1 was found to be significantly up-regulated on T cells , macrophages and NK cells within the thymus , spleen , lymph nodes and liver in response to MHV-3 infection . To determine the exact role of PD-1 in the pathogenesis of FH , PD-1-deficient mice were used to establish an infection model . Interestingly , MHV-3-induced liver damage in PD-1-deficient mice occurred rapidly and the lesion area was much larger than in their WT littermates . We then extended our investigation to the thymus , spleen and lymph nodes , where increased PD-1-positive cells were observed post-infection . Surprisingly , severe tissue necrosis and substantial apoptosis was observed in these organs of PD-1-deficient mice at 72 h after MHV-3 infection . In contrast , these organs from WT mice exhibited relatively normal histology , a finding in agreement with previously reported results [27] . Taken together , these results suggested that PD-1 deficiency promoted expansion of the pathological damage from the liver to the lymph organs , including the spleen , lymph node and thymus in this FH model , thereafter , the absence of PD-1 was associated with higher mortality rates in response to MHV-3 infection . Murine FH induced by MHV-3 is a recognized and validated model for studying host resistance/susceptibility to human hepatitis virus , and several studies have shown that BALB/c or C57BL/6 mice have an innate susceptibility to the infection [3] , [4] . FGL2 has been proposed as a critical mediating factor of lethality in the MHV-3-induced FH mice due to the fact that it has the capacity to induce fibrinogen deposition , which in turn activates the coagulation cascades and induces procoagulant activity [5] . To clarify whether the tissue necrosis we observed in PD-1-deficient mice following infection was also mediated by FGL2 , the expression of FGL2 was analyzed . Results showed that the expression of FGL2 was principally associated with CD31-positive endothelial cells , CD68-positive macrophages and CD11c-positive DCs . Surprisingly , significantly higher levels of FGL2 were observed after infection in all of PD-1-deficient organs , including the liver , thymus , spleen , lymph nodes , and serum than that in those from WT littermates . In addition , increased fibrinogen deposition was observed in the organs of PD-1-deficient mice . Although we currently have no direct data to evidence that FGL2 directly mediates the mortality of our PD-1-deficient mice , data from other researchers have clearly shown that FGL2 promoted mouse mortality in response to MHV-3 infection [5] , [28] , [29] , [30] . Considering this , our results strongly indicate that the mortality of PD-1-deficient mice post-MHV-3 infection is due to the higher level of FGL2 secretion and increased fibrinogen deposition . Indeed , it has been reported that both FGL2 and PD-1 are expressed on T cells , macrophages , and DCs , and that targeted deletion of fgl2 or PD-1 leads to impaired T cell activity , and these events are related to the development of autoimmune diseases [6] , [31] , [32] . We here also observed PD-1 expression as being enhanced on T cells ( both CD4+ and CD8+ T cells ) . It was reasonable to propose that the expression of FGL2 may have been directly regulated by PD-1 signals . Unexpectedly , our FGL2/PD-1 dual staining showed that PD-1-positive cells in the liver , thymus , spleen and lymph nodes did not co-express FGL2 , indicating that the expression of FGL2 was not directly regulated by PD-1 . On the other hand , the expression of FGL2 is believed to be induced by IFN-γ and TNF-α [21] , [22] , while PD-1 signaling has the capacity to inhibit IFN-γ and TNF-α secretion from PD-1-positive immune cells [6] . Therefore , we evaluated and compared the status of IFN-γ and TNF-α in both PD-1-deficient and WT mice . Definitively , the transcription of IFN-γ and TNF-α genes was rapidly enhanced post-MHV-3 infection in PD-1-deficient mice , as compared to WT controls . In particular , a higher level of IFN-γ was observed in NK cells but not in CD3+ T cells of PD-1-deficient liver post-MHV-3 infection , indicating that the PD-1 signal can inhibit IFN-γ secretion from NK cells under such condition . Conversely , injection of a the combination of anti-IFN-γ and anti-TNF-α blocking mAbs was able to successfully inhibit fgl2 mRNA transcription and protein expression , resulting in reduced tissue damage and significantly protecting against MHV-3-mediated mortality in these mice . These results demonstrated that up-regulation of FGL2 in PD-1-deficient mice after MHV-3 infection was controlled , at least partially , by IFN-γ and TNF-α . Recently , the secretion of FGL2 from naturally occurring CD4+Foxp3+ regulatory T cells ( Tregs ) was demonstrated and it was reported that deficiency of Treg-produced FGL2 resulted in increased effector T cell proliferation [32] . More interestingly , Levy and colleagues showed that the frequency of FGL2+ Tregs was higher in lymphoid tissues of MHV-3 infected mice , and treatment with FGL2-specific antibodies reversed MHV-3-induced liver injury and mortality in vivo . These findings demonstrated that FGL2 is an important effector cytokine of Tregs that contributes to MHV-3-induced FH [30] . PD-1 signaling has also been described as participating in regulation of Treg differentiation and function [33] , [34] . In our study , we also analyzed the status of Foxp3+ cells in both PD-1-deficient and WT controls . However , the number of Foxp3-positive cells in the liver , spleen , lymph node or thymus was not significantly different between PD-1-deficient mice and their WT littermates after 72 h of MHV-3 infection ( Fig . S3 ) . Therefore , Foxp3+ cells are unlikely to be involved in the mortality of PD-1-deficient mice . However , the functional status of these Tregs ( for example , the level of FGL2 secretion ) in PD-1-deficienct mice requires further investigation , and such studies are in progress in our lab . In conclusion , we have determined that PD-1 signaling can limit the immunopathological damage induced by MHV-3 infection in a mouse FH model . Our results suggest that enhancing the PD-1 signal by an immunotherapeutic approach might be a useful treatment for FH .
All experiments were approved by and conducted in accordance with the guidelines of the Animal Care and Use Committee of the Third Military Medical University . All efforts were made to minimize animals' suffering . PD-1-KO-N10 ( strain: BALB/cJ ) mice were kindly provided by Prof . T . Honjo ( Department of Immunology and Genomic Medicine , Kyoto University , Japan ) . The WT control mice were purchased from the Animal Center of Beijing University School of Medicine . All mice were maintained in micro-isolator cages and housed in the animal colony at the Animal Center , Third Military Medical University , standard laboratory chow diet and water was supplied ad libitum . Mice were used in experimental analysis at age of six weeks and at an average weight of 17 g ( range: 16∼18 g ) . MHV-3 was kindly provided by Prof . Q . Ning ( Institute of Infectious Disease , Tongji Hospital of Tongji Medical College , Wuhan , China ) . The virus was plaque-purified and then expanded in murine L2 cells . Virus-containing supernatants were collected and stored at -80°C until use . Mice were intraperitoneally ( i . p . ) injected with 10 PFU/mouse in a total volume of 200 µl . In some experiments , PD-1-deficient mice were infected with MHV-3 ( 10 PFU ) and simultaneously treated with a combination injection of anti-IFN-γ ( 200 µg/mouse per day , clone: R4-6A2 , eBioscience , San Diego , CA , USA ) and anti-TNF-α ( 200 µg/mouse per day , clone: MP6-XT22 , eBioscience ) mAbs , tissues were isolated for hematoxylin and eosin ( H&E ) staining to detect damage , and for fgl2 mRNA transcription measured by qPCR ( see below ) . Serum ALT and AST levels were measured by an AU5400 automatic biochemistryanalyzer ( OLYMPUS , Japan ) . In order to monitor the mortality , anti-IFN-γ and anti-TNF-α blocking mAbs or rat IgG1 control mAbs were injected everyday for a total of 6 days . Paraffin-embedded tissue blocks were cut into 5 µm slices which were mounted on polylysine-charged glass slides . Endogenous peroxidase activity was blocked by exposure to 3 . 0% H2O2 for 15 min . Antigen retrieval was performed in a citrate buffer ( pH 6 . 0 ) at 120°C for 10 min . Sections were then incubated at 4°C overnight with anti-mouse FGL2 ( 1∶100 , mouse IgG , Santa Cruz , CA , USA ) , PD-1 ( 5 µg/ml , goat IgG , R&D Systems , Minneapolis , MN , USA ) , CD11c ( 1∶50 , rabbit IgG , Santa Cruz ) , CD68 ( 1∶50 , rat IgG , eBioscience ) , CD31 ( 1∶100 , rabbit IgG , Santa Cruz ) or fibrinogen ( 1∶100 , mouse IgG1 , Dako , Capenteria , CA , USA ) . Immunoreactivity was detected by using a fluorescein isothiocyanate ( FITC ) -conjugated ( 1∶100 , Zymed , San Francisco , CA , USA ) or Cy3-conjugated secondary antibodies ( 1∶200; Jackson ImmunoResearch , West Grove , PA , USA ) . Results were analyzed by fluorescence microscopy ( Axioplan 2 , Zeiss , Germany ) . For immunohistochemical staining , HRP-conjugated anti-mouse , anti-goat or anti-rabbit IgG ( 1∶200 , Zymed ) was used , and the results were visualized with diaminobenzidine ( DAB , Dako ) . Sections incubated with secondary antibodies only were used as isotype controls . PD-1- , FGL2- or Foxp3-positive cells were determined by image analysis of histological sections . Photomicrographs were obtained in high-power fields ( hpf , 0 . 625 mm2 ) and captured for analysis using Image Pro-Plus 5 . 0 software ( Media Cybernetics , Silver Spring , MD , USA ) . The distribution of FGL2 and Foxp3 in thymus and lymph node is restricted in medulla and only the positive cells in this area were calculated . The number of PD-1- , FGL2- or Foxp3-positive cells per hpf were counted and expressed as the mean ± standard error of the mean ( SEM ) . Moreover , analysis of tissue damage was based on H&E staining . The expression of PD-1 on immune cells ( CD4 , CD8 , NK and macrophages ) from different organs was assessed by flow cytometry ( FACsAria Cytometer; Becton Dickinson , Germany ) . Briefly , cell suspensions of liver , spleen , blood and thymus tissues were washed and resuspended in PBS . Cells were then incubated for 30 min at room-temperature in the dark using primary antibodies ( PE-PD-1 , FITC-CD4 , FITC-CD8 , FITC-NK1 . 1 and FITC-CD68 . eBioscience ) . To analyze the source of IFN-γ in the liver , PD-1-deficient and WT mice were treated with MHV-3 ( 10 PFU ) . After 72 h , liver tissues were isolated and mechanically homogenized , lymphocytes were collected thereafter . Cells were then treated with Brefeldin A solution ( BFA ) for 4 h , and FITC-NK1 . 1 , FITC-CD3 or PE-IFN-γ mAbs ( eBioscience ) were added and the solution incubated for an additional 1 h . For each analysis , 10000 cells were evaluated . Flow cytometric data were analyzed with CellQuest Pro Software . The microarray experiment was performed under contact by Kangcheng Co . Ltd . ( Shanghai , China ) . Briefly , total RNA was isolated by Trizol from liver tissue of PD-1-deficient and WT mice treated with 10 PFU MHV-3 for 72 h . RNA concentration was measured on the ND-1000 spectrophotometer ( Nanodrop , Wilmington , DE , USA ) and quality evaluated by denaturing gel electrophoresis . Samples were then amplified and labeled using a NimbleGen One-Color DNA Labeling Kit and hybridized using the NimbleGen Hybridization System ( Roche Applied Science , Shanghai , China ) . After hybridization and washing , the processed slides were scanned by the Axon GenePix 4000B microarray scanner . Three independent experiments were performed , and for each test and control sample , two hybridizations were carried out by a reverse fluorescent strategy . Only genes whose alteration tendency was concordant between both microarray assays were selected as differentially expressed genes . Total RNA from the liver and spleen of WT and PD-1-deficient mice was isolated by Trizol ( Invitrogen , Carlsbad , CA , USA ) , according to the manufacturer's instructions . RNA samples were quantitated by measurement of optical density at 260 nm . Total mRNA ( 2 µg ) was reverse-transcribed to cDNA using the RevertAid H Minus First Strand cDNA Synthesis Kit ( Fermentas China , Shenzhen City , China ) , in accordance with the manufacturer's instructions . qPCR was performed to quantitatively analyze the gene transcription levels of fgl2 , IFN-γ and TNF-α genes . The primers for fgl2 were: sense 5′-TGGACAACAAAGTGGCAAATCT-3′ and anti-sense 5′-TGGAACACTTGCCATCCAAA-3′ . The primers for IFN-γ were: sense 5′-TCAAGTGGCATAGATGTGGAAG-3′ , and anti-sense 5′-CGCTTATGTTGTTGCTGATGG-3′ . The primers for TNF-α were: sense 5′-CACGCTCTTCTGTCTACTGAAC-3′ and anti-sense 5′-ATCTGAGTGTGAGGGTCTGG-3′ . The primers for β-actin ( internal control ) were: sense 5′-CACTATCGGCAATGAGCGGTTCC-3′ and anti-sense 5′-CAGCACTGTGTTGGCA TAGAGGTC-3′ . The qPCR was performed at 95°C for 10 s followed by 40 cycles of 95°C for 5 s , 60°C for 15 s , and 72°C for 15 s . The specificity of PCR product was examined by a dissociation curve , and results were analyzed by the 2−ΔΔCT method [35] . The expression of FGL2 in liver from MHV-3 infected ( 72 h ) PD-1-deficient mice or their WT littermates was determined by Western-blot; the protocol has been described previously [36] . The serum FGL2 level from mice infected with or without MHV-3 was detected by using the mouse FGL2 ELISA Kit ( Cat: E90512Mu; Uscn Life Science Inc . , Wuhan , China ) and following the manufacturer's instructions . All results shown are representative of at least three separate experiments . Unpaired student's t-test ( two-tailed ) or the Mann-Whitney test was used for comparison of two groups where appropriate . Kaplan Meier curve with log-rank test ( GraphPad Prism 4 . 03 software ) was used to analyze the mortality rate . p-value <0 . 05 was considered as statistically significant . | The principal characteristic of fulminant viral hepatitis ( FH ) induced by the murine hepatitis virus strain-3 ( MHV-3 ) is severe hepatocellular necrosis , which is mediated by the fibrinogen-like protein 2 ( FGL2 ) , a molecule that has the capacity to promote fibrinogen deposition and activate the coagulation cascades . Here , we report that MHV-3 infection of program death-1 ( PD-1 ) -deficient mice results in tissue damage throughout multiple organs , including the liver , spleen , thymus and lymph nodes . The liver damage , in particular , occurred earlier and was more severe in PD-1-deficient mice than in their wild type ( WT ) littermates . Further investigation determined that MHV-3 infection was associated with high levels of IFN-γ and TNF-α in the damaged organs of PD-1-deficient mice . Conversely , intraperitoneal injection of a combination of anti-IFN-γ and anti-TNF-α blocking mAbs led to inhibition of FGL2 expression , greatly attenuated tissue lesions and reduced mortality . Our results demonstrate that PD-1 signaling controls immunopathological damage following MHV-3 infection , indicating that manipulation of the PD-1 signal might represent a useful strategy for FH immunotherapy . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"virology/animal",
"models",
"of",
"infection",
"pathology/immunology",
"virology/effects",
"of",
"virus",
"infection",
"on",
"host",
"gene",
"expression",
"immunology/immunity",
"to",
"infections",
"pathology/molecular",
"pathology"
]
| 2011 | Programmed Death (PD)-1-Deficient Mice Are Extremely Sensitive to Murine Hepatitis Virus Strain-3 (MHV-3) Infection |
Macrolides are used to treat various inflammatory diseases owing to their immunomodulatory properties; however , little is known about their precise mechanism of action . In this study , we investigated the functional significance of the expansion of myeloid-derived suppressor cell ( MDSC ) -like CD11b+Gr-1+ cells in response to the macrolide antibiotic clarithromycin ( CAM ) in mouse models of shock and post-influenza pneumococcal pneumonia as well as in humans . Intraperitoneal administration of CAM markedly expanded splenic and lung CD11b+Gr-1+ cell populations in naïve mice . Notably , CAM pretreatment enhanced survival in a mouse model of lipopolysaccharide ( LPS ) -induced shock . In addition , adoptive transfer of CAM-treated CD11b+Gr-1+ cells protected mice against LPS-induced lethality via increased IL-10 expression . CAM also improved survival in post-influenza , CAM-resistant pneumococcal pneumonia , with improved lung pathology as well as decreased interferon ( IFN ) -γ and increased IL-10 levels . Adoptive transfer of CAM-treated CD11b+Gr-1+ cells protected mice from post-influenza pneumococcal pneumonia . Further analysis revealed that the CAM-induced CD11b+Gr-1+ cell expansion was dependent on STAT3-mediated Bv8 production and may be facilitated by the presence of gut commensal microbiota . Lastly , an analysis of peripheral blood obtained from healthy volunteers following oral CAM administration showed a trend toward the expansion of human MDSC-like cells ( Lineage−HLA-DR−CD11b+CD33+ ) with increased arginase 1 mRNA expression . Thus , CAM promoted the expansion of a unique population of immunosuppressive CD11b+Gr-1+ cells essential for the immunomodulatory properties of macrolides .
Macrolides have immunomodulatory properties in addition to their antibacterial effects [1 , 2] . Indeed , previous clinical trials have shown that macrolides exhibit clinical benefits for various pulmonary diseases , including diffuse panbronchiolitis [3] , chronic obstructive pulmonary disease [4] , acute respiratory distress syndrome [5] , cystic fibrosis [6] , and non-cystic fibrosis bronchiectasis [7] . However , the precise mechanism underlying their favorable immunomodulatory effects remains largely unknown . Previous studies of macrolides have focused on their in vitro effects on isolated cells , such as immune cells , epithelial cells , endothelial cells , and fibroblasts [8] . In particular , macrolides inhibit the production of pro-inflammatory cytokines [9]; however , the effects of macrolides on specific immune cell populations have not been examined . We hypothesized that the immunomodulatory properties of macrolides are physiological in origin , as they exhibit globally beneficial effects for various inflammatory diseases with different pathogenesis . Myeloid-derived suppressor cells ( MDSCs ) , a heterogeneous population of myeloid progenitors , are expanded in various diseases , including cancer , bacterial and parasitic infections , acute and chronic inflammation , sepsis , transplantation , and autoimmunity , and can potently suppress T-cell responses [10–12] . In mice , MDSCs express both CD11b and Gr-1 , whereas human MDSCs mainly express Lineage−HLA-DR−CD11b+CD33+ or CD11b+CD14−CD33+ , but lack markers homologous to mouse Gr-1 [10 , 11] . Here , we present evidence that clarithromycin ( CAM ) , a 14-membered macrolide , expands the immunosuppressive CD11b+Gr-1+ cell population to ameliorate lipopolysaccharide ( LPS ) -induced shock and post-influenza pneumococcal pneumonia in mice . We also demonstrate a CAM-induced expansion of the human Lineage−HLA-DR−CD11b+CD33+ cell population in healthy volunteers similar to that observed for CD11b+Gr-1+ cells in mice . Our findings shed new light on the novel immunomodulatory mechanism of macrolides and will contribute to the development of novel approaches to treat inflammatory diseases .
We first examined whether macrolides , in this case CAM , could modulate immune cell populations in wild-type ( WT ) mice . Naive WT mice were intraperitoneally administered CAM daily for 3 days , and single-cell suspensions prepared from the spleen and lungs were analyzed by flow cytometry on the day after the last injection . We did not detect significant differences in the numbers of NK1 . 1+ , B220+ , CD4+ , and CD8+ lymphocytes in both the spleen and lungs between vehicle- and CAM-treated mice ( Fig 1A and 1B ) . Conversely , CAM administration markedly expanded the splenic and lung CD11b+Gr-1+ cell population in naive mice ( Fig 1C–1E ) . This significant increase in CD11b+Gr-1+ cell number was also observed after oral administration of CAM for 2 weeks ( Fig 1F and 1G ) . Next , we intraperitoneally injected CAM at various doses daily for three consecutive days and analyzed cell expansion 24 h after the last injection . As expected , CAM administration increased the CD11b+Gr-1+ cell population in a dose-dependent and time-dependent manner ( S1 Fig ) . We further analyzed two populations of Gr-1+ cells , i . e . , the Ly-6G+ and Ly-6C+ cell populations . Indeed , CAM-treated CD11b+Gr1+ cells were divided into two subsets: CD11b+Ly-6G−Ly-6C+ cells and CD11b+Ly-6G+Ly-6Clow cells ( Fig 1H ) . A cytospin analysis revealed that CD11b+Ly-6G−Ly-6C+ cells displayed a monocytic morphology ( Fig 1I ) , whereas CD11b+Ly-6G+Ly-6Clow cells showed a granulocytic morphology ( Fig 1J ) . The lungs were harvested from mice administered CAM daily for three consecutive days 24 h after the last CAM injection and stained with Gr-1 antibody . Of note , only Gr-1+ immunostaining was used because a previous flow cytometry analysis revealed that nearly all Gr-1+ cells in the lungs after CAM treatment were CD11b+ cells ( Fig 1D ) . Interestingly , Gr-1+ cells increased markedly after CAM treatment and were mainly located in the extravascular space of the lungs ( i . e . , the lung parenchyma or interstitium ) ( S2 Fig ) . To further characterize the CAM-treated CD11b+Gr-1+ cell phenotype , we performed a microarray analysis with sorted splenic CD11b+Gr-1+ cells obtained from vehicle- and CAM-treated mice . We detected a 310-fold increase in the expression of arginase-1 expression , which is a key enzyme inhibiting T cell proliferation and an MDSC marker ( Fig 2A ) . We confirmed that both intraperitoneal injection and oral administration of CAM induced arginase-1 mRNA expression in splenic CD11b+Gr-1+ cells ( S3 Fig ) . We further measured arginase activity in the spleen and arginase-1 expression in the lungs . High arginase activity was detected in CAM-treated CD11b+Gr-1+ cells , whereas no activity was detected in vehicle-treated CD11b+Gr-1+ cells ( Fig 2B ) . In addition , immunostaining analysis confirmed that Gr-1+ cells produce arginase-1 in the lungs of CAM-treated mice ( Fig 2C ) . We next measured the concentration of splenic nitric oxide ( NO ) , a crucial parameter for MDSC biology and identification [13] . The NO concentration was significantly higher in the spleen of CAM-treated mice than in that of vehicle-treated mice ( Fig 2D ) . To further investigate the function of CAM-treated CD11b+Gr-1+ cells in comparison with neutrophils and monocytes/macrophages , elastase activity , MPO activity , and phagocytic activity were measured ( S4 Fig ) . Elastase activity in CAM-treated CD11b+Gr-1+ cells was lower than that in any other group ( vehicle-treated CD11b+Gr-1+ cells [mainly steady-state non-activated neutrophils and monocytes/macrophages] , LPS-treated CD11b+Gr-1+ cells ( mainly activated neutrophils and monocytes/macrophages ) , thioglycolate-elicited sterile peritonitis-induced activated neutrophils , and isolated peripheral non-activated neutrophils ) ( S4A Fig ) , consistent with the notion that elastase activity in MDSCs is lower than that in neutrophils . MPO activity in CAM-treated CD11b+Gr-1+ cells was greater than that in vehicle-treated CD11b+Gr-1+ cells and in peripheral neutrophils , which appeared to be non-activated steady-state neutrophils or monocytes/macrophages . However , it was lower than that in LPS-treated CD11b+Gr-1+ cells ( activated neutrophils and monocytes/macrophages ) and thioglycolate-elicited sterile peritonitis-induced activated neutrophils ( S4B Fig ) . In addition , phagocytic activity in CAM-treated CD11b+Gr-1+ cells was lower than that in any other group ( S4C Fig ) , supporting the notion that CAM-treated CD11b+Gr-1+ cells are not neutrophils or monocytes/macrophages , but MDSC-like cells . To verify that CAM-treated CD11b+Gr-1+ cells suppressed T cell activity , the minimal functional characteristic necessary to identify cells as MDSCs [13] , we performed carboxyfluorescein succinimidyl ester ( CFSE ) T cell proliferation assays . The histogram of CD3+ T cells co-cultured with CAM-treated CD11b+Gr-1+ cells shifted to the right along the X-axis when compared to that for the vehicle-treated CD11b+Gr-1+ cells , but the degree of the shift was modest . These results indicate that CAM-treated CD11b+Gr-1+ cells might hinder CD3+ T cell proliferation ( S5 Fig ) . We next examined the CAM-treated CD11b+Gr-1+ cells for the presence of various surface markers expressed on MDSCs and/or myeloid cells , including CD244 [14] , CTLA-4 [15] , PD-1 [16] , PD-L1 [17] , CXCR2 [11] , CXCR4 [11] , CD80 [11] , CD115 [11] , and CX3CR1 [18] . There were no dramatic differences in the expression of these markers , but CD244 and CTLA-4 expression tended to be slightly higher in CAM-treated CD11b+Gr-1+ cells than in vehicle- treated CD11b+Gr-1+ cells ( S6 Fig ) . Because CD244 is specifically expressed on granulocytic MDSCs [14] , we evaluated CD244 in CD11b+Ly-6G+ cells . We confirmed high CD244 expression levels in CAM-treated CD11b+Ly6G+ cells ( Fig 2E ) . To further investigate the functions of these cell populations in vitro , bone marrow-derived macrophages ( BMDMs ) were co-cultured with CAM- or vehicle-treated CD11b+Gr-1+ cells stimulated with LPS ( i . e . , with cell–cell contact ) , and cytokine secretion was assessed 12 h later . Notably , CAM-treated CD11b+Gr-1+ cells suppressed pro-inflammatory cytokine production ( TNF-α and IFN-γ ) and increased anti-inflammatory IL-10 levels ( Fig 2F–2H ) . Collectively , these in vitro results indicate that CAM-treated CD11b+Gr-1+ cells potentiate immunomodulatory properties related to cytokine production . Previous studies have suggested that 14- and 15-membered ring , but not 16-membered ring macrolides have immunomodulatory properties [19 , 20]; thus , we evaluated the immunomodulatory properties of various macrolides by assessing CD11b+Gr-1+ cell expansion . Interestingly , CAM ( 14-membered ring ) and azithromycin ( 15-membered ring ) , but not josamycin ( JOS ) ( 16-membered ring ) , increased the percentage and number of splenic and lung CD11b+Gr-1+ cells ( S7 Fig ) . Because 14- and 15- , but not 16-membered macrolide antibiotics can decrease pro-inflammatory cytokine and chemokine secretion , these results suggest that the extent of macrolide-induced CD11b+Gr-1+ cell expansion is an indicator of their immunomodulatory effects . To examine the immunomodulatory properties of CAM-treated CD11b+Gr-1+ cells in vivo , mice were subjected to LPS-endotoxin shock . Pretreatment with CAM significantly increased survival ( Fig 3A ) . Moreover , TNF-α and IFN-γ serum levels were significantly lower in CAM-treated mice 12 h after LPS injection , and IL-10 levels were significantly higher than those in vehicle-treated counterparts ( Fig 3B–3D ) . A flow cytometry analysis revealed that CAM accelerated the LPS-dependent expansion of CD11b+Gr-1+ cells in both the spleen and lungs ( Fig 3E and 3F ) . The numbers of CD11b+Gr-1+ cells in the spleen and lungs were slightly , but significantly higher in CAM-treated mice than in vehicle-treated mice ( Fig 3G ) . To investigate the role of the CD11b+Gr-1+ cell population in the LPS-shock model , we attempted to deplete CD11b+Gr-1+ cells . The experimental schema for the analysis summarized in Fig 3H and 3I is shown in S8 Fig . Because >85% of splenic and lung Gr-1+ cells in CAM-treated LPS-shock mice were CD11b+ ( Fig 3E and 3F ) , we used an anti-Gr-1 ( including Ly-6G and Ly-6C ) antibody to deplete this cell population . The improved survival observed for CAM treatment in the LPS-shock model was reversed by Gr-1+ cell depletion at 24 h before LPS treatment ( Fig 3H ) . In addition , we obtained similar results when the Gr-1+ cell population was depleted 1 h before the first CAM treatment ( i . e . , from the beginning of CAM treatment ) ( Fig 3I ) . These results indicate that the CAM-treated Gr-1 population has protective effects in this model . Next , we adoptively transferred CD11b+Gr-1+ cells isolated from CAM- or vehicle-treated donors to LPS-challenged recipient mice . Adoptive transfer of CAM-treated CD11b+Gr-1+ cells protected mice from LPS-induced lethality , as compared to vehicle-treated CD11b+Gr-1+ cells ( Fig 3J ) and PBS control injections ( S9 Fig ) , and was associated with decreased TNF-α and IFN-γ and increased IL-10 serum levels ( Fig 3K–3M ) . Collectively , these results indicated that CAM ameliorates LPS-induced lethality via CAM-treated CD11b+Gr-1+ cells . Because CAM-treated CD11b+Gr-1+ cells potentiated anti-inflammatory IL-10 secretion in response to LPS stimulation ( Fig 2F ) , we hypothesized that IL-10 production in CD11b+Gr-1+ cells contributed to the beneficial effects on LPS-induced shock . To evaluate our hypothesis , we adoptively transferred CAM-treated CD11b+Gr-1+ cells from WT and IL-10 knockout ( Il10-/- ) donors into syngeneic WT recipients . Notably , WT CD11b+Gr-1+ cell grafts improved survival in the LPS-induced shock model , whereas those from Il10-/- donors failed to convey these effects ( Fig 4A ) , indicating that IL-10 production from CAM-treated CD11b+Gr-1+ cells contributed to the protective effects of macrolides in LPS-induced endotoxin shock . To confirm these effects in vitro , we measured supernatant cytokine levels in co-cultures of BMDMs and CAM-treated WT or Il10-/- CD11b+Gr-1+ cells 12 h after LPS stimulation . As expected , the TNF-α and IFN-γ levels were significantly higher in cells co-cultured with Il10-/- cells than in those from WT counterparts ( Fig 4B and 4C ) , indicating that the decreased pro-inflammatory cytokine production by CAM-treated CD11b+Gr-1+ cells was at least partially IL-10-dependent . We further investigated the role of IL-10 in expansion of the CAM-treated CD11b+Gr1+ cell population . CAM still expanded CD11b+Gr-1+ cells in Il10-/- mice , suggesting that the major expansion mechanism of CD11b+Gr1+ cells by CAM is not related to IL-10 ( S10 Fig ) . We examined the immunomodulatory effects of CAM-treated CD11b+Gr-1+ cells in a mouse model of post-influenza pneumococcal pneumonia . In particular , we used a clinically isolated strain of Streptococcus pneumoniae ( serotype 3 ) that harbors penicillin and macrolide resistance genes , although the strain was classified as a penicillin-intermediate resistant S . pneumoniae ( PISP ) , as described in the Materials and Methods section and S1 Table . CAM treatment significantly improved survival when compared to that of vehicle- or ampicillin ( ABPC ) -treated mice ( Fig 5A ) , whereas JOS treatment did not improve survival ( S11 Fig ) . Interestingly , both CAM and ABPC significantly reduced total cell counts in bronchoalveolar lavage fluid ( BALF ) as compared to those in vehicle-treated mice ( Fig 5B ) , as well as the bacterial load in lungs 18 h after infection ( Fig 5C ) . Blood cultures showed lower bacterial loads in CAM-treated mice than in vehicle- and ABPC-treated mice , but the differences were not significant ( Fig 5D ) . Moreover , a histopathological examination of the lungs 18 h after pneumococcal infection indicated that both CAM and ABPC ameliorated white blood cell recruitment ( mainly neutrophils ) , alveolar wall thickening , and edema when compared to those of vehicle-treated mice ( Fig 5E ) . Furthermore , IFN-γ levels were significantly suppressed in the BALF and serum in CAM-treated mice as compared to vehicle-treated and ABPC-treated mice , respectively , whereas IL-10 was markedly increased in both the BALF and serum of CAM-treated mice ( Fig 5F–5I ) . Similarly , adoptive transfer of splenic CD11b+Gr-1+cells into mice with post-influenza pneumococcal pneumonia immediately after pneumococcal inoculation revealed that CAM-treated CD11b+Gr-1+ cells prolonged the survival of recipient mice when compared to ABPC- or vehicle-treated counterparts ( Fig 5J ) . Previous studies of post-influenza pneumococcal pneumonia have reported that IFN-γ production is detrimental to secondary bacterial infection [21 , 22]; thus , because our results showed that CAM and CAM-treated CD11b+Gr-1+ cells suppress IFN-γ production in post-influenza pneumococcal pneumonia , we hypothesized that IFN-γ might have a detrimental effect on survival in our model mice . Indeed , CAM-mediated prolonged survival was partially reversed by intranasal administration of recombinant IFN-γ ( Fig 5K ) . To further investigate the roles of IFN-γ in CAM-treated CD11b+Gr-1+ cells , IFN-γ deficient ( Ifng-/- ) and WT mice were subjected to post-influenza pneumococcal pneumonia . We first confirmed better survival in vehicle-treated Ifng-/- mice than in vehicle-treated WT mice ( Fig 5L ) , consistent with previous results showing that Ifng-/- mice are resistant to post-influenza pneumococcal pneumonia [22] . We also observed slightly higher survival in CAM-treated Ifng-/- mice than in vehicle-treated Ifng-/- mice and CAM-treated WT mice ( Fig 5L ) , suggesting that the protective effects of CAM in this model depend , at least in part , on suppression of IFN-γ production . Collectively , these data demonstrate that CAM improves the overall survival of mice with post-influenza pneumococcal pneumonia , at least in part via suppression of IFN-γ production by CAM-treated CD11b+Gr-1+ cells . Previous studies have indicated that type I IFN is also detrimental in a post-influenza pneumococcal pneumonia model using interferon-α/β receptor ( IFNAR ) knockout ( Ifnar1-/- ) mice [23] . To investigate the role of type I IFN in CAM-treated CD11b+Gr-1+ cells , Ifnar1-/- mice and WT mice were subjected to post-influenza pneumococcal pneumonia . We did not observe a significant difference between WT mice and Ifnar1-/- mice . In addition , CAM did not significantly improve survival in Ifnar1-/- mice , but it resulted in a slight improvement in survival in Ifnar1-/- mice ( S12 Fig ) . These results may be explained by the strong virulence of influenza in Ifnar1-/- mice . Accordingly , we could not determine whether the protective effects of CAM in post-influenza pneumococcal pneumonia are independent or partially dependent on type I IFN , and additional studies with larger sample sizes are needed . We next explored the mechanism underlying CAM-induced CD11b+Gr-1+ cell expansion , and initially focused on effectors known to induce the differentiation and proliferation of CD11b+Gr-1+ cells , including IL-17 , CCL2 , CXCR4 , and toll-like receptor ( TLR ) /MyD88 signaling [10] . Thus , we examined CAM-mediated CD11b+Gr-1+ cell expansion in Il17a-/- , Ccl2-/- , and Myd88-/- mice , as well as WT mice treated with a CXCR4 antagonist ( AMD3100 ) . However , no differences were observed between these mice and WT controls ( S13 Fig ) , indicating that IL-17 , CCL2 , CXCR4 , and TLR/MyD88 signaling pathways are not involved in this process . The prokineticin family member Bv8 ( also known as prokineticin 2 [Prok2] ) was previously reported to induce CD11b+Gr-1+ cells in an autocrine and paracrine STAT3-dependent manner [24] . Notably , our microarray data identified that Bv8 ( Prok2 ) was highly expressed ( 12 . 3-fold higher relative to WT mice ) in splenic CD11b+Gr-1+ cells obtained from CAM-treated mice ( Fig 2A ) . A protein analysis of splenic cells after CAM treatment revealed higher Bv8 expression than that in vehicle-treated mice ( Fig 6A ) , as well as increased phosphorylated ( p ) -STAT3 ( Fig 6B ) . To investigate the role of Bv8 in CAM-induced CD11b+Gr-1+ cell induction , mice were administered an anti-Bv8 antibody at 4 days and 1 day before , and then during , the CAM treatment period ( three consecutive days ) . Notably , treatment with the anti-Bv8 antibody reduced the percentage of CD11b+Gr-1+ cells in both the spleen ( 16 . 5% vs . 6 . 59%; 60 . 1% reduction ) ( Fig 6C ) and lungs ( 14 . 3% vs . 8 . 04%; 43 . 8% reduction ) ( Fig 6D ) . Because Bv8 and STAT3 form a feed-forward loop that promotes CD11b+Gr-1+ myeloid cell proliferation [24] , we investigated this as a potential induction mechanism in Stat3flox/flox/Mx1-Cre ( Stat3ΔMx1 ) mice . As predicted , the CAM-treated CD11b+Gr-1+ cell populations in the spleen and lungs were reduced in Stat3ΔMx1 mice ( Fig 6E and 6F ) . Consistently , Bv8 gene expression was significantly reduced in splenic CD11b+Gr-1+ cells isolated from CAM-treated Stat3ΔMx1 mice when compared to expression in Stat3flox/flox control mice ( Fig 6G ) . Taken together , these results indicate that STAT3-Bv8 signaling facilitates the CAM-mediated expansion of the CD11b+Gr-1+ cell population . Macrolide administration causes substantial changes in the composition of the gut microbiota [25] . Accordingly , we hypothesized that the modulation of the gut microbiota by CAM might contribute to the expansion of CD11b+Gr-1+ cells . To investigate this hypothesis , germ-free mice and specific pathogen-free ( SPF ) mice were administered CAM daily for three consecutive days , and the CD11b+Gr-1+ cell populations were quantified the day after the last injection . The baseline CD11b+Gr-1+ cell population in untreated germ-free mice was significantly lower than that in untreated SPF mice ( S14 Fig ) , suggesting that the resident CD11b+Gr-1+ cell population is influenced by the gut commensal microbiota . Moreover , expansion of the CD11b+Gr-1+ cell population by CAM was also observed in both germ-free mice and SPF mice; however , the relative expansion of splenic and lung CD11b+Gr-1+ cells was significantly lower in CAM-treated germ-free mice than in SPF counterparts , suggesting that commensal gut microbiota enhance CD11b+Gr-1+ cell population expansion in response to CAM treatment ( S14 Fig ) . To determine whether the immunosuppressive MDSC-like population also exhibits CAM-induced expansion in humans , six healthy volunteers were administered 800 mg of CAM daily for 7 days and relative changes in the Lineage−HLA-DR−CD11b+CD33+ MDSC population were examined in peripheral blood by flow cytometry ( Fig 7A ) . Although the Lineage−HLA-DR−CD11b+CD33+ cell population appeared to increase following CAM treatment , the difference was not significant ( p = 0 . 0699 ) ( Fig 7B ) ; however , arginase-1 gene expression was markedly elevated in these cells after 7 days of CAM treatment compared to the baseline control measurements ( Fig 7C ) . These results suggest that human Lineage−HLA-DR−CD11b+CD33+ MDSCs were polarized to a more immunosuppressive phenotype in response to CAM treatment .
In this study , we demonstrated that CAM stimulates the expansion of the immunosuppressive CD11b+Gr-1+ cell population to enhance survival in a LPS-induced shock and post-influenza pneumococcal pneumonia mouse model , in a STAT3/Bv8 axis-dependent manner . We also showed that human Lineage−HLA-DR−CD11b+CD33+ cells correspond to the CD11b+Gr-1+ population in mice , as they both exhibited a significant increase in arginase-1 mRNA expression levels following CAM treatment . Increasing evidence suggests that macrolides have immunomodulatory properties , beyond their antibacterial effects [1 , 2]; however , the overuse of these drugs for this purpose is criticized because the mechanism of action is unclear [26] . Therefore , the elucidation of their exact immunomodulatory effects is needed . Numerous in vitro studies have demonstrated that macrolides modulate the release of humoral factors from immune , epithelial , and other structural cells [8] . Our study is distinguished from previous reports in that it identified the unique immunosuppressive cell population mediating these effects in mice , and implicates a homologous cell population in humans . Our analyses indicated that the CAM-treated CD11b+Gr-1+ cells were MDSC-like cells . MDSCs are a population comprised of heterogeneous cell clusters consisting of myeloid progenitor cells and immature myeloid cells , e . g . , immature neutrophils [10] . The differences between granulocytic-MDSCs and mature neutrophils are often debatable [27 , 28] . Here , we confirmed that the CAM-treated CD11b+Gr-1+ cell population includes both CD11b+Ly6G+Ly6Clow cells ( granulocytic-MDSC-like cells ) and CD11b+Ly-6G−Ly-6C+ cells ( monocytic-MDSC-like cells ) . We cannot estimate the magnitude of the contributions of these two populations ( granulocytic- and monocytic-MDSC-like cells ) to the beneficial effects of CAM treatment because we used an anti-Gr-1 antibody ( RB6-8C5 clone , anti-Ly-6G and Ly-6C antibody ) to deplete the Gr-1+ cell population . The CAM-treated CD11b+Gr-1+ cell population can be discriminated from mature neutrophils in their anti-inflammatory effects , as mature neutrophils are generally inflammatory in nature . We showed that CD11b+Gr-1+ cells actively secrete the anti-inflammatory cytokine IL-10 and suppress the production of the inflammatory cytokines IFN-γ and TNF-α , although selected neutrophils have also been reported to produce IL-10 [29] . In addition , our results indicating that elastase activity and MPO activity were lower in CAM-treated CD11b+Gr-1+ cells than in activated neutrophils/monocytes/macrophages , including LPS-treated CD11b+Gr-1+ cells and thioglycolate-induced neutrophils , are consistent with the notion that CAM-treated CD11b+Gr-1+ cells are not activated neutrophils , but MDSC-like cells . Importantly , our work demonstrated that CAM-treated CD11b+Gr-1+ cells potentiate the production of IL-10 , an anti-inflammatory cytokine that plays a central role in infection by limiting the immune response to pathogens and preventing damage to the host [30] . Many earlier studies found that IL-10 promotes survival in sepsis and endotoxin shock [31–33] , consistent with our results indicating beneficial effects of IL-10 in LPS-induced shock . IL-10 has been reported to have detrimental roles in post-influenza pneumococcal pneumonia [34 , 35]; however , another previous study found no significant differences in susceptibility to post-influenza secondary pneumococcal infection between WT and Il10-/- mice [22] . We did not determine the detailed mechanisms underlying increased IL-10 levels in the post-influenza pneumococcal pneumonia model in the present study , but this should be evaluated in further studies . We also showed that CAM-treated CD11b+Gr-1+ cells distinctly suppressed IFN-γ production both in vitro and in vivo , corroborating earlier studies of the 14-membered ring macrolide erythromycin in a mouse model of influenza virus infection [36] . The significance of suppressed IFN-γ production in LPS-induced endotoxin shock was not evaluated in the present study; however , IFN-γ has been reported to be detrimental in LPS-endotoxin shock by hampering endotoxin tolerance [37 , 38] . Considering these previous reports and our results indicating the partial restoration of IFN-γ production in macrophages co-cultured with CAM-treated Il10-/- CD11b+Gr-1+ cells ( Fig 4C ) , we speculate that enhanced IL-10-mediated IFN-γ suppression by CAM would also protect against LPS-induced endotoxin shock . In terms of significance , a previous study indicated that IFN-γ production by CD4+ and CD8+ T cells in response to influenza virus infection enhanced susceptibility to secondary pneumococcal infection [22] . In addition , accumulating evidence indicates that CD11b+Gr-1+ cells suppress IFN-γ production by T cells [39 , 40] and NK cells [41] . As such , we speculate that the suppression of IFN-γ production in these cells , as well as macrophages , by CAM treatment would have protective effects in post-influenza pneumococcal pneumonia . Consistent with this notion , we confirmed that exogenous administration of recombinant IFN-γ partially reversed the survival benefit in post-influenza pneumococcal pneumonia model mice . We also observed slightly greater survival in CAM-treated Ifng-/- mice than in vehicle-treated Ifng-/- mice and CAM-treated WT mice . These results suggest that the protective effects of CAM in this model at least partially depend on suppression of IFN-γ production . We speculate that , in addition to IFN-γ suppression , other mechanisms , such as increased IL-10 , contribute to the protective effects of CAM in this model . Furthermore , we found that the STAT3/Bv8 axis is actively involved in the mechanism underlying CAM-induced CD11b+Gr-1+ cell expansion . Bv8 induces CD11b+Gr-1+ cells in an autocrine and paracrine STAT3-dependent manner [24] , and regulates CD11b+Gr-1+ cell-dependent tumor angiogenesis [42] . A Bv8 antagonist was found to inhibit angiogenesis and myeloid cell infiltration in mouse models of glioblastoma and pancreatic cancer [43] . These studies , in combination with our study , suggest that the STAT3/Bv8 axis is essential to myeloid cell recruitment , including that of CD11b+Gr-1+ cells . In addition , because STAT3 plays a crucial role in the IL-10-mediated anti-inflammatory response [44] , and considering our results for IL-10 induction from CAM-treated CD11b+Gr-1+ cells , it is likely that the expanded CD11b+Gr-1+ cell population mediates its immunomodulatory effects via STAT3-dependent IL-10 production . Previous studies have clearly demonstrated that 14- and 15-membered , but not 16-membered macrolide antibiotics , can decrease pro-inflammatory cytokine and chemokine secretion [8 , 19 , 20] . Similarly , we observed expansion of CD11b+Gr-1+ cells following treatment with 14- and 15-membered , but not 16-membered macrolide antibiotics , suggesting that the degree of CD11b+Gr-1+ cell expansion is correlated with their immunomodulatory properties . Thus , macrolide-treated CD11b+Gr-1+ cells might be a good biological marker for evaluating the potency of their immunomodulatory properties . Oral administration of 25–100 mg/kg CAM in mice is comparable to clinically achievable levels in humans [45] . As the area under the curve ( AUC ) was more than 2-fold higher for intraperitoneal administration of 10 mg/kg CAM than for oral administration of 10 mg/kg CAM ( personal communication ) , we believe that the intraperitoneal administration of CAM at 100 mg/kg in the present study is clinically relevant for humans . The overuse of macrolide antibiotics could result in the evolution of macrolide-resistant organisms . To address this concern , immunomodulatory macrolides lacking antimicrobial effects , e . g . , EM900 and EM703 , have been in development [46 , 47] . The presence of immunosuppressive CD11b+Gr-1+ cells can be a good surrogate marker to measure the immunomodulatory effects of these types of drugs in early development . Finally , we tried to identify the human immunosuppressive MDSC population that corresponds to murine CD11b+Gr-1+ cells . Exact human MDSC surface markers are yet to be identified , partly owing to the extensive heterogeneity in the human MDSC population [11] . However , based on previous studies [11] , we identified Lineage-HLA-DR-CD11b+CD33+ cells as a major human MDSC population . In addition , we observed upregulation of arginase-1 mRNA in the human MDSC population after oral administration of CAM , indicating that CAM likely expands immunosuppressive MDSC-like cells in humans . In the present study , peripheral blood was used to assess CAM-treated human MDSCs , but these effects could be evaluated with much greater clarity in lung or spleen cell populations , as in the model mice . In the present study , we used an anti-Gr-1 antibody ( RB6-8C5 clone , anti-Ly-6G and Ly-6C antibody ) to deplete the Gr-1+ cell population , resulting in depletion of both monocytic- and granulocytic-MDSC-like cell populations . This may not be a good approach for the depletion of MDSC-like cells owing to the lack of selectivity , i . e . , myeloid cells such as neutrophils and monocytes/macrophages can also be depleted by treatment with the anti-Gr-1 antibody . Although several parameters have been suggested for distinguishing granulocytic-MDSCs from neutrophils [14] , they are insufficient [48] . Further investigations are needed to better differentiate these cells . A working hypothesis for the mechanism underlying CAM-mediated expansion of the CD11b+Gr-1+ cell population is summarized in S15 Fig . In this system , CAM expands the CD11b+Gr-1+ cell population in a STAT3/Bv8-dependent manner . The anti-inflammatory effects of these cells , including increased IL-10 and decreased IFN-γ , are sufficient to protect mice from death by LPS-induced shock and post-influenza pneumococcal pneumonia . In conclusion , we identified a novel mechanism underlying CAM-mediated expansion of the immunosuppressive CD11b+Gr-1+ cell population , sufficient to attenuate LPS-induced lethal shock and post-influenza pneumococcal infection . Considering that the immunosuppressive effects of CD11b+Gr-1+ cells may also ameliorate various inflammatory diseases , our findings provide a new strategy for the treatment of inflammatory diseases and the development of new classes of macrolides and novel drugs . Further detailed studies would contribute to a complete understanding of the mechanism underlying the beneficial effects of CAM .
Male WT C57BL/6J mice ( 8–10 weeks old ) were purchased from CLEA Japan , Inc . ( Tokyo , Japan ) . Il10−/− mice and Ccl2-/- mice were purchased from Jackson Laboratory ( Bar Harbor , ME , USA ) . Il17a-/- and Ifng-/- mice were kindly provided by Prof . Yoichiro Iwakura ( Tokyo University of Science , Noda , Japan ) . Ifnar1-/- mice were kindly provided by Prof . Shigekazu Nagata ( Kyoto University , Kyoto , Japan ) [48 , 49] . Stat3flox/flox mice and Myd88-/- mice were purchased from Oriental BioService , Inc . ( Kyoto , Japan ) Mx-1-Cre mice were kindly provided by the Mouse Genetics Cologne Foundation in Germany . Stat3ΔMx1 mice were generated by mating Stat3flox/flox mice with Mx-1-Cre mice . To remove the floxed Stat3 allele , 8-week-old Stat3ΔMx1 mice were intraperitoneally injected with 250 μg of polyinosinic-polycytidylic acid ( [Poly ( I:C ) ]; Sigma-Aldrich , St . Louis , MO , USA ) three times every other day . Littermate Stat3flox/flox mice served as controls . Germ-free mice were purchased from CLEA Japan , Inc . , and were kept in the germ-free facility at Keio University School of Medicine . Other mice were housed under specific pathogen-free ( SPF ) conditions . The present study was performed in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . All animal experiments were approved by the Animal Care and Use Committee at the Keio University School of Medicine ( Protocol No . 12110 ) . The human study was approved by the Ethics Committee of Keio Center for Clinical Research in November 2012 ( Protocol No . 20120250 ) . All participants of the study ( i . e . , blood donors ) were informed about the risks of participation in the study and written consent was obtained from all of them . All participants were adults and non-minors . Clarithromycin ( CAM ) was kindly provided by Taisho Toyama Pharmaceutical Co . , Ltd . ( Tokyo , Japan ) , and diluted with PBS containing 5% gum arabic . The CAM or vehicle solution ( 5% gum arabic in PBS ) was intraperitoneally injected at 100 mg/kg daily throughout the course of the experiment . In oral administration experiments , the CAM and vehicle solutions were administered at 100 mg/kg by gavage daily for 14 days . The dose of CAM was determined based on preliminary experiments summarized in S1 Fig and previous reports [45 , 50 , 51] . Azithromycin ( Sigma-Aldrich ) , josamycin ( Sigma-Aldrich ) , and ampicillin ( Sigma-Aldrich ) were dissolved in dimethyl sulfoxide ( DMSO; Sigma-Aldrich ) at 10 mg/mL and further diluted in sterile PBS . Antibiotics were administered in vivo at 100 mg/kg , 200 mg/kg , and 100 mg/kg , based on previous reports [51–54] . In mouse studies , single-cell suspensions were prepared for flow cytometry as follows; spleens and lungs were cut into small fragments and digested for 45 min at 37°C with RPMI 1640 containing 0 . 05 mg/mL Liberase ( Roche Diagnostics , Basel , Switzerland ) and 250 U/mL DNAse I ( Worthington Biochemical Corporation , Lakewood , NJ , USA ) . Single cells were washed with fluorescence activated cell sorting ( FACS ) buffer ( 2% fetal calf serum [FCS] and 1 mg/mL sodium azide in PBS ) and incubated with an anti-CD16/CD32 monoclonal antibody ( eBioscience , San Diego , CA , USA ) for 10 min followed by staining with the following antibody conjugates for 20 min: PE-Cy7-CD45 ( 30-F11; BioLegend , San Diego , CA , USA ) , APC-Cy7-CD45 ( 30-F11; BD Biosciences , Franklin Lakes , NJ , USA ) , PerCP-Cy5 . 5-CD45 ( 30-F11; BD Biosciences ) , PE-Cy7-CD3 ( 17A2; BioLegend ) , PerCP-Cy5 . 5-CD4 ( GK1 . 5; BioLegend ) , APC-Cy7-CD8a ( 53–6 . 7; BioLegend ) , APC-TCR-γδT ( B1; BioLegend ) , PE-NK1 . 1 ( PK136; BioLegend ) , APC-Gr-1 ( RB6-8C5; BioLegend ) , FITC-CD11b ( M1/70; BioLegend ) , APC/Cy7-Ly-6G ( 1A8; BioLegend ) , PerCP/Cy5 . 5-Ly-6C ( HK1 . 4; BioLegend ) , PE-CD244 . 2 ( 2B4 B6; eBioscience ) , PE-CTLA-4 ( BioLegend ) , PE-PD1 ( eBioscience ) , PE-PDL1 ( eBioscience ) , PE-CXCR2 ( BioLegend ) , PE-CXCR4 ( eBioscience ) , PE-CD80 ( BioLegend ) , PE-CD115 ( eBioscience ) , and PE-CX3CR1 ( R&D Systems , Minneapolis , MN , USA ) . The cells were then washed twice and subsequently stained with PE-Texas Red propidium iodide ( PI; BD Biosciences ) . The antibody conjugates PE-CD33 ( HIM3-4; BD Biosciences ) , Alexa700-CD11b ( CR3A; BioLegend ) , PE/Cy7-HLA-DR ( BioLegend ) , FITC-Lineage ( CD3 , CD19 , CD56 ) ( eBioscience ) , and APC/Cy7-CD45 ( HI30; BioLegend ) were used for human cell analyses . The cells were washed twice and stained with PE-Texas Red PI ( BD Biosciences ) . Stained cells were processed by flow cytometry using a BD FACS Aria II system ( BD Biosciences ) or MoFlo XDP ( Beckman Coulter , Brea , CA , USA ) and the data were analyzed using FlowJo 7 . 6 . 2 ( Tree Star , Inc . , Ashland , OR , USA ) . Lung tissue was fixed with 4% paraformaldehyde and embedded in paraffin . The lung sections were deparaffinized in xylene and rehydrated in a graded ethanol series , and antigen retrieval was performed by heating sodium citrate buffer at 95°C for 20 min . Non-serum protein block ( Dako ) was applied for 30 min . Slides were then stained with rat anti-Gr-1 primary antibody ( RB6-8C5; R&D Systems ) or rabbit anti-arginase-1 primary antibody ( ABS535; Merck , Kenilworth , NJ , USA ) followed by Alexa Fluor 488- or Alexa Fluor 555-coupled secondary antibody , respectively ( Invitrogen , Carlsbad , CA , USA ) . Sections were counterstained with DAPI ( Vector Labs ) and analyzed using an Axio Imager ( Zeiss , Oberkochen , Germany ) . After mice were anesthetized by intraperitoneal injection of xylazine ( 10 mg/kg ) and ketamine ( 100 mg/kg ) , an APC-Cy7-CD45 antibody conjugate ( 0 . 4 μg/2 μL diluted in 100 μL of PBS ) was administered 5 min before euthanasia together with heparin ( 10 IU/g ) . After 5 min , mice were exsanguinated and intratracheally injected with a PerCPCy5 . 5-CD45 antibody conjugate ( 0 . 4 μg/2 μL diluted in 100 μL of PBS ) . After an additional 5 min , the lungs were harvested and a lung single cell suspension was prepared , followed by staining with a PE-Cy7-CD45 antibody conjugate ex vivo . Total RNA was isolated using the RNeasy Mini Kit ( QIAGEN , Hilden , Germany ) according to the manufacturer’s protocol . The isolated RNA was quantified , and quality was assessed using an Agilent 2100 BioAnalyzer with the RNA Nano Chip ( Agilent , Santa Clara , CA , USA ) . Then , 2 μg of total RNA was used to generate single-stranded antisense cDNA using the Ovation Biotin System ( NuGEN Technologies , San Carlos , CA , USA ) according to the manufacturer’s instructions . Labeled targets were hybridized to Affymetrix MOE430 2 . 0 GeneChip microarrays ( Affymetrix ) for 16 h at 45°C . The arrays were washed and scanned according to standard Affymetrix protocols . GeneSpring 12 . 0 ( Silicon Genetics , Redwood City , CA , USA ) was used for data analysis . Microarray data sets have been deposited in the National Center for Biotechnology Information ( NCBI ) Gene Expression Omnibus under accession number GSE81241 . Splenic CD11b+Gr-1+ cells sorted from CAM- or vehicle-treated mice were lysed in 50 μL of RIPA Lysis and Extraction Buffer ( Thermo Fisher Scientific , Waltham , MA , USA ) supplemented with Halt Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktail 2 ( Sigma-Aldrich ) . The lysate was collected by centrifugation at 14 , 000 × g at 4°C for 15 min and 100 μL of ddH2O was added . Arginase activity was determined using the QuantiChrom Arginase Assay Kit ( BioAssay Systems , Hayward , CA , USA ) following the manufacturer's instructions . The NO concentration in spleen extracts was determined using the Quantichrom Nitric Oxide Assay Kit ( BioAssay Systems ) according to the manufacturer’s instructions . CD3+ T cells were isolated from splenocytes by negative selection , using a magnetic cell sorting system ( MACS; Miltenyi Biotec , Bergisch Gladbach , Germany ) . Sorted cells were washed and resuspended in 5 mL of PBS/0 . 1% bovine serum albumin ( BSA ) . T cell proliferation assays were performed using the CellTrace CFSE Cell Proliferation Kit ( Thermo Fisher Scientific ) according to the manufacturer’s instructions . Briefly , the sorted CD3+ T cells were incubated with CFSE for 20 min , washed , and then stimulated with anti-CD3 antibody ( 5 μg/mL ) ( BioLegend ) and anti-CD28 antibody ( 1 μg/mL ) ( BioLegend ) . The CFSE-labeled CD3+ T cells were then co-cultured with sorted splenic CAM-treated or vehicle-treated CD11b+Gr-1+ cells for 3 days prior to fluorescence intensity analysis by flow cytometry . The Neutrophil Elastase Activity Assay Kit ( Cayman , Ann Arbor , MI , USA ) was used to measure elastase activity according to the manufacturer’s instructions . The Myeloperoxidase ( MPO ) Activity Assay Kit ( Colorimetric ) ( ab105136; Abcam , Cambridge , UK ) was used to measure MPO activity according to the manufacturer’s instructions . The Phagocytosis Activity Assay Kit ( IgG FITC ) ( Cayman ) was used to measure phagocytic activity according to the manufacturer’s instructions . To prepare LPS-treated CD11b+Gr-1+ cells , mice were intraperitonially injected with LPS ( 50 mg/kg ) , and splenic CD11b+Gr-1+ cells 12 h after LPS treatment were sorted using anti-Gr-1 magnetic beads . To prepare thioglycolate-induced neutrophils , mice were intraperitoneally injected with 3% thioglycolate in 2 mL of PBS . Peritoneal neutrophils were recovered 4 h after the injection . For isolation of peripheral neutrophils , the Cell-based Assay Neutrophil Isolation Histopaque kit was used according to the manufacturer’s instructions ( Cayman ) . To prepare monocytes , Monocyte Isolation Kit ( BM ) was used according to the manufacturer’s instructions ( Miltenyi Biotec ) . Bone marrow cells were harvested from 8–10-week-old-male C57BL/6J mice by flushing the femur and tibia with RPMI 1640 medium . Recovered cells were then cultured in bone marrow cell medium ( 20% FCS , 30% L-cell supernatant , 2 mM l-glutamine , 1% penicillin/streptomycin , 0 . 25 μg/mL amphotericin B in RPMI 1640 ) . Fresh bone marrow cell medium was added on day 3 . On day 6 , adherent cells were replated in RPMI 1640 medium supplemented with 10% FCS , 2 mM l-glutamine , and 1% penicillin/streptomycin for use as BMDMs . On day 7 , the medium was removed and BMDMs were cultured with or without CD11b+Gr-1+ cells ( either CAM-treated or vehicle-treated ) for 1 h , followed by stimulation with LPS for 12 h . TNF-α , IFN-γ , and IL-10 levels were measured using a DuoSet ELISA Kit ( R&D Systems ) according to the manufacturer’s instructions . Mice received a single intraperitoneal ( i . p . ) injection of 50 mg/kg LPS ( Escherichia coli O111:B5; Sigma-Aldrich ) . For CAM pre-treatment , mice were intraperitoneally injected with CAM ( 100 mg/kg ) once daily for three consecutive days , followed by LPS challenge the day after the last injection of CAM . For the pharmacological depletion of Gr-1+ cells , mice received an i . p . injection of an anti-Gr-1 antibody ( 250 μg/mouse ) ( clone RB6-8C5; BioLegend ) or isotype control antibody 24 h before LPS challenge or 1 h before starting CAM treatment ( i . e . , 73 h before LPS challenge ) . The antibody dose was determined according to our previous study [55] . For adoptive transfer experiments , splenic CD11b+Gr-1+ cells ( 1 × 106/mouse ) in 0 . 1 mL of PBS sorted from CAM- or vehicle-treated mice were intravenously transferred into mice by tail vein injection . Influenza virus A/H1N1/PR8/34 ( PR8 ) was kindly provided by Dr . Hideki Hasegawa ( National Institute of Infectious Diseases , Tokyo , Japan ) [56] . The PR8 virus titer was measured by a plaque forming assay using Madin-Darby canine kidney ( MDCK ) cells ( CCL-34; American Type Culture Collection ) . S . pneumoniae was kindly provided by Dr . Kimiko Ubukata ( Keio University , Tokyo , Japan ) . The strain was clinically isolated ( serotype 3 ) and was both penicillin- and macrolide-resistant , harboring penicillin resistance ( penicillin binding protein [pbp1a , pbp2x , and pbp2b] ) and macrolide resistance ( mefA and ermB ) genes . Detailed drug susceptibility testing of the strain is shown in S1 Table . The minimum inhibitory concentration ( MIC ) of Penicillin G was 1 , and therefore the strain was identified as clinically penicillin-intermediate S . pneumoniae ( PISP ) according to Clinical and Laboratory Standards Institute ( CLSI ) definition M100 S26:2016 . To establish the post-influenza secondary S . pneumoniae infection mouse model , mice were intranasally inoculated with five plaque-forming units ( PFUs ) of PR8 . Seven days after PR8 infection , the mice were intranasally infected with 1 × 104 colony-forming units ( CFUs ) of S . pneumoniae suspended in 50 μL of PBS . Mice were then intraperitoneally injected with CAM ( 100 mg/kg ) diluted in PBS with 5% gum arabic or vehicle daily starting from 1 h after pneumococcal infection throughout the course of the experiment . Mice were monitored for survival for up to 5 days after pneumococcal infection . A 7-day interval between influenza infection and subsequent S . pneumoniae inoculation was used as a lethal model in the present study , based on previous studies [21 , 57] . For IFN-γ administration , mice were intranasally inoculated with 16 μg/kg recombinant IFN-γ ( PeproTech ) or vehicle at 30 min and 24 h after pneumococcal infection . The dose of recombinant IFN-γ was determined according to a previous report [58] and survival was monitored for 7 days . Bronchoalveolar lavage fluid ( BALF ) was collected with 0 . 7 mL of saline , and total cell counts were determined [59] . Whole left lungs were harvested and homogenized with 1 mL of sterile PBS with 0 . 1% Triton X and serially diluted with sterile PBS . Equal volumes of each dilution of the lung homogenates and blood were applied to trypticase soy agar ( TSA ) plates containing 5% sheep blood ( BD Trypticase Soy Agar II with 5% Sheep Blood; BD Diagnostics ) and incubated overnight at 37°C . After overnight incubation , bacterial colonies were counted and expressed as CFUs per mL of solution . Lung specimens harvested 18 h after pneumococcal infection were fixed with 4% paraformaldehyde . After dehydration and embedding in paraffin , 3-μm sections were stained with hematoxylin and eosin ( H&E ) . Whole lung tissues were homogenized using a Dounce tissue homogenizer on ice in RIPA Lysis and Extraction Buffer ( Thermo Fisher Scientific ) . Homogenates were centrifuged at 12 , 000 × g at 4°C for 20 min , and the supernatants were collected and stored at −80°C until use for protein analysis . Bicinchoninic acid ( BCA ) protein assays ( Thermo Fisher Scientific ) were used to measure the total protein concentration in lung homogenates . Equal amounts ( 15–30 μg ) of cell lysates were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( Bio-Rad Laboratories ) , and the proteins were transferred to a polyvinylidene fluoride membrane ( Bio-Rad Laboratories ) . After overnight incubation with antibodies against prok2 ( Millipore , Billerica , MA , USA ) , STAT3 , or p-STAT3 ( both from Cell Signaling Technology , Danvers , MA , USA ) , the membrane was counterstained with horseradish peroxidase-conjugated rabbit IgG antibody and visualized with enhanced chemiluminescence detection reagents ( GE Healthcare ) . Images were analyzed using ImageJ 1 . 45v ( National Institutes of Health , Bethesda , MD , USA ) . The anti-Bv8 antibody was kindly provided by Genentech . Mice were intraperitoneally injected with the anti-Bv8 antibody ( 5 mg/kg ) or control IgG at 4 days and 1 day before CAM treatment and once daily for three consecutive days during CAM treatment . Total RNA was isolated from the lung or spleen using the RNeasy Mini Kit according to the manufacturer’s instructions ( QIAGEN ) . Total RNA was reverse-transcribed using the High-Capacity RNA-to-cDNA cDNA Kit ( Life Technologies ) . Real-time quantitative polymerase chain reaction ( PCR ) was performed using SYBR Green chemistry on a 7500 Fast Real-Time PCR system ( Applied Biosystems ) . The primer sequences for each target gene are provided in S2 Table . Peripheral blood cells were isolated from healthy subjects . Red blood cells ( RBCs ) were removed from 40 mL of heparinized peripheral blood using Dextran T-500 ( Pfizer ) . After removal any remaining RBC using ammonium-chloride-potassium ( ACK ) lysis buffer ( Lonza , Basel , Switzerland ) , peripheral cells were stained with antibodies described above . Data are expressed as the mean ± SEM and analyzed using the Mann–Whitney-U-test , Wilcoxon matched-pairs signed rank test , or ANOVA , followed by Tukey’s tests for multiple comparisons . Log-rank tests were performed for survival studies . p < 0 . 05 was considered statistically significant . | Myeloid-derived suppressor cells ( MDSCs ) are a heterogeneous population of anti-inflammatory myeloid progenitors that expand in response to acute and chronic inflammation as well as in various diseases , such as autoimmune diseases and cancer . The macrolide antibiotic clarithromycin has immunomodulatory effects in various inflammatory diseases , distinct from its antimicrobial effects , but the mechanism underlying these effects is unknown . The present study demonstrates that clarithromycin treatment induces a marked expansion of CD11b+Gr-1+ MDSC-like cells in the spleen and lungs , sufficient to protect mice from LPS-induced lethality and clarithromycin-resistant bacterial pneumonia via increased IL-10 and decreased IFN-γ levels . Clarithromycin-induced CD11b+Gr-1+ cell expansion was dependent on STAT3-mediated Bv8 production . Moreover , expansion of the immunosuppressive MDSC-like cell population was observed following clarithromycin treatment in humans . Collectively , these results suggest that the immunomodulatory effects of clarithromycin can be attributed to the induction of CD11b+Gr-1+ MDSC-like cells via the STAT3/Bv8 axis . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"blood",
"cells",
"flow",
"cytometry",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"immune",
"cells",
"spleen",
"drugs",
"immunology",
"pulmonology",
"immunosuppressives",
"animal",
"models",
"pneumonia",
"routes",
"of",
"administration",
"model",
"organisms",
"experimental",
"organism",
"systems",
"pharmacology",
"neutrophils",
"research",
"and",
"analysis",
"methods",
"white",
"blood",
"cells",
"animal",
"cells",
"t",
"cells",
"mouse",
"models",
"spectrophotometry",
"cytophotometry",
"cell",
"biology",
"intraperitoneal",
"injections",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"spectrum",
"analysis",
"techniques"
]
| 2018 | Clarithromycin expands CD11b+Gr-1+ cells via the STAT3/Bv8 axis to ameliorate lethal endotoxic shock and post-influenza bacterial pneumonia |
Integrins are heterodimeric ( αβ ) cell surface receptors that are activated to a high affinity state by the formation of a complex involving the α/β integrin transmembrane helix dimer , the head domain of talin ( a cytoplasmic protein that links integrins to actin ) , and the membrane . The talin head domain contains four sub-domains ( F0 , F1 , F2 and F3 ) with a long cationic loop inserted in the F1 domain . Here , we model the binding and interactions of the complete talin head domain with a phospholipid bilayer , using multiscale molecular dynamics simulations . The role of the inserted F1 loop , which is missing from the crystal structure of the talin head , PDB:3IVF , is explored . The results show that the talin head domain binds to the membrane predominantly via cationic regions on the F2 and F3 subdomains and the F1 loop . Upon binding , the intact talin head adopts a novel V-shaped conformation which optimizes its interactions with the membrane . Simulations of the complex of talin with the integrin α/β TM helix dimer in a membrane , show how this complex promotes a rearrangement , and eventual dissociation of , the integrin α and β transmembrane helices . A model for the talin-mediated integrin activation is proposed which describes how the mutual interplay of interactions between transmembrane helices , the cytoplasmic talin protein , and the lipid bilayer promotes integrin inside-out activation .
Integrins are cell surface receptors involved in many essential cellular processes , such as cell migration , and in pathological defects , such as thrombosis and cancer [1] . Integrins are αβ heterodimers . Each subunit has a large ectodomain , a single transmembrane ( TM ) helix and a short flexible cytoplasmic tail [2] . Integrins are crucial for many signal transduction events [3]–[6] . Unusually , they can transmit signals in both directions across the cell membrane [2] , [7] . In the inside-out activation pathway , formation of a complex between talin , the integrin β cytoplasmic tail , and the membrane is thought to shift the integrin conformational equilibrium towards an active state [8]–[15] . Activation is believed to proceed via changes in TM helix packing [2] , [3] coupled to substantive conformational changes in the integrin ectodomain . Talin consists of a head domain ( ∼50 kDa ) and a large rod domain ( ∼220 kDa ) [16] , [17] . A crystal structure of the talin head domain revealed a novel linear arrangement of four subdomains , F0 , F1 , F2 and F3 ( Fig . 1A ) . Although the head domain has sequence homology with other FERM ( band four-point-one , ezrin , radixin , moesin ) domains [18] , the linear arrangement of the subdomains , an inserted loop in the F1 domain , and the extra F0 domain confer on talin significantly different features from canonical FERM domains [19] . Experimental studies of the talin head have yielded valuable insights into the role of the different subdomains [8] , [12] , [17] , [20] , [21] . A key step in integrin activation is binding of the F3 subdomain to the integrin β tail [9] , [10] , [12]–[14] . The other subdomains have also been shown to contribute to integrin activation , although they do not bind directly to the integrin β subunit [8] . The F2 subdomain has a positively charged patch which has been shown to enhance activation by interacting with negatively charged lipid headgroups in the membrane [11] , [21] . The F0 and F1 subdomains have an ubiquitin-like fold [17] , [20] along with a flexible , positively charged loop of ∼40 residues inserted in the F1 domain . This loop ( which was removed to facilitate structure determination of the head domain [17] ) has been proposed to form a transient helix stabilized by interactions with acidic phospholipids [20] . Structural studies of the talin head domain and fragments have also suggested that the F2–F3 and the F0–F1 domain pairs are relatively rigid but the pairs are connected by a flexible linker [17] . The inactive state of integrins is in part maintained by interactions between the α and β TM helices . Important interactions are defined by regions known as the outer ( OMC ) and inner ( IMC ) membrane clasps ( Fig . 1B ) . In the OMC , a GxxxG motif in the α integrin TM region allows close packing with the β integrin helix whereas in the IMC there are hydrophobic interactions involving a cluster of phenylalanines and a salt bridge between the α and β chains [22] . Various models for TM helix rearrangements , including ‘scissor’ or ‘piston’ movements and increased helix separation have been proposed to explain activation and trans-membrane signaling [10] , [23]–[29] but there remains a need to distinguish and refine these models . Molecular dynamics ( MD ) simulations allow us to explore the conformational dynamics and lipid interactions of membrane proteins [30] . Multiscale approaches combine coarse-grained molecular dynamics ( CG-MD ) simulations [31] , [32] , which extend the time scales that can be studied , with subsequent all-atom simulations , which allow refinement of the system [33] . We previously used this approach to develop a model that explained how the F2–F3 fragment of the talin head domain associates with the plasma membrane in a way that led to a scissoring movement of the two integrin TM helices [34] . In the current study , we build upon these studies to elucidate the interactions between the complete talin head domain ( i . e . domains F0–F3 ) , a lipid bilayer , and the α/β integrin TM regions . Our results provide novel information about the orientation of the intact talin head in complex with the lipid bilayer , and about changes induced in the TM helical regions . Overall , the results reveal how binding of talin to the membrane and to integrins tails leads to integrin activation .
We have used a serial multiscale MD simulation [35] approach to explore the dynamics of the talin head ( F0–F3 ) domain , its interaction with lipid bilayers , and the resultant conformational changes of a talin/integrin TM complex embedded in a phospholipid bilayer . This approach has previously been used to explore the interactions of a number of peripheral proteins with membrane surfaces and lipids [11] , [36]–[39] , of TM helices within a lipid bilayer [40] , [41] , and of more complex signaling and related assemblies within membranes [34] . It enables one to combine coarse-grained ( CG ) simulations of membrane association and related events , with more detailed atomistic simulations to refine the resultant models . It thus provides a complementary approach to extended atomistic simulations [42] . The principal simulations underlying the current study are summarized in Table 1 and in Fig . 2 . The talin head domain crystal structure ( PDB:3IVF [17] ) lacks a long loop region in the F1 domain and therefore atomistic simulations of the talin head domain ( tal-sol-AT ) in solution were first used to explore potential internal flexibility between the four component sub-domains ( F0–F3 ) along with possible conformations of the F1 domain insertion . For these simulations the insertion in the F1 domain ( res: 134–172 ) was modeled in a random coil conformation , and was located away from the F0–F1 pair ( see Fig . 3B and Fig . S1A ) using Modeller 9v8 ( http://salilab . org/modeller/ ) [43] , [44] . This configuration of the loop allows exploration of all possible conformations/orientations and selection of a preferred conformation/orientation relative to the talin head domain . The conformation of the talin head domain suggested by the above simulations was used to model the association of the talin head domain with a phospholipid bilayer . Since NMR studies suggested a helical propensity for the region involving residues 154–167 [20] this region was modeled as an α-helix ( tal-h2F0-CG and Fig . S1 ) . Subsequently , the same multiscale simulation approach was used to explore the dynamic behavior of a talin/TM integrin complex in a bilayer ( αβ-talh2-CG ) on a multi-microsecond timescale . The talin head domain/αβ complex was constructed as described in Kalli et al . [34] using the αΙΙbβ3 TM region NMR structure [22] , the F2–F3/β1D complex crystal structure [12] and the talin head domain configuration obtained from the talin head domain simulations described in this study . An ‘open’ model generated by these CG simulations was subsequently explored via a microsecond duration atomistic simulation ( αβ-talh2o-AT ) . A number of control simulations were also performed to evaluate the robustness/sensitivity of the results and to explore the contributions of different regions and interactions ( e . g . flexibility within the domain , electrostatic interactions and other helical conformations in the F1 loop ) to the binding of the talin head to anionic lipid bilayers . Detailed descriptions of these simulations are provided in the Supporting Information ( Tables S1 and S2 ) . In total our study amounts to ca . 60 µs of CG-MD and ca . 2 µs of atomistic molecular dynamics simulations ( AT-MD ) simulation time . To study the conformational dynamics of the talin head domain prior to the association with the bilayer atomistic ( AT-MD ) simulations of the talin head domain ( i . e . subdomains F0 to F3 ) in aqueous solution in the absence of a bilayer were performed ( tal-sol-AT in Table 1 ) . During these simulations the flexible linker between the F2–F3 and F0–F1 pairs allowed transient displacement of the F0–F1 subdomain relative to the F2–F3 subdomain with the angle defined in Fig . 3A . This angle , equal to 0° for a linear arrangement of F0-F1-F2-F3 as seen in the crystal structure , ranged from 0° to 90° in the simulations . During these simulations the long loop in the F1 domain moved closer to the F0–F1 pair , and adopted an extended conformation on the same side of the protein as the positively charged patch on F2 ( Fig . 3B ) . Calculation of the electrostatic field around the talin head conformation observed at the end of these simulations suggests that localization of this loop close to the F0–F1 pair creates an extensive positively charged surface on one side of protein; this could facilitate strong talin/bilayer interactions ( Fig . S2 ) . Despite experimental evidence for an α-helical propensity in the F1 loop [17] , no helix formation was detected ( this might be due to insufficient simulation time for a coil-to-helix conformational transition to occur ) . Although there was a relatively large change in the angle between the F0–F1 and F2–F3 domain pairs , no significant angle change was observed within either the F0–F1 or the F2–F3 domain pair , suggesting that each pair behaves approximately as a rigid body . Having established in the tal-sol-AT simulations ( see above ) that the F1 loop interacts with F0–F1 to form a positively charged surface that extends the positive patch on F2–F3 , CG simulations with the loop in this location were performed to explore the nature of the interactions of the complete talin head domain with an anionic lipid bilayer . Note that in this simulation system a small helical region ( h2 helix; see Fig . S1 ) was included within the F1 loop as indicated by NMR data [20] . During this modeling of the loop the remainder of the structure , with the exception of the region modeled as helical ( res: 154–167 ) , was restrained to maintain the talin conformation derived from the above simulations . These restraints were removed during the simulations . In the tal-h2F0-CG simulation ( Table 1; Fig . 4 and Fig . 5 ) , talin was observed to associate with the bilayer in four out of five simulation and in all four of these simulations talin bound to the bilayer initially via a basic loop ( res: 318–330 ) in the F3 domain , and subsequently via the positively charged patch in the F2 domain ( res: 255–285 ) ( Fig . S3A ) . Contacts were defined by using a distance cut-off of 7 Å between the protein residues and the lipids . These two regions have been identified previously [11] to promote productive binding of the isolated F2–F3 fragment to an anionic lipid bilayer . The additional surface created by the F1 loop also interacted with the bilayer ( Fig . 4C ) . Interestingly , in all simulations which resulted in a talin/bilayer complex , the talin head domain with the F1 loop adopted a V-shaped conformation due to rotation of the F0–F1 pair relative to the F2–F3 pair in the bilayer plane ( Fig . 4A ) . The reorientation of the F2–F3 and F0–F1 domain pairs prior to the binding to the bilayer was also observed in other simulations in which a different starting conformation of talin was used ( e . g . the talin crystal structure; data not shown ) . In contrast to the more dynamic variations in the angle between the F0–F1 and F2–F3 observed when talin was in solution ( see above ) , the V-shaped conformer was stabilized by association with the bilayer ( Fig . 4B ) . This conformation optimizes talin/lipid interactions and induces a more compact arrangement of domains , although this new arrangement is still different from the linear arrangement in the X-ray structure [17] and the canonical FERM domain packing of F0 to F3 [18] . During the AT-MD simulations ( tal-h2F0-AT; see Table 1 ) that started from the final snapshot of the tal-h2F0-CG simulation , the V-shaped conformation of talin was retained , with talin interacting preferentially with the headgroups of the anionic POPG lipids ( Fig . S6B ) . No restrictions in the position/flexibility of the loop or the domains were imposed in the AT-MD simulations . Simulations of the talin head domain with a neutral bilayer ( containing 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidyl-choline ( POPC ) lipids ) resulted in no association of talin with the bilayer ( Fig . S3B and Text S1 ) . These results are in good agreement with the available experimental data [11] , [12] , [17] and augment our previous observation that electrostatic interactions are important in regulating the formation of a talin/membrane complex . Control CG simulations starting with the talin head domain crystal structure ( i . e . without the F1 loop ) with the same POPC/POPG ( 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidyl-choline/1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidyl-glycerol ) bilayer showed that the positively charged region of F0–F1 augmented by the F1 loop promotes direct F0–F1/bilayer interactions ( data not shown ) . CG models of talin with restricted flexibility between the F0–F1 and F2–F3 domain pairs did not bind to the bilayer in a way that would facilitate binding of F2–F3 to the β-tail in any of these simulations ( see Text S1 ) . Disruption of the electrostatic interactions between talin and the membrane ( simulating with experimentally tested mutations [10] , [12] ) also resulted in the ‘non-productive’ orientation of the talin head domain relative to the membrane ( see Methods for description of mutations ) . An orientation is judged here to be ‘non-productive’ when the talin/bilayer complex formed is incompatible with binding to the β integrin cytoplasmic region in a manner similar to that observed by Anthis et al . [12] ( see Text S1 ) . Overall , our simulations suggest that optimal association of talin with the membrane is enhanced by conformational flexibility within the head domain , especially between the F0–F1 and F2–F3 pairs . This flexibility facilitates optimal interactions of the V-shaped conformation of talin with the headgroups of anionic lipids . Reduction in the anionic lipid content significantly decreased the talin/bilayer association and , in the presence of mutations that disrupted the talin/lipid electrostatic interactions , talin bound to the bilayer in a perturbed orientation ( Fig . S5 ) . Having established the nature of the interactions of the intact talin head domain with a bilayer , we set out to explore the impact of these interactions on the structure of the integrin TM region . To this end , the talin head domain with optimal bilayer interactions ( i . e . from simulation tal-h2F0-CG ) was modeled in a complex with the α/β integrin TM complex ( see Methods for details of modeling this complex ) . This complex was inserted in a POPC/POPG bilayer ( Fig . 6A ) , and five extended CG-MD simulations ( αβ-talh2-CG; Table 1 ) were performed . In these simulations all restraints between the integrin α subunit TM domain and the F2–F3/β TM complex were removed to allow the α TM domain to move relative to the rest of the complex . Closer examination of the movement of the individual domains within the complex during the simulations revealed a reorientation of the talin head domain relative to the bilayer surface , comparable to the reorientation observed in our earlier studies of the F2–F3/αβ complex ( Fig . 6B ) . This rotation , in turn , induced a ∼25° rotation of the β TM helix perpendicular to the bilayer normal , similar to that seen in our previous studies ( Fig . 6B ) [34] . This resulted in the disruption of the interactions in the OMC and IMC regions of the α/β TM segment which were shown to maintain the integrin inactive state [22] , [27] , [29] , [45]–[52] . In particular , the β TM helix rotation perturbed the close packing in the OMC region ( because of the 972GxxxG976 motif in the α TM helix ) and disrupted the hydrophobic interactions in the IMC region formed by F992 and F993 residues and the αIIb995β3723 salt bridge . Calculation of the angle between the F2–F3 and the F0–F1 domain pairs during the simulations revealed a stable angle of ∼60° ( the definition of the angle is the same as described above ) , indicating that the V-shaped conformation of the talin head was retained in all the simulations ( Fig . S7A ) . We note that simulations of the α/β dimer and the isolated β integrin tails in the same bilayer were performed in our previous study [34] . To explore the talin/αβ complex in more detail , a structure which represented an “open” state of the integrin TM domain was selected from the αβ-talh2-CG simulation ( using similar criteria to those described in our previous simulations of the α/β dimer and the isolated β integrin tails in a bilayer [34] ) and converted to an AT representation . In this “open” integrin model , the interactions in both the IMC and OMC regions are disrupted . We have previously suggested that the scissoring motion of integrin TM helices may be “a trigger for inside-out activation” [34] . To explore the consequences of activating this trigger , we performed an extended ( microsecond ) AT-MD simulation of the complex ( αβ-talh2o-AT; Table 1 ) . At the end of this atomistic simulation the V-shaped conformation of the talin head domain was retained . Calculation of the inter-helical distances in the OMC and IMC regions revealed an increase in the helix-helix distance in both the IMC and OMC regions ( Fig . 7A , B ) , corresponding to dissociation of the α and β TM helices . Dissociation was preceded by a ‘scissoring’ movement similar to that seen in our previous studies of the F2–F3/αβ complex but with more extensive movement of the TM helices and a large increase in the tilt angle of the β helix relative to the bilayer normal , to a final value of 40° , ( Fig . S7B ) . On a similar simulation timescale , the F2–F3 fragment alone induced a much smaller separation of the α/β subunits ( Fig . S8B ) with a comparable increase in the β TM helix tilt angle to that observed in our previous studies [34] . This suggests that the F2–F3 domains are sufficient to modulate the β tail tilt angle but the entire head is more effective in producing TM helix separation . The increase in tilt angle induced by the talin head increases the extent of membrane-embedding of the β tail TM region . In the final orientation seen in the simulation with the talin/αβ complex , residues from K716 up to K725 are embedded in the membrane ( Fig . 8 ) . A similar increase in the extent of the membrane-embedded region was observed in a recent experimental study [53] . Despite a large increase in β-TM tilt angle , the T715 sidechain remains oriented toward the lipid phosphate atoms , but the K716 sidechain is no longer in contact with the lipid headgroups . This is in good agreement with experimental data that identified interactions between the β3 K716 ε-amine group and the lipid phosphate in the integrin inactive state , suggesting that this interaction controls the tilt angle of the β3 integrin tail [54] . Mutation of K716 in these experimental studies shifted the integrin conformational equilibrium towards an active state , possibly by perturbing the β subunit tilt angle and the α/β TM region crossing angle . Thus , in the active state one would expect weaker K716/lipid headgroup interactions , as observed in our simulations of the integrin active state . The tendency of the β TM helix to adopt a tilted orientation in the membrane is also suggested by other experimental [55] and computational [21] data . Our studies show that conformational changes in the talin head on binding to anionic phospholipid membranes mediate transmembrane signaling by the integrin TM helix dimer . Simulations have revealed how optimization of the interactions of the complete talin head domain ( F0–F3 ) with an anionic phospholipid bilayer promotes a rearrangement of the subdomains , which in turn initiates a conformational rearrangement of the integrin TM region . In particular , on binding to the membrane the F0–F3 talin head domain undergoes a conformational change to a V-shape , rather than the linear arrangement of the F0-F1-F2-F3 subdomains observed in the crystal structure . Formation of a complex between the rearranged talin head and the integrin TM domain subsequently triggers separation of the TM helices i . e . disassembly of the TM helix dimer . The first stage of this disassembly is an initial scissoring motion of the helices , as seen in our previous studies [34] . The current study reveals how the flexible linker between the F0–F1 and F2–F3 pairs allows talin to adopt a conformation which optimizes its contacts with anionic headgroups of lipids ( Fig . 4A ) . This correlates well with the studies by Bouaouina et al . [8] which have suggested that the activation responses of β3 and β1 integrins have different dependencies on talin head fragments , indicating that formation of an optimal configuration of the talin/membrane complex could also have mechanistic importance . Our simulations reveal how electrostatic interactions between the protein and anionic lipid headgroups orient the talin head domain and optimize its interactions with the membrane . The cationic surface of the talin head , which binds to the anionic lipids , is formed by the F2 and F3 domains plus the F1 loop . In silico mutations ( i . e . K324D and K256E , R277E , K272E , R274E ) of this surface perturb binding of talin to the membrane , in agreement with experimental studies [12] , [17] , [20] , [21] . This binding surface is consistent with NMR [10] , crystallographic [12] and TIRF microscopy studies [21] . Experimental studies indicate that inserted loop in the F1 domain is dynamic in nature [17] , [20] . Our simulations suggest that it provides a binding surface close to the cationic surface of F2–F3 . In all simulations that yielded a ‘productive’ orientation of the talin head domain on the membrane , the F2–F3 pair always associated prior to F0–F1 . This agrees with recent data that the association constant of the complete talin head domain with lipids is similar to that of the F2–F3 fragment [56] . We note that in kindlin , a homolog of talin that co-activates integrins , an even longer lysine-rich loop is inserted in the F1 domain . This lysine rich loop in kindlin is highly conserved and is believed to support binding to anionic phospholipid head groups [57] . Simulations that included the entire talin head/membrane/TM complex revealed that talin interactions with the integrin β tail and the membrane surface disrupt the interactions of both the OMC and IMC regions , resulting in eventual dissociation of the TM helices . Calculation of the hydrogen bonds between the lipids and the membrane ( see Fig . S7C ) after the disruption of interactions in the TM region show an increase in the number of lipid/talin hydrogen bonds . This suggests a stronger association between talin and the bilayer after formation of the talin/αβ complex . This could explain why the intact talin head domain has more dramatic effect than F2–F3 alone . The crystal structure of the integrin ectodomain and cysteine disulfide mapping of an intact integrin [58] , [59] suggest that the α and β domains are in close proximity to one another in the inactive state . Thus a scissoring movement followed by dissociation of the two TM helices provides a plausible model for how the TM domain may trigger a conformational change in the ectodomain leading subsequent adoption of an extended active state . From a more general perspective , this study reveals the interplay of membrane interactions and conformational changes involved in transmembrane signaling by receptors and associated proteins and/or domains . In particular , it may be compared with recent simulation studies , e . g . of the EGF receptor [42] , which suggested that substantive repacking of the TM helix dimer and interactions between the intracellular kinase domain and anionic lipids play a key role in signaling across a membrane . The mechanisms in these two classes of membrane receptors ( i . e . integrins and receptor tyrosine kinases respectively ) may be compared with movements of TM helices thought to mediate signaling in GPCRs ( as revealed by crystallographic , NMR and simulation studies [60] ) . One possible consequence of the extensive movements of TM helices is that signaling mechanisms are likely to be modulated by changes in ( local ) lipid bilayer properties [61] . This clearly merits further investigation by both computational and biophysical approaches .
The CG-MD simulations were performed using a local variant [31] , [62] of the MARTINI forcefield [32] . A mapping of approximately 4∶1 heavy atoms to CG particles was used . Harmonic restraints ( i . e . an elastic network model; ENM ) between backbone particles within a cut-off distance of 7 Å was applied with a harmonic restraint force constant of 10 kJ/mol/Å2 . In the tal-l25-CG and tal-l50-CG simulations ( Table S1 ) the force constant for the ENM was set to 25 kJ/mol/Å2 and 50 kJ/mol/Å2 respectively and the cut-off distance was increased to 10 Å . The bilayer was constructed by self-assembly CG-MD simulations . In these simulations the lipids were placed randomly within a simulation box and solvated with CG water molecules and ions to neutralize the system . Subsequently , a production simulation was performed for 200 ns . After the first 10–15 ns of simulation the bilayer formed with an equal distribution of lipids in the two leaflets . For the simulation systems discussed here , two different bilayers were constructed . The first bilayer contained 832 zwitterionic POPC lipids and the second had 512 POPC and 320 POPG lipids ( ratio of 3∶2 ) . In the CG simulations the center of mass of the protein was placed 120 Å from the center of mass of the preformed bilayer ( Fig . S3A ) . This starting distance between protein and bilayer was chosen to be much larger than the cut-off distance used for the electrostatic and the van der Waals terms in the CG forcefield . All systems were subsequently solvated with CG water molecules and neutralized with CG sodium particles , energy minimized for 250 steps and equilibrated for 5 ns with the protein Cα particles restrained ( force constant 10 kJ/mol/Å2 ) . Finally , CG-MD simulations were performed . The final snapshot of the CG-MD simulation was converted to an atomistic ( AT ) representation , using a fragment-based approach [33] , for further refinement . For the simulations with the talin/αβ complex , the same POPC/POPG bilayer as above was used . The TM region of the talin/αβ complex was inserted in the bilayer using GROMACS . The lipids that overlapped with the integrin TM region were removed . The same energy minimization and equilibration steps were performed as described above . A modified CG model was used where all the ENM restraints between the integrin α TM region and the rest of the complex were removed . All CG-MD simulations were performed using GROMACS 4 . 5 ( www . gromacs . org ) [63] , [64] . A Berendsen thermostat [65] was used for temperature coupling with a coupling constant of 1 . 0 ps and a reference temperature of 310 K . The Lennard-Jones and Coulombic interactions were shifted to zero between 9 Å and 12 Å , and 0 to 12 Å respectively . The time step was 20 fs . A Berendsen barostat was used for pressure coupling . The coupling constant was 1 . 0 ps , the compressibility was 5 . 0×10−6 bar−1 and the reference pressure was 1 bar . The AT-MD simulations were performed using the GROMOS96 43a1 forcefield [66] . The Parrinello-Rahman barostat [67] and the Berendsen thermostat [65] were used for pressure and temperature coupling , respectively . The bond length was constrained using the LINCS algorithm [68] and the particle mesh Ewald ( PME ) algorithm [69] was used to model long-range electrostatic interactions . A cut-off distance of 10 Å was used for the van der Waals interactions . All the AT simulation systems were energy minimized using a steepest descent algorithm and equilibrated for 2 . 5 ns with the protein Cα atoms restrained ( force constant 10 kJ/mol/Å2 ) . Subsequently , unrestrained AT-MD simulations were performed . All the analyses were performed using GROMACS ( www . gromacs . org ) [63] , [64] , VMD [70] and locally written codes . | Transmission of signals across the cell membrane is an essential process for all living organisms . Integrins are one example of cell surface receptors ( αβ ) which , uniquely , form a bidirectional signalling pathway across the membrane . Integrins are crucial for many cellular processes and play key roles in pathological defects such as cardiovascular diseases and cancer . They are activated to a high affinity state by the intracellular protein talin in a process known as ‘inside-out activation’ . Despite their importance and the existence of functional and structural data , the mechanism by which talin activates integrin remains elusive . In this study we use a multi-scale computational approach , which combines coarse-grained and atomistic molecular dynamics simulations , to suggest how the formation of the complex between the talin head domain , the cell membrane and the integrin moves the integrin equilibrium towards an active state . Our results show that conformational changes within the talin head domains optimize its interactions with the cell membrane . Upon binding to the integrin , talin facilitates rearrangement of the integrin TM region thus promoting integrin activation . This study also provides a demonstration of the strengths of a computational multi-scale approach in studies of membrane interactions and receptor conformational changes and associated proteins that enable transmembrane signaling . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
]
| []
| 2013 | Conformational Changes in Talin on Binding to Anionic Phospholipid Membranes Facilitate Signaling by Integrin Transmembrane Helices |
Mitochondrial DNA ( mtDNA ) variants have been traditionally used as markers to trace ancient population migrations . Although experiments relying on model organisms and cytoplasmic hybrids , as well as disease association studies , have served to underline the functionality of certain mtDNA SNPs , only little is known of the regulatory impact of ancient mtDNA variants , especially in terms of gene expression . By analyzing RNA-seq data of 454 lymphoblast cell lines from the 1000 Genomes Project , we found that mtDNA variants defining the most common African genetic background , the L haplogroup , exhibit a distinct overall mtDNA gene expression pattern , which was independent of mtDNA copy numbers . Secondly , intra-population analysis revealed subtle , yet significant , expression differences in four tRNA genes . Strikingly , the more prominent African mtDNA gene expression pattern best correlated with the expression of nuclear DNA-encoded RNA-binding proteins , and with SNPs within the mitochondrial RNA-binding proteins PTCD1 and MRPS7 . Our results thus support the concept of an ancient regulatory transition of mtDNA-encoded genes as humans left Africa to populate the rest of the world .
Genetic variants in the nuclear and mitochondrial genomes have been traditionally used to trace the ancient global migration paths of different human populations [1 , 2] . Such studies assumed most variants to be neutral and hence merely reflective of the age of the populations studied . However , accumulating evidence suggests that many common variants in the mitochondrial and nuclear genomes have functional impacts [3] . Specifically , ancient mitochondrial DNA ( mtDNA ) variants and genetic backgrounds ( haplotypes , haplogroups ) have been associated with an altered tendency to develop a variety of complex traits [3–5] that affected mitochondrial activity in cell culture experiments [6–9] and which conferred adaptive advantages over the course of human evolution [10–12] . Whereas it has been shown that mtDNA variants affected mitochondrial protein activities , such as oxidative phosphorylation ( OXPHOS ) or the production of reactive oxygen species ( ROS ) , the impact of mtDNA variants on the regulation of gene expression has drawn little attention . In the nucleus , variants that affect gene expression ( eQTLs ) can be mapped within intergenic regions , introns and exons [13 , 14] . Unlike the nuclear genome ( nDNA ) , most ( ~93% ) of the human mtDNA contains intron-less genes , with the majority of known gene regulatory elements being mapped within the major mtDNA non-coding control region , the D-loop . These transcriptional regulatory elements include the heavy strand and light strand promoters , as well as the three conserved sequence blocks ( CSBs I-III ) . The only known mtDNA gene regulatory element that maps within the coding region is recognized by members of the transcription termination factor family ( mTERF ) [15] . These proteins correspond to the main regulatory elements that modulate the expression of the entire set of mtDNA genes ( N = 37 ) , including those encoding 13 protein subunits of the OXPHOS pathway and 24 RNA components of the mitochondrial translation machinery ( 22 tRNAs and two rRNAs ) . All known regulators of mtDNA transcription are imported as proteins from the nucleus [16] , namely mitochondrial RNA polymerase ( POLRMT ) , mitochondrial transcription factors A ( TFAM ) and B2 ( TFB2 ) , and mTERF [15] . Recently , however , we and others have shown that additional nDNA-encoded transcription factors , such as MEF2D , the estrogen receptor , c-Jun and Jun-D are imported into mitochondria , where they bind the mtDNA within the coding region outside the D-loop to regulate transcription [17–20] . These findings not only suggest that mtDNA transcriptional regulation is more complex than once thought but also imply that the quest for genetic variants that affect the regulation of mitochondrial gene expression should not be limited to non-coding mtDNA sequences . The study of eQTLs in the mtDNA lags far behind that of nDNA eQTLs . We were the first to show that an ancient mtDNA control region variants affected in vitro transcription and mtDNA copy numbers in cells sharing the same nucleus but differing in their mtDNAs ( i . e . , cytoplasmic hybrids or cybrids ) [21] . Subsequently , two studies that measured mtDNA transcript levels ( among other mitochondrial activities ) in cybrids revealed differences among certain mtDNA haplogroups [22 , 23] . Despite these advances , a worldwide overview of the landscape of mtDNA transcriptional differences in human populations , similar to what is known of nDNA gene expression [24–26] , remains lacking , as do mechanistic explanations for the specific observations made . Such an overview is an essential step towards highlighting candidate mtDNA eQTLs . With this aim in mind , we analyzed RNA-seq data that was recently made available as part of the 1000 Genomes Project for 454 unrelated individuals of Caucasian and sub-Saharan African origin . Our analyses indicated that samples carrying a combination of certain mtDNA variants presented a distinct mtDNA gene expression pattern . Strikingly , the most prominent finding was that all of samples carrying the African mtDNA haplogroup L diverged from the rest in their pattern of expression , suggesting that mtDNA gene expression diverged between people who left Africa and those who remained in the continent , supporting an ancient regulatory difference . Furthermore , the association of such mtDNA gene expression patterns with SNPs within known regulators of mtDNA gene expression shed light on the possible mechanism underlying this phenomenon .
Levels of gene expression can vary among individuals , tissues and species [27] . As such , we utilized RNA-seq experiments to assess differential mitochondrial gene expression patterns among individuals and ethnicities ( Fig 1 ) . To this end , we sought RNA-seq studies addressing a variety of human populations . As a first step , we attempted to compile available RNA-seq datasets from various populations [26 , 28–31] to generate the largest and most diverse studied cohort . However , expression pattern clustering analysis grouped RNA-seq samples according to the study of origin , even when considering the same samples that were separately sequenced and analyzed independently by different groups ( S1 Fig ) , thus arguing against co-analysis of RNA-seq data generated by different protocols . Hence , although Sudmant et al . [27] recently showed that differences in gene expression patterns between tissues are greater than are differences between studies , our results reveal that while focusing on a single tissue , differences in gene expression patterns between studies exceeds differences among individuals . Therefore , to avoid such artifacts , we focused our analysis on the largest of the relevant studies , encompassing 462 publicly available RNA-seq samples from Caucasians and sub-Saharan Africans [26] , all part of the 1000 Genomes Project [32] . This dataset included results from mRNAs and rRNAs sequencing libraries , here referred as the ‘long RNA’ dataset , as well as short-reads sequencing libraries that includes mtDNA-encoded tRNAs ( i . e . the tRNA dataset ) . In that study , all samples were randomly distributed to seven laboratories and RNA-seq data was generated following an identical shared protocol . In considering the 462 RNA-seq samples , eight of the long RNA dataset did not successfully map to human nDNA and mtDNA reference genomes . Our analysis indicated that this problem stems from uneven numbers of paired reads ( STAR mapping criterion ) , which may reflect lower data quality . To avoid possible technical biases we excluded the mentioned 8 samples from further analysis . The number of reads per base that mapped to mtDNA in the remaining 454 long RNA samples ranged from several hundred in the case of tRNA genes , to nearly half a million for some protein-coding genes ( S2A Fig ) . Sequencing reads corresponding to tRNAs were under-represented in the long RNA dataset likely due to the library preparation protocols used , which involved a size selection step . We partially overcame this limitation by analyzing the tRNA dataset . Here , 16 of the 22 tRNA genes were represented in the tRNA dataset with sufficient numbers of mapped reads for analysis in at least 90% of the samples . For the sake of consistency , we included only those individuals who were represented in the long RNA dataset when considering the tRNA dataset , thereby retaining 440 samples with coverage of up to tens of thousands mapped reads per mtDNA base ( S2B Fig ) . Lappalainen et al . [26] reported that there were no significant differences between the same samples that were generated in different laboratories . This enabled us to divide the entire dataset into two groups matched according to ethnicity and gender ratio , which were separately treated as biological replicates . Recently , we , and others , identified RNA-DNA differences ( RDD ) in three mtDNA sites common to all human individuals and human tissues tested to date [33 , 34] , as well as in 90% of the vertebrates [35] . Our sequence analysis of the long RNA dataset also identified these three RDDs in all analyzed samples , while excluding sequencing and mapping errors by analysis filters , as previously outlined [33] , thus further supporting the quality of our analyzed data . The polycistronic nature of mtDNA transcription permitted RNA-seq reads that covered the complete mtDNA of all tested individuals . This enabled reconstruction of the entire mtDNA sequence in all analyzed samples , which were aligned to reveal polymorphic positions and reconstruct a phylogenetic tree ( S3A Fig ) . Such analysis enabled the assignment of all individuals to specific mtDNA haplogroups . Since the analyzed samples are part of the 1000 Genomes Project , we extracted the mtDNA sequence from the DNA sequence database of the same samples and used these to construct a phylogenetic tree ( S3B Fig ) . Notably , the topologies and distribution of haplogroups throughout the RNA and DNA-based trees were nearly identical and were in agreement with previously published trees [12 , 36] . Therefore , putative human RNA-DNA sequence differences did not affect overall mtDNA tree topology . We and others previously showed that nDNA harbors a repertoire of mtDNA sequence fragments ( NUMTs ) that were transferred from the mitochondria during the course of evolution [37 , 38] . NUMTs potentially pose an obstacle to mtDNA gene expression assessment , as a subset of RNA reads might originate from NUMTs rather than from the active mtDNA . As a first step to control for such a scenario , we remapped the RNA-seq reads of each sample against their own reconstructed mtDNA sequence ( personalized mapping ) . This approach also controlled for a second possible bias . It is conceivable that gene expression level differences could be affected by exclusion of sequencing reads due to mapping of the RNA-seq reads to a single European reference sequence ( the revised Cambridge Reference Sequence ( rCRS ) , i . e . rCRS mapping ) , resulting in the exclusion of numerous variants [39] . Since the mtDNA is highly variable and since ~130 , 000 years separates the appearance of the L haplogroup from the remaining mtDNA genetic backgrounds analyzed [10 , 40] , our sample-specific analysis enforced increasingly accurate and unique read mapping while excluding erroneous mapping to more than a single locus . Secondly , paired-end technology enabled us to exclude reads whose paired read partner mapped to a non-mtDNA locus . Finally , we repeated our read mapping while avoiding the unique-mapping step and compared the results obtained to those realized by unique-mapping analysis . The latter analysis did not reveal any skew in the expression pattern observed for those samples analyzed . We , therefore , concluded that NUMTs had little or no impact on our data . We asked whether certain mtDNA SNPs associate with differential expression levels of mtDNA-encoded genes . Since we analyzed multiple mtDNA SNPs ( including both singletons and lineage-defining SNPs ) , Bonferroni correction for multiple testing was applied . As mentioned above , initial analysis was performed while randomly dividing the samples into two groups while retaining the proportions of gender and ethnicity . Such analysis , using the personalized mapped samples , revealed correlation between certain SNPs and a distinct expression pattern . Close inspection revealed that all these SNPs corresponded to mtDNA haplogroup L ( Fig 2 , S1 Table , S2 Table ) . It is worth noting that analysis based on either the personalized- or rCRS-mapped samples led to comparable expression patterns ( S4 Fig ) . This was despite the fact that the personalized mapping exhibited with excess of mapped reads in L halogroup samples , i . e . a mean of additional 26 , 197 reads per sample–a 0 . 09% increase , in the personalized mapping samples . Similarly , there was a slight increase in the number of reads in personalized mapped Caucasian samples , i . e . a mean of additional 5 , 279 reads per sample–a 0 . 02% increase . Taken together , regardless of the mapping approach , we conclude that L haplogroup individuals displayed reduced levels of mtDNA gene expression . For the sake of simplicity further analyses were performed using the personalized mapped samples . To control for possible bias underlying the trend towards lower levels of L-haplogroup mtDNA transcript expression , we considered the expression patterns of nDNA-encoded genes in Africans versus non-Africans . We found 2 , 380 nDNA-encoded genes that are differentially expressed in Africans ( S3 Table ) , yet unlike the mtDNA genes ~54% showed higher expression , while the rest showed lower expression in the African group ( S5 Fig ) . These findings suggest a lack of bias in the expression pattern of mtDNA-encoded transcripts . To control for possible group assignment bias , we randomly re-divided the samples 500 times , while retaining constant proportions of gender and ethnicities . Following group assignment , we repeated the gene expression normalization process and SNP association analysis . Our results revealed that in more than 60% of the replicated divisions , ten mtDNA-encoded genes ( MT-TH , MT-TI , M-TL2 , MT-CO2 , MT-ND2 , MT-ND6 , MT-CO1 , MT-ATP6 , MT-ND3 and MT-ND1 ) consistently showed significantly reduced expression levels in L-haplogroup samples ( Fig 3 ) . These results confirm that African L-haplogroup individuals possess a distinct mtDNA gene expression pattern . Since mtDNA transcription and replication are coupled in human mitochondria [41] , we included mtDNA copy number as one of the covariates in all our eQTL analyses . Nevertheless , we tested whether the differences in expression levels associated with variations in mtDNA copy numbers . We found that variations in mtDNA copy numbers did not differ between L- and non L-haplogroup mtDNAs ( Fig 4 ) . This suggests that the variation we observed in mtDNA gene expression patterns was independent of mtDNA copy numbers , a finding in agreement with previous results [42] . We reasoned that the highly significant gene expression differences between Africans and Caucasians may mask intra-population expression variation . To address this possibility we repeated the gene expression analysis separately for Caucasians and Africans . Although this analysis did not reveal any significant intra-population differences while considering the long RNA dataset ( S4 Table , S5 Table ) , our results indicate that in Africans , individuals belonging to haplogroup L3b had significantly higher expression of cysteine tRNA ( Fig 5A and S6 Table ) . While analyzing the Caucasian samples ( Fig 5B–5F and S7 Table ) , we found that tRNA Leucine ( 2 ) had higher expression in individuals belonging to haplogroup U5 . Secondly , higher expression of tRNA arginine was found in individuals belonging to haplogroup T . Finally , tRNA glycine had higher expression level in individuals sharing SNPs that define haplogroup cluster WI , in individuals harboring a guanine as compared to those having an adenine allele in mtDNA position 10 , 398 ( shared by haplogroups J , K and I ) , and in individuals with either an adenine or a cytosine in mtDNA position 16 , 129 as compared to those with a guanine in this position . Hence , our intra-population analysis revealed much significant variation in mtDNA gene expression that was previously masked by the more prominent differential expression between Africans and Caucasians . Such differences may stem , at least in part , from variation in the impact of certain alleles on gene expression , depending on their linked haplotypes ( Fig 6 ) . This is best exemplified by the relatively high expression of tRNA glycine in Caucasian haplogroup cluster WI individuals ( with the 12 , 705T allele ) as compared to individuals with the 12 , 705C allele ( see also Fig 5D ) ; all Africans harbor the 12 , 705T allele , which exhibits even lower tRNA glycine expression than the Caucasian 12 , 705C allele . The latter caused lack of significance while calculating the impact of 12 , 705 SNPs on gene expression considering Africans and Caucasians together ( Fig 6B ) . Taken together , the impact of mtDNA SNPs on gene expression differences is modified , at least in part , by their linked genetic background . Since the regulation of all mitochondrial activities is governed by nDNA-encoded factors , we asked which nDNA-encoded genes are the best candidates to modulate the most prominent distinct mtDNA gene expression pattern–that of the L-haplogroup . As a first step in addressing this question , we screened for nDNA-encoded genes that were co-expressed with the mtDNA genes ( Pearson correlation ) . These genes were then subjected to Matrix eQTL [43] analysis to identify differential expression between mtDNA genetic backgrounds ( i . e . , L or non-L haplogroups ) , while including gender , mtDNA copy number , and sample resource lab as covariates . After further correction for multiple testing ( Bonferroni correction ) , a list of 2 , 380 genes remained ( S3 Table ) . GO analysis revealed enrichment in RNA-binding proteins and poly-adenylated RNA-binding proteins ( Table 1 , S8 Table ) , suggesting that RNA stability likely plays a significant role in the differential expression of mitochondrial genes in L-haplogroup individuals . To further explore candidate genes that best explain the distinct L-haplogroup mtDNA expression pattern , we screened for nDNA SNP association . To avoid statistical power issues , we focused our SNP association study on the most comprehensive list of human RNA-binding and transcription-associated mitochondrial genes available [16] , supplemented by additional transcription factors that were recently localized to the human mitochondria and bind the mtDNA [18 , 19] . This analysis showed 7 , 665 correlations between a total of 511 non-redundant nDNA SNPs and 15 mtDNA-encoded genes , of which only five SNPs in four nDNA-encoded genes remained significant after correction for multiple testing in both test groups . Notably , these SNPs correlated with the expression of four mtDNA-encoded genes , namely MT-ND2 , MT-CO2 , MT-CO1 and MT-ND6 ( Fig 7A , Table 2 , S9 Table ) . The indicated SNPs map within the nDNA-encoded genes PTCD1 , MRPS7 , PNPT1 , HEMK1 and GRSF1 . Three of these nDNA genes have already been analyzed for their involvement in mitochondrial gene expression . PNPT1 is involved in RNA import into the mitochondria and is a part of the mitochondrial RNA degradosome [44 , 45] . PTCD1 takes part in processing primary mtDNA polycistronic transcripts [46] and GRSF1 is involved in processing both the full and tRNA-less polycistrons [47] . The fourth gene , MRPS7 , is a mitochondrial ribosomal protein , and therefore takes part in the mitochondrial translation machinery [48] and HEMK1 methylates the mitochondrial translation release factor HMRF1L [49] . In conducting the same analysis on the tRNA dataset , two of the identified SNPs correlated with the differential expression of the L-haplogroup in both groups of RNA-seq samples , namely PTCD1 and MRPS7 ( Fig 7B , Table 3 , S10 Table ) . In summary , our eQTL association analysis of reduced mtDNA gene expression in L-haplogroup individuals revealed consistent association of SNPs within PTCD1 and MRPS7 , both in the large and the tRNA datasets , suggesting a common mechanism involved .
In this study , while examining RNA-seq data from 454 individuals of African and Caucasian origin from the 1000 Genomes Project , we found distinct patterns of transcript abundance in individuals with certain mtDNA SNPs , corresponding to the mtDNA L-haplogroup ( the most abundant African haplogroup ) . L-haplogroup individuals presented an overall lower expression of certain mtDNA genes ( both mRNAs and tRNAs ) , as compared to individuals corresponding to non-L-haplogroups . Such distinct mtDNA gene expression patterns of the L-haplogroup suggest an ancient regulatory trait that likely existed prior to out-of-Africa migrations . In the future , when large Asian RNA-seq data become available , it will be possible to corroborate this interpretation . Our mtDNA eQTL analysis , which revealed a significant expression pattern difference between Africans and Caucasians , was based on SNP-expression pattern association , and was not based on prior division into populations . Furthermore , while performing intra-population eQTL analysis we found distinct mtDNA gene expression pattern for specific haplogroups , only while considering the tRNA genes . Finally , we noticed that the expression of tRNA glycine was elevated in individuals belonging to haplogroups W and I , as well as in individuals with a guanine allele in mtDNA position 10 , 398 and in individuals with either an adenine or a cytosine in position 16 , 129 ( which is found in people belonging to multiple haplogroups ) . Interestingly , all haplogroup I individuals harbor a 10 , 398G allele , suggesting that haplogroup I SNPs play a major role in determining differential expression of tRNA glycine . Alternatively , since the SNPs in positions 10 , 398 and 16 , 129 occurred multiple independent times during the evolution of man , it is possible that these SNPs have direct functional impact on the expression of tRNA glycine . We interpret all these results to mean that the intra-population expression differences might be context-dependent , and therefore , they were masked by including both Africans and Caucasians in the same analysis . Specifically , masking of such expression differences may occur , at least in part , due to differential impact of certain SNPs on expression , depending on their linked genetic backgrounds ( Fig 6 ) . Such differential impact could further be explained by: ( A ) the eQTL has a small functional impact which is enhanced by additional eQTLs , or ( B ) the causal eQTL is in linkage with the identified eQTL . However , currently we cannot differentiate which of the suggested explanations is the most plausible . Taken together , our eQTL analysis was not confounded by populations , and therefore revealed candidate mtDNA-encoded eQTLs . A recent study of mitochondrial activity in six cell lines sharing the same nDNA but diverging in their mtDNAs ( i . e . , cybrids ) , revealed differences in activity and transcript abundance among three L-haplogroup and three H-haplogroup cybrids [23] . Similarly , Gomez-Duran and colleagues identified expression pattern differences between haplogroup H cybrids when compared with those of the haplogroup Uk , 5 cell lines each [22] . Since we studied a much larger sample size from highly diverse individuals , we argue that our study better represents the natural population rather than focusing on specific haplogroups . This further underlines the future need to expand our study to include Asians so as to shed further light on mitochondrial regulatory differences from a world-wide perspective . Once cybrid technology has been adapted for high throughput analysis , it would be of interest to apply our genomic analysis to a large collection of cybrids with diverse mitochondrial genomes . Since the distinct L-haplogroup mtDNA expression pattern was shared between tRNAs and long RNAs that are encoded by both mtDNA strands , it is plausible that the observed differences stem either from early stage transcription or from polycistron stability . Alternatively , since expression pattern differences were limited to certain mtDNA-encoded genes , the underlying mechanism could involve differences in the RNA stability of the mature transcripts or during transcript maturation , as previously suggested [50] . With this in mind , both analysis of co-expressed nDNA-encoded genes and our eQTL association study revealed that RNA-binding proteins with mitochondrial function ( i . e . , PTCD1 and MRPS7 ) best explain the distinct mtDNA gene expression patterns of L-haplogroup individuals . Although a lack of association with SNPs in the vicinity of known mtDNA transcription regulators was observed , one cannot exclude future detection of such association when more mtDNA transcription regulators are identified . In summary , the distinct mtDNA transcript expression pattern observed in African individuals supports an ancient mitochondrial phenotype , as humans left Africa to populate the rest of the world . We identified several candidate nDNA-encoded modulators of this expression pattern , although their direct functional impact remains to be studied . Nevertheless , expression differences were only seen in certain mtDNA genes , despite the fact that all mtDNA genes are co-transcribed in two polycistrons corresponding to the light and heavy strands . This finding , along with the observed association of SNPs in mitochondrial RNA-binding genes , suggests that RNA decay is the best candidate mechanism for modulating the observed expression pattern .
RNA-seq data from lymphoblastoid cell lines ( LCL ) were obtained from several inter-population studies [26 , 28–31] . Datasets were downloaded from: citation 28 - www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE19480; citation 29 - www . ebi . ac . uk/arrayexpress/experiments/E-MTAB-197/; citation 26 - www . ebi . ac . uk/ena/data/view/ERR188021-ERR188482 and www . ebi . ac . uk/ena/data/view/ERR187488-ERR187939; citation 30 - www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE54308; citation 31 - www . ncbi . nlm . nih . gov/sra/ ? term=SRP026597 . However , since transcript abundance assessment ( see below and in the Results section ) revealed mtDNA gene expression clusters according to these studies , our analysis focused on data in the largest recently published study so as to avoid analysis artifacts [26] . In that study , RNA-seq libraries were constructed and data were generated in seven different labs that followed the same experimental protocol , relying on randomly distributed samples . This dataset included results from sequenced mRNA and rRNA ( ‘long RNA’ ) libraries , as well as small RNA libraries . The largest RNA dataset contained samples from 462 individuals from five world-wide populations ( 91 CEU , 95 FIN , 94 GBR , 93 TSI and 89 YRI ) . Eight of the samples were not successfully mapped to the human genome due to too much unpaired reads and were , therefore , excluded , thus leaving 454 samples for further analysis . Small RNA sequences generated from a subset of these samples ( N = 452 ) encompassed most of the individuals described above ( 87 CEU , 93 FIN , 94 GBR , 89 TSI and 89 YRI ) . Three of the samples were not included in the long RNA dataset and were , therefore , excluded from further analysis , in addition to eight samples whose RNA sequences could not be mapped to the long RNA dataset . The same was true for an additional sample that was not successfully mapped to the dataset of small RNA libraries ( tRNAs in the mtDNA ) . This left 440 samples for further analysis . Data were downloaded from Arrayexpress ( E-GEUV-1 and E-GEUV-2 for long RNA and tRNA , respectively ) . The long RNA dataset was downloaded from: http://www . ebi . ac . uk/ena/data/view/ERR188021-ERR188482 The small RNA sequence libraries were downloaded from: http://www . ebi . ac . uk/ena/data/view/ERR187488-ERR187939 Taking advantage of the polycistronic transcription of the mitochondrial genome [51] , mtDNA genomes were reconstructed from each of the RNA-seq samples using MitoBamAnnotator [52] . Each reconstructed mtDNA sequence underwent haplogroup assignment using HaploGrep [53] . Finally , multiple sequence alignment and phylogenetic analysis ( neighbor joining , 1000 X bootstrap and default parameters ) of all reconstructed mtDNA genomes were performed using MEGA 5 [54] . RNA-seq reads of the long RNA dataset were mapped onto the entire human genome reference sequence ( GRCh 37 . 75 ) using STAR v2 . 3 [55] and the available genome files at the STAR website ( code . google . com/p/rna-star/ ) . Mapping was performed using default parameters , in addition to the [—outFilterMultimapNmax 1] parameter to achieve unique mapping . Since mtDNA sequence variability can impact the number of mapped RNA-seq reads , the reads were remapped against the same human genome files after replacing the mitochondrial reference sequence by the reconstructed mtDNA of each of the relevant analyzed individuals . To this end , a revised index was generated for the new reference genome by replacing the human mtDNA sequence with the reconstructed version . This was conducted separately for each tested sample , with all other files being retained . Most of the parameters used were retained , with one exception . Apart from replacing the mtDNA reference , we further increased our mapping accuracy by allowing fewer mismatches [—outFilterMismatchNmax 8] , while analyzing couples of paired reads ( a total length of 150bp ) . The tRNA dataset was mapped using the same parameters and references as in the remapping process described above , with the single exception of no mismatches allowed [—outFilterMismatchNmax 0] so as to reduce mapping errors [56] . Alignment files ( SAM format ) were compressed to their binary form ( BAM format ) using Samtools [57] with the default parameter [view -hSb] selected , and sorted using the [sort] parameter . Mapped reads were counted using HTSeq-count v0 . 6 . 1 . p1 [58] , employing the [-f bam -r pos -s no] parameters . Reads were normalized to library size using DESeq v1 . 14 . 0 [59] and the default parameters . This protocol was employed for both the long RNA and tRNA datasets . mtDNA sequences of all individuals were aligned to identify polymorphic positions . In the tRNA dataset , some tRNA genes had no reads in a subset of our analyzed samples . Therefore , only genes presenting with reads in more than 90% of the samples were used , thus leaving 16 tRNA genes for further analysis . For each polymorphic position , the samples were divided into groups according to their allele assignment . As described in Lappalainen [26] et al . , using the linear model implemented in the Matrix eQTL R package [43] , eQTL mapping was calculated according to the allele assignment , while considering gender , mtDNA copy number and sample resource ( i . e . lab of origin ) as covariates . A Bonferroni correction was employed to correct for multiple testing . To reduce false positive discovery rate we focused on SNPs shared by at least 10 individuals . To identify possible associations of nDNA-encoded genes with differential expression patterns of mtDNA genes , the analysis focused on known SNPs ( as listed by the 1000 Genomes Project ) in the dataset of genes with known mitochondrial RNA-binding activity [16] . This dataset was supplemented by transcription factors and RNA-binding proteins that were recently identified in human mitochondria but were not included in MitoCarta ( i . e . , c-Jun , JunD , CEBPb , Mef2D ) [18 , 19] . Such prioritization was employed to enable detection of possible correlations with sufficient statistical power . Since our analysis of the mtDNA revealed sites with more than 2 alleles ( i . e . 3 alleles at the most ) we performed our analysis such that the major allele frequency will not exceed 95% , thus enabling the discovery of the two other minor alleles . To assess co-expression of nDNA-encoded genes with the mtDNA-encoded ones , Pearson correlation was employed ( p < 5 . 23e-8 , after Bonferroni-correction for multiple testing ) . Briefly , co-expression was sought between the 15 mtDNA-encoded mRNA and rRNA genes and 63 , 662 nDNA-encoded genes . Then , Matrix eQTL [43] was employed to identify differential expression between L and non-L genetic backgrounds among the significantly co-expressed genes , while including gender , mtDNA copy number , and sample resource lab as covariates . Finally , only nDNA-encoded genes that were significantly co-expressed , and were differently expressed between African and Caucasians ( p < 5 . 5e-6 , Bonferroni corrected for multiple testing ) , were subjected to gene ontology ( GO ) analysis . To this end , PANTHER [60] was used , categorized according to molecular function . As the cell lines that were used in the analyzed RNA-seq study were derived from individuals included in the 1000 Genomes Project [61] , mapped DNA reads files were downloaded from the 1000 Genomes Project ftp website in BAM format ( ftp://ftp . 1000genomes . ebi . ac . uk/vol1/ftp/data ) , using Samtools with the [view -hb] parameter . Bedtools was employed to estimate the read coverage of nDNA regions in each individual , using the BAM and global BED files of that individual . Using these files , read coverage over most of the mtDNA sequence ( mtDNA positions 1–16 , 499 ) was compared to that of randomly selected sets of 100 , 000 bases from each autosomal chromosome ( nucleotide coordinates 20 , 100 , 000–20 , 200 , 000 in each of the 22 autosomes ) . Specifically , the above-mentioned read coverage values were used to calculate the ratio between mtDNA and nDNA read coverage . This ratio equals the estimated mtDNA copy numbers . | The mitochondrion is an organelle found in all cells of our body and plays a significant role in the energy and heat production . This is the only organelle in animal cells harboring its own genome outside of the nucleus . Mitochondrial DNA ( mtDNA ) variants have been traditionally used as neutral markers to trace ancient population migrations . As a result , the functional impact of human mtDNA population variants on gene regulation is poorly understood . To address this question , we analyzed available data of mtDNA gene expression pattern in a large group of individuals ( 454 ) from diverse human populations . Here , we show for the first time that the ancient migration of humans out of Africa correlated with differences in mitochondrial gene expression patterns , and could be explained by the activity of certain RNA-binding proteins . These findings suggest a major mitochondrial regulatory transition , as humans left Africa to populate the rest of the world . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"sequencing",
"techniques",
"transfer",
"rna",
"rna-binding",
"proteins",
"mitochondrial",
"dna",
"gene",
"regulation",
"population",
"genetics",
"non-coding",
"rna",
"forms",
"of",
"dna",
"mitochondria",
"dna",
"bioenergetics",
"population",
"biology",
"molecular",
"biology",
"techniques",
"rna",
"sequencing",
"cellular",
"structures",
"and",
"organelles",
"haplogroups",
"research",
"and",
"analysis",
"methods",
"proteins",
"gene",
"expression",
"molecular",
"biology",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"energy-producing",
"organelles",
"evolutionary",
"biology"
]
| 2016 | Ancient Out-of-Africa Mitochondrial DNA Variants Associate with Distinct Mitochondrial Gene Expression Patterns |
Transcription factors ( TFs ) regulate gene expression through specific interactions with short promoter elements . The same regulatory protein may recognize a variety of related sequences . Moreover , once they are detected it is hard to predict whether highly similar sequence motifs will be recognized by the same TF and regulate similar gene expression patterns , or serve as binding sites for distinct regulatory factors . We developed computational measures to assess the functional implications of variations on regulatory motifs and to compare the functions of related sites . We have developed computational means for estimating the functional outcome of substituting a single position within a binding site and applied them to a collection of putative regulatory motifs . We predict the effects of nucleotide variations within motifs on gene expression patterns . In cases where such predictions could be compared to suitable published experimental evidence , we found very good agreement . We further accumulated statistics from multiple substitutions across various binding sites in an attempt to deduce general properties that characterize nucleotide substitutions that are more likely to alter expression . We found that substitutions involving Adenine are more likely to retain the expression pattern and that substitutions involving Guanine are more likely to alter expression compared to the rest of the substitutions . Our results should facilitate the prediction of the expression outcomes of binding site variations . One typical important implication is expected to be the ability to predict the phenotypic effect of variation in regulatory motifs in promoters .
The regulation of gene expression is mediated mainly through specific interactions of TF proteins with DNA promoter elements . TF binding sites ( TFBS ) are short ( typically of length 6–20 bases ) and imprecise; unlike restriction enzymes which recognize unique nucleotide sequences , a single TF protein may interact with a range of related sequences . For most TFs , there appears to be no distinct sequence of nucleotide bases that is shared by all recognized binding sites . However , there are typically clear biases in the distribution of bases that occur at each binding site position . These biases are commonly represented by position weight matrices ( PWMs ) , whose components give the probabilities of finding each nucleotide at each binding site position [1] . Given the degenerate nature of genuine binding sites , highly similar sites within the same genome may be recognized by the same TF or by distinct TFs . This is also true for the genomes of related species , where slight changes in binding site sequence , occurring throughout evolution , may in some cases maintain the specificity of the site and in others lead to its loss or to the creation of a site targeted by a different TF [2] , [3] . The desire to distinguish between ‘neutral’ binding site variations , which do not change the recognition range of the site , and ‘functional’ variations , which may affect gene expression by altering protein-DNA interactions , lays at the heart of this work . Such a distinction may have several implications . Firstly , it should greatly improve the performance of scanning algorithms , which search promoter sequences for matches to predefined PWMs . These algorithms typically regard all mismatches between a promoter sequence and a given PWM's preferences as equal ( c . f . ScanACE [4] , Match™ [5] , MAST [6] ) . More reliable predictions may be obtained if such mismatches are differentially weighed based on their expected effects on expression . Identification of genuine sites is also crucial when comparing the promoters of orthologous genes - some across-species variations may change the functionality of a motif in some of the organisms . Another intriguing application is the detection of regulatory site variations , which have the potential to reduce fitness , and cause diseases , through altering gene expression . Disease-causing binding site variations are known to occur [7] , [8] , however so far no attempts have been made for their prediction on a genome wide scale . Most efforts to distinguish disease-causing variations from neutral ones have focused on coding single nucleotide polymorphisms ( SNPs ) [9]–[15] . Estimates show that the human population contains thousands of cis-regulatory variations [16] . Such high numbers justify a dedicated effort for the development of computational means for predicting deleterious regulatory variations . The present work lays the foundations for the development of such methods , experimented here in yeast , introducing measures for quantifying the effects of binding site variations on gene expression . We have first constructed a putative binding site motif collection using our previously introduced expression coherence ( EC ) score , which quantifies a motif's regulatory effect by measuring the extent to which genes that contain it display similar expression at a given biological condition [17]–[19] . We next exploit this quantitative measures of a motif's regulatory effect , in order to systematically compare the expression patterns of genes containing binding sites differing by a single nucleotide position . We accumulate statistics for many substitutions across multiple putative binding sites , and observe that not all nucleotide substitutions are similar in severity: We found that substitutions of the type A->N , and substitutions of the type N->A ( where N is any nucleotide but A ) tend to be “benign” , i . e . they relatively rarely change the expression patterns of the regulated genes . On the other hand , substitutions of the type G->N , and N->G are more likely than other substitutions to lead to binding site loss .
A single base mutation in the binding site of a TF can result in one of three scenarios: ( i ) Binding site conservation - the mutant is also recognized by the same TF and the substitution from wild-type to mutant is thus expected to have a very mild effect ( Figure 1 , green arrows ) . ( ii ) Binding site switching - the mutant is no longer recognized by the original TF , but is recognized by an alternative TF ( Figure 1 , blue arrows ) . ( iii ) Binding site loss: the mutant binding site is no longer recognized by any TF ( Figure 1 , red arrows ) . The primary goal of this work was to computationally distinguish these three scenarios . As a first step towards establishing a general scheme for the assessment of the effect of single-base substitutions on the binding sites of transcription factors , we examined computationally the effects of such substitutions on the consensus site of the yeast sporulation factor Ndt80 , the primary transcriptional activator of middle sporulation genes . Using our motif landscape analysis tool [17] we analyzed the effects of all single-base substitutions on Ndt80 binding ( Figure 1 , right panel ) . For each ‘mutant’ motif , this tool answers two questions: can it potentially constitute a binding site for a TF ? and more specifically , is it likely to bind the same TF as the ‘wild-type’ motif ? To answer the first question , the tool employs the previously described Expression Coherence ( EC ) score [17]–[19] , which assesses the effect of a promoter sequence motif on the corresponding genes' expression profiles . The EC score measures the extent to which a set of genes ( in this case the set is defined by a common sequence motif in the genes' promoters ) displays similar expression profiles at a given biological condition . The statistical significance ( p-value ) of an EC score is the estimated probability of obtaining the observed or higher EC score by chance and is dependent on the size of the set of genes considered [17] . The answer to the second question is obtained by computing the Pearson correlation distance between the mean expression profile , under a relevant condition , of the set of genes whose promoters contain the ‘wild-type’ motif and that of the set of genes whose promoters contain the ‘mutant’ motif . Using the S . cerevisiae sporulation expression data [20] , our landscape analysis predicted that two out of the three possible substitutions in the second position will have only a minor effect on expression whereas an A->G substitution at the same position will have a harsher effect ( see Figure 1 legend for details ) . When averaging over all possible single nucleotide substitutions ( Figure 2 ) , the second position seems to be the most tolerant towards substitutions ( mean expression distance , i . e . 1-Pearson correlation coefficient , is 0 . 0358 ) , and the seventh position – the most sensitive ( mean expression distance 0 . 7715 ) . One possible reason for such a marked difference between the tolerance of different positions within the same motif to substitutions may be that the binding transcription factor forms different contacts with the DNA at each of the positions . Particularly , we may expect the positions that form tight contact to be less permissive to substitutions . Indeed , our results are in good agreement with the structural data of Ndt80 bound to its DNA target [21]: the second ‘permissive’ motif position is the only position which does not form a direct contact with the protein . But do these differences affect TF function ? Reassuringly , these results are also supported by published in vivo reporter expression experiments and in vitro binding assays of Ndt80 mutants [22] . This experiment represents the ‘wet’ analog to our computational experiment – each of the nucleotide positions in the Ndt80 consensus site was replaced with all possible three alternatives . These results too showed that the second position is most permissive to substitutions , and that , as predicted by us , G is the only nucleotide that when placed at this position weakens binding affinity and reduces expression level of the reporter gene [22] . This implies that the computational measures used here can complement and predict the outcome of ‘wet’ mutation experiments . The use of Ndt80 as a test case also provided us with the opportunity to examine whether we can computationally distinguish between binding site switching and binding site loss . Ndt80 recognizes variations of a site termed middle sporulation element ( MSE ) , whose consensus sequence is GNCRCAAW . Interestingly , variations of the MSE are also recognized by Sum1 , a transcriptional repressor of middle sporulation genes during mitosis and early sporulation . Through a combination of in vivo reporter expression assays and in vitro binding assays of Ndt80 and Sum1 mutants , Pierce et al . defined the specific binding preferences of these two TFs [22] . They found that while positions 3–5 of the MSE are important for binding of both Ndt80 and Sum1 , there is a difference in binding preferences at positions 6–7 . For these positions , Ndt80 requires strictly an A , whereas Sum1 binds equally to an A and to a T . Indeed , our landscape analysis ( Figure 1 , right panel ) shows that mutating position 6 from A to T results in a change in expression profile , yet coherence remains significant ( p-value 0 . 0083 ) . This may be explained by the binding site switching from Ndt80 to Sum1 . Transitions of the same position into C or G result in binding site loss ( p-values of the EC score of the substituted motifs are 0 . 4 and 0 . 3 , respectively ) . The same applies for position 7 , in which transition from A to T maintains a relatively significant EC score ( p-value 0 . 019 ) , yet the expression profile is changed relative to the genes containing the consensus motif . On the other hand , substitutions at this position to both C and G lead to complete loss of coherence ( EC p-values are 0 . 4 for both variants ) . This position also scored as the most sensitive to mutations – any change will abolish the Ndt80 site , either by switching or complete loss . Encouraged by our ability to predict the effects of binding site substitutions within a single motif , we attempted to generalize these predictions in order to define universal properties of substitutions that alter gene expression . Towards this end we compiled a dataset of motifs that are likely to participate in the regulation of gene expression . This study was conducted in the S . cerevisiae genome , for which vast TFBS knowledge is available . However , in order to both broaden this knowledge and form a quantifiable connection between binding site sequence and the expression profiles of the regulated genes , we compiled a new comprehensive motif dataset . This dataset is unbiased by prior knowledge and is based on the premise that any nucleotide sequence that resides in the promoter of a gene may contribute to the regulation of the gene's expression . A further advantage of our dataset is that it is not limited to TF binding sites – motifs found by our methodology are relevant to transcription , but not all are necessarily TF binding sites . Some , for example , may be involved in DNA bending . We constructed our motif dataset by integrating whole genome promoter sequences of S . cerevisiae with expression patterns of the corresponding genes in 40 natural and perturbed biological conditions including cell cycle , sporulation and various stress responses . Each biological condition was represented by a time series of microarrays ( see methods ) . To obtain the most comprehensive dataset , we systematically scanned all k-mers ( k ranges from 7 to 11 ) that appear in S . cerevisiae promoters . For each k-mer and each of the 40 biological conditions , we computed the EC score of the set of genes that contain it in their promoters . A p-value was assigned to each EC score and a false discovery rate ( FDR ) [23] of 0 . 1 was applied to correct for multiple hypotheses ( see methods ) . The EC score was used not only to assess the biological significance of the scanned k-mers , but also to assign them with likely regulatory functions , in the form of the set of biological conditions in which they operate , and the regulatory effects they exert in each condition ( e . g . increased expression following stress , or peak in expression level at a particular cell cycle stage ) . A total of 8 , 610 sequence motifs appeared significant in at least one of the examined biological conditions . These comprise the ‘core’ of our dataset ( hereafter referred to as the ‘core dataset’ ) . This dataset represents potential cis-regulatory elements , whereas the rest of the scanned k-mers represent a control set of presumably non-functional elements . A list of our core motif sequences , along with their EC scores and p-values in the biological condition in which each motif obtained the most significant score is provided in the supporting information ( supporting Table S1 ) . We performed several analyses to validate the ability of our method to identify biologically significant motifs ( see supporting Text S1 for complete details ) . First , we compared the core dataset to the well accepted reference collection of yeast TFBS published by Harbison et al . [24] ( supporting Text S1 and Figure S1 ) . Briefly , the Harbison set was obtained by experimentally determining the genomic occupancy of DNA-binding transcriptional regulators under rich medium as well as other growth conditions . Using motif-discovery algorithms the information from genome-wide location data was then combined with phylogenetically conserved sequences and prior knowledge to derive for each regulator its probable specificities [24] . Each motif in the Harbison set is represented by a positional weight matrix ( PWM ) , which specifies for each position in the regulator binding site and each of the four possible nucleotides , the likelihood of observing the specific nucleotide at that particular position . Our dataset covers 99 out of the 102 PWMs published by Harbison et al . We also found a tendency for genes , which contain the same motif from our core dataset in their promoters , to be associated with similar functions ( supporting Text S1 ) . We next used our core dataset to investigate characteristics that may be of relevance to motifs' biological function . For this purpose we compiled a control set of 190 , 211 low scoring k-mers , that were insignificant in all 40 examined biological conditions , and in addition scored especially low ( p-value > 0 . 8 , gene set size > 8 ) in at least one of these conditions . We considered various features that may be important for the function of a regulatory motif and for each such feature , defined a quantitative measure , and tested whether it can significantly differentiate between our highly scoring motifs and the control set . Compared to the control set our significant motifs were found to have high GC content ( relative to the yeast AT rich genomic background ( supporting Figure S2 ) , to have high entropy ( supporting Figure S3 ) , to appear in higher copy numbers ( supporting Figure S4 ) and to display a preference to distinct positions relative to the transcriptional start site ( TSS ) in different promoters ( positional bias ) ( supporting Figure S5 ) . Additionally , our motifs were found to be evolutionarily conserved in the promoters of four closely related Saccharomyces species ( Figure 3 and supporting Figure S6 ) . Reassuringly , some of these properties are known to characterize functional binding sites ( positional bias [19] , [25] , multiplicity of sites [17] , [26] , evolutionary conservation [27] , [28] ) . The analysis of evolutionary conservation revealed , that while the core dataset is significantly more conserved than a randomized motif set , there is a substantial number of motifs in the core dataset that are not significantly conserved , as manifested by an overlap in the two distributions of conservation rates ( Figure 3 ) . This implies that the core dataset likely contains many false positives . Therefore , in order to obtain a distilled signal in the analyses that follow we applied a conservation-based filter to the core dataset . Using the 95th percentile of the conservation rate ( see methods ) of a control motif set as a threshold , we filtered out those core dataset motifs that had a conservation rate below this threshold . The new dataset , referred to as the filtered core dataset , constitutes 1 , 036 motifs ( see supporting Table S1 for the identity of the motifs that were included in this set ) . We next exploited the filtered core dataset in order to study the functional outcomes of binding site variations . For each motif in this dataset we exhaustively enumerated all possible single-base substitutions and examined their effect on expression patterns , in view of the three previously described scenarios ( Figure 1 , right panel ) . The process of assessment of the effect of a substitution is best described in terms of a decision tree . Figure 4 summarizes the distribution of substitutions along the leaves of this decision tree . The first decision addresses the possibility of binding site loss . If the substitution results in a k-mer sequence that is outside of the filtered core dataset we consider the substitution to result in a binding site loss , since the genes that contain that version of the motif are not significantly coherent , whereas the genes that contain the original motif from the filtered core are coherent . Next , we attempt to discover clear cases of binding site switches . These are defined as cases where the original motif and the substituted sequence are both members of the filtered core dataset , but the sets of biological conditions in which they are inferred to exert their effect are non-overlapping . Finally , substitutions that survive the two filters can constitute either a conservation of the binding site ( i . e . benign mutations ) or a binding site switch , depending on their effect on the expression profile in the biological conditions in which both the motif and its substituted variant are coherent . Interestingly , examination of the distribution of correlation coefficients between the mean expression profiles of the motif and its substituted variant for those substitutions that fell into the last category revealed that these correlation coefficients tend to be high and statistically significant ( Figure 5 ) . Thus , it seems that cases similar to the one observed for Ndt80 and SUM1 , in which a substitution results in a switch to a binding site that is active in the same biological condition as the original binding site , are quite rare . Table 1 describes the few cases found to constitute such a switch . Consequently , we define the substitutions that fall into the last category as benign substitutions that will have very little effect on expression . Overall from the decision tree it is apparent that the majority ( ∼97 . 5% ) of the single nucleotide substitutions result in a loss of a functional binding site . It should be emphasized , though , that the our criteria for inclusion in the high confidence core dataset were very strict ( both in terms of expression coherence and evolutionary conservation ) , and that more loose criteria could have resulted in a lower proportion of binding site loss events . Using the classification of substitutions obtained for the filtered core dataset we next attempted to generalize these predictions in order to define properties of substitutions that alter gene expression . Our goal was to define substitution types that are more radical than others ( in analogy to amino acid substitutions where there are conservative changes that maintain the chemical properties of the residue versus radical changes that result in a residue with different characteristics ) . For this purpose we used chi-square tests to compare the distribution of the different substitution types among the cases of binding site loss with that found for cases with benign effects on expression ( the last category in the decision tree which corresponds to binding site conservation or switch ) . The results are summarized in Tables 2–4 . We found that substitutions from A to T ( or from T to A ) are the most underrepresented among the cases of binding site loss , implying that such substitutions tend to be benign ( Table 2 ) . On the other hand , substitutions from G to T ( or from T to G ) seem to be very radical as they are the most overrepresented among cases of binding site loss ( Table 2 ) . In general it seems that substitutions that involve a G ( either as the source nucleotide or as a target nucleotide , i . e . N->G , or G->N , where N is any nucleotide , but G ) tend to lead to a binding site loss , whereas substitutions that involve an A ( as a source or as a target ) have a tendency to be benign ( see Tables 3 and 4 for source and target statistics , respectively ) . Also , the identity of the source nucleotide has less of an impact on the outcome than the identity of the target nucleotide , as attested by the higher chi-square scores in Table 4 compared to Table 3 . Finally , by taking into account the chi-square deviation observed for each of the twelve possible substitution types , as well as the information whether it is found more or less than expected among the cases of binding site loss , we can rank the substitution types from most benign to most severe ( Table 2 ) . These statistics also allow us to rank in terms of severity the three possible substitutions from every source nucleotide . For instance , in a position containing a C as the source , the most severe substitution would be to a G , followed by T , and then by A . We further examined the cases where two motifs that differ by one nucleotide are coherent in a shared condition . In such cases the genes that contain one version of a motif and the genes that contain an alternative version of the same motif may display a similar , or dissimilar expression pattern ( in the same condition where they both show high coherence ) . We thus examined the corresponding expression profiles in such conditions and asked if they tend to be similar or not for the different substitution types . To explore this possibility we examined separately the distributions of correlation coefficients and corresponding p-values for the different substitution types ( supporting Figure S7 ) . Since we cannot distinguish which motif corresponds to the ‘wild-type’ and which corresponds to a ‘mutant’ in this case the distributions for complementary substitution types were pooled together . Visual examination of these distributions revealed that the distribution of p-values for substitutions between C and G tends to higher ( i . e . less significant ) values relative to the corresponding distributions for substitutions among other pairs of nucleotides ( median p-value for C⇔G substitutions is 1 . 08e-04 , whereas the highest median for the other pairs of nucleotides is 6 . 60e-05 ) . In other words , substitutions between G and C tend to generate more cases of dissimilar expression patterns compared to other substitutions . Using a rigorous comparison between the distributions corresponding to different pairs of nucleotides we found that no two distributions differ significantly from each other when correcting for multiple hypotheses ( supporting Tables S2 , S3 , and S4 ) . Nonetheless , the comparisons between the distribution of p-values corresponding to C⇔G substitutions to the other distributions yield much lower p-values than comparisons among the distributions of other pairs of nucleotides ( p-values range from 0 . 0336 to 0 . 1606 for comparisons involving the C⇔G substitutions , whereas for other comparisons the lowest p-value is 0 . 5382 ) . This may indicate that C⇔G substitutions have a greater tendency than other substitutions to weaken the affinity of the regulatory protein to the binding site .
By accumulating statistics for many substitutions across the motifs in our filtered motif dataset we observed that not all nucleotide substitutions are similar in severity: In the S . cerevisiae genome substitutions involving a G have a harsher effect on average than those involving an A . This observation may be perhaps explained by the fact that although both G and A participate in specific protein-DNA recognition through hydrogen bonds between the side chains of nucleotides and amino acids , the amino acids that preferentially bind G ( Lysine , Arginine , Serine and Histidine ) are much more prevalent in DNA-protein contacts than those that preferentially recognize A ( Asparagine and Glutamine ) [29] . Also , the observation that substitutions between A and T are under-abundant in cases of binding site loss may be explained by the fact that both these nucleotides participate in ring-stacking interactions with proline and phenylalanine in protein-DNA complexes [29] . It would be interesting to check whether the rules we have found apply to additional genomes . An intriguing follow-up on this study would be to test additional features that may affect the sensitivity of a binding site position to substitutions , such as its evolutionary conservation and its proximity to the protein in the DNA-protein complex . Many such features may be ultimately integrated in order to form a prioritization scheme that would allow the ranking of existing genome variations by their disease-causing potential . The approach presented here demonstrates for the first time how a huge amount of data , in this case the promoter sequences of all yeast genes as well as expression data for all genes across multiple conditions , can be harnessed and utilized for taking the first step towards assessing the effects of nucleotide substitutions on regulatory binding sites . A conceptual analogue of this endeavor for assessing the effects of amino acid substitutions on protein function could amount to mutating many proteins , say enzymes , in many different ways , and for each mutation measuring the reduction , or change , in biochemical activity and specificity . Since data for such an effort is not even close to becoming available , the methodology presented here utilizes in a unique way data that is available for its domain . While the main advantage of our methodology is the huge sample size , the disadvantage is that we are unable to control for other differences between promoters of analyzed genes ( i . e . differences that are outside of the substituted position ) . The fact that we obtain statistically significant differences between the effects of different types of substitutions on expression likely indicates that , despite uncontrolled sources of variation , we extracted genuine signals . An additional application of the present approach may be in algorithms that assign PWMs to promoters ( e . g . PRIMA [30] ) as it should provide means to differentially weigh mismatches between the PWM preferences and the promoter sequence , based on expected effect on expression . Particularly , at least in the S . cerevisiae genome , if the mismatch between the PWM and the sequence examined involves a G , the sequence is less likely to be a functional binding site than if the mismatch involves an A .
Promoter sequences for 5 , 651 S . cerevisiae genes were taken from the Saccharomyces Genome Database ( SGD ) [31] . Whole-genome mRNA expression data of 40 time series experiments in S . cerevisiae , were downloaded from ExpressDB [32] . These time series represent a wide range of natural ( e . g . cell cycle ) and perturbed conditions . This set of conditions was utilized by us before [33] and a complete list of conditions is available at http://longitude . weizmann . ac . il/TFLocation/conditions_explist . html . Yeast promoters were systematically scanned for all occurrences of every possible k-mer ( k varies from 7–11 ) , resulting in an index file listing for each k-mer the set of genes that contain it in their promoters , along with the positions and orientations ( strand ) of each occurrence . Bidirectional promoters were taken twice in different orientations and associated with the corresponding genes . Following the k-mer indexing step , EC scores ( and corresponding p-values ) in various experimental conditions were calculated for the sets of genes containing each of the k-mers in their promoters . The correction for multiple hypotheses was performed separately for each condition using a false discovery rate ( FDR [23] ) of 0 . 1 ( allowing 10% false positives ) . In addition to the EC scores and corresponding p-values , each k-mer was characterized by the expression profile it dictates; this was defined , at each time point as the average expression level of all genes assigned to the k-mer . Such averaged profiles were defined for each k-mer in each of the 40 time series experiments , resulting in 40 vectors per motif . A fundamental assumption made by our method was that a regulatory protein recognizes and binds double stranded DNA , and would therefore bind a motif instance equally whether it appears on the forward strand or the reverse strand of the promoter . Thus , in generating our motif dataset we considered a specific k-mer and its reverse complement as a single motif . As a consequence , pairs of genes with divergent promoters always ended up together in the sets of genes used to calculate the EC score of a motif . To address the question of what impact the above assumption had on the resulting dataset of motifs with statistically significant EC scores ( the ‘double-stranded dataset’ ) , we reapplied our method to the promoters of the S . cerevisiae this time counting forward instances and reverse instances of the same k-mer separately . We refer to the dataset of motifs with statistically significant EC scores generated in this way as the ‘single-stranded dataset’ . Supporting Table S5 gives a detailed comparison of the double-stranded and single-stranded datasets . While the two datasets are of similar size , the overlap between them is quite low: 4728/8610 motifs are unique to the double-stranded dataset ( both the forward and reverse instance of the motif is absent from the single-stranded dataset ) and 3767/8280 motifs are completely unique to the single-stranded dataset . This indicates that the answer to the question of whether TF binding sites are directional is not a straight-forward one . The high fraction of motifs unique to the double-stranded dataset implies that in many cases both forward and reverse instances of a motif generate the same expression pattern , and may not be detectable using the single stranded methodology because of lack of statistical power due to splitting of the regulated genes into two gene sets . On the other hand , the high fraction of motifs unique to the single-stranded dataset , as well as the high number of cases where only one direction of a double-stranded motif is included in the single-stranded dataset implies that in many cases the directionality of the motif is important to the recognition by the regulatory protein . For these motifs consideration of their reverse complement in the calculation of EC scores in the double-stranded methodology adds noise that in many cases cannot be overcome by the signal . Thus , the most comprehensive approach in the future would probably be to consider the union of the results obtained using both the double-stranded and the single-stranded approaches . For simplicity , we chose to limit the current study to those motifs generated by the double-stranded methodology . Future studies may consider also those motifs generated using the single-stranded methodology . Our method for compiling our dataset of potential regulatory motif is based on the detection of motifs whose presence in the promoters of a group of genes is associated with a similar expression pattern for the genes in question . A potential concern could be that paralogs resulting from very recent duplication events , and that have therefore not diverged sufficiently both in their promoters and expression patterns , would lead to false positives in our motif dataset . To address this concern we examined , using the data of Kafri et al . [34] , the relationship between time of duplication ( as assessed by Ks , the rate of synonymous substitutions within the coding sequence ) and the overlap in known motif content of pairs of paralogs ( supporting Figure S8 ) . We found that time of duplication is not a good predictor for the tendency of paralogs to share regulatory motifs . Therefore , it seems unlikely that the similarity in promoter sequences of paralogous genes would drive false discovery of potential regulatory motifs . The formal definition of the EC score is the fraction of gene pairs in a given set S , for which the Euclidean distance between normalized expression profiles falls bellow a threshold D . The threshold D is determined based on the distribution of pairwise distances between expression profiles of all genes in the genome ( or more precisely of all genes for which expression level was monitored ) . The original definition of the EC score [18] used the 5th percentile as the cutoff for defining “close” expression profiles ( D ) . This definition may create a bias towards TFs that exert a very tight regulation and miss regulatory motifs that correspond to factors exerting a more lose regulation . We therefore tested a range of EC definitions , with cutoffs corresponding to the 5th , 10th , 20th , 30th , 40th and 50th percentiles of the pair-wise distance distribution . For each definition of EC cutoff we assigned a significance p-value separately . P-values were calculated by random sampling . For each of the 40 expression time series and for each gene set size ( varying from 3–50 genes ) , we selected 100 , 000 random gene sets and computed an EC score for each such set at each cutoff definition . We define the p-value of a given EC score as the fraction of random sets ( of the same size and at the same condition ) that scored similarly or higher ( note that this sets a lower bound of 10−5 on the significance that can be assigned to a given EC score ) . Because we assume that for a given EC score , the probability to get the same score for random sets of genes drops with the set size , gene sets larger than 50 were assigned an upper bound approximated p-value , using the randomly sampled sets of size 50 . See supporting Figure S9 for the distribution of set sizes used for the different k-mers . Promoter data for four closely related Saccharomyces species S . cerevisiae , S . mikatae , S . kudriazevii and S . bayanus were taken from Cliften et al . [27] . Reference lists of motifs that were defined solely based on phylogenetic footprinting were taken from both Cliften et al . [27] and Kellis et al . [28] . The motif conservation calculation was adapted from Xie et al . [35] . Motif conservation was defined as the fraction of motif positions that are identical across all 4 species . We defined the motif conservation rate separately for each motif as the ratio of conserved motif instances to total occurrences of the motif in the genome . We regarded a motif instance as conserved if it displayed at least 90% conservation . Note that since the promoter alignments do not cover whole promoters for all genes the conservation rate doesn't factor in all occurrences of the motif in the S . cerevisiae genome , and in particular some of the motifs did not appear in any of the alignments and their conservation rate is thus undetermined . The distribution of conservation rates obtained for our core dataset of motifs was compared to a control distribution , which was obtained by calculating the conservation rate for randomized versions of the motifs ( in order to preserve GC content ) . We took the 95th percentile of the control set distribution as the cutoff defining high conservation . Supporting Protocol S1 displays the results obtained with different cutoffs for the definition of high conservation . It can be seen that using different cutoffs the general trends observed in Table 2 are preserved , albeit less strictly in the more permissive cutoffs . | A prime mode of control of transcription is the binding of transcription factors to promoter elements . These elements are often imprecise – often more than one , yet typically not all , of the nucleotides may be tolerated . In another field of protein structures , only some of the amino acid substitutions are tolerated . Are there parallels to this situation in transcription motifs ? For instance , could substitutions between Adenine and Guanine have relatively little effect on transcription ? An experimental approach here is daunting , requiring a library of reporter genes , each under the regulation of a different motif version . Yet genomes contain genes with natural variations on motifs , while micro-array data provide the expression pattern of those genes . We inspected the S . cerevisiae genome to derive nucleotide substitution severity “rules” . We compared the expression of genes containing motifs to the expression of genes that contain motif variations , and assessed how drastically each substitution affected expression . The power of the approach comes from the statistics – we gathered data for thousands of genes and motifs in dozens of growth conditions . We found that not all nucleotide substitutions are equal: e . g . substituting Adenines is more likely to be benign , compared to Guanines . It is possible that the different chemical nature of nucleotides could explain these findings . One future implication of this type of work is that it may aid in predicting which human mutations in promoter elements are more likely to cause a disease by affecting transcription . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"genetics",
"and",
"genomics/functional",
"genomics",
"genetics",
"and",
"genomics/gene",
"expression",
"computational",
"biology/genomics",
"computational",
"biology/transcriptional",
"regulation"
]
| 2008 | Functional Characterization of Variations on Regulatory Motifs |
Shoot apical meristems ( SAM ) are resistant to most plant viruses due to RNA silencing , which is restrained by viral suppressors of RNA silencing ( VSRs ) to facilitate transient viral invasion of the SAM . In many cases chronic symptoms and long-term virus recovery occur , but the underlying mechanisms are poorly understood . Here , we found that wild-type Cucumber mosaic virus ( CMVWT ) invaded the SAM transiently , but was subsequently eliminated from the meristems . Unexpectedly , a CMV mutant , designated CMVRA that harbors an alanine substitution in the N-terminal arginine-rich region of the coat protein ( CP ) persistently invaded the SAM and resulted in visible reductions in apical dominance . Notably , the CMVWT virus elicited more potent antiviral silencing than CMVRA in newly emerging leaves of infected plants . However , both viruses caused severe symptoms with minimal antiviral silencing effects in the Arabidopsis mutants lacking host RNA-DEPENDENT RNA POLYMERASE 6 ( RDR6 ) or SUPPRESSOR OF GENE SILENCING 3 ( SGS3 ) , indicating that CMVWT induced host RDR6/SGS3-dependent antiviral silencing . We also showed that reduced accumulation of the 2b protein is elicited in the CMVWT infection and consequently rescues potent antiviral RNA silencing . Indeed , co-infiltration assays showed that the suppression of posttranscriptional gene silencing mediated by 2b is more severely compromised by co-expression of CPWT than by CPRA . We further demonstrated that CPWT had high RNA binding activity leading to translation inhibition in wheat germ systems , and CPWT was associated with SGS3 into punctate granules in vivo . Thus , we propose that the RNAs bound and protected by CPWT possibly serve as templates of RDR6/SGS3 complexes for siRNA amplification . Together , these findings suggest that the CMV CP acts as a central hub that modulates antiviral silencing and VSRs activity , and mediates viral self-attenuation and long-term symptom recovery .
RNA silencing is a well-established plant antiviral response triggered by viral double-stranded RNAs or highly structured single-stranded RNAs in host plants . Host Dicer-like ( DCL ) enzymes cleave both RNA types into 21 to 24 nucleotide ( nt ) small interfering RNAs ( siRNAs ) that are subsequently sorted into Argonaute-containing RNA-induced silencing complexes ( RISC ) to guide specific cleavage of the cognate viral RNAs [1–7] . Some cleavage products serve as templates for host RNA-directed RNA polymerase ( RDR ) 1 , or RDR6 , to synthesize abundant de novo dsRNAs that are processed by DCLs into secondary siRNAs that enhance antiviral RNA silencing [5 , 7–11] . In addition , a plant-specific RNA binding protein , Suppressor of Gene Silencing 3 ( SGS3 ) , is required for siRNA amplification through forming complexes with RDR6 [12] . RNA silencing , as a major defense mechanism , occurs in all virus-infected tissues and is an extensive feature in newly emerging tissues [11 , 13–15] . Recent studies have revealed that symptom recovery from viral infection is generally concomitant with induction of RNA silencing in the shoot apices of infected plants and this depends in part on RDR activity . For example , Potato virus X ( PVX ) , Turnip crinkle virus ( TCV ) , and Potato spindle tuber viroid ( PSTVd ) transiently invade the meristems of plant mutants defective in host RDR6 [11 , 13 , 14] . As a counter-defense against host RNA silencing , many plant viruses have evolved VSRs to block various RNA silencing steps [16–18] . Some VSRs are also viral pathogenicity factors during systemic infections . For instance , the 2b protein of CMV and the 16K protein of Tobacco rattle virus ( TRV ) facilitate shoot apical meristem ( SAM ) invasion by blocking antiviral RNA silencing [19–21] . Ectopic expression of VSRs , like the potyvirus HC-Pro protein , enhance viral RNA accumulation of two distinct nepoviruses and prevent symptom recovery [22 , 23] . Nonetheless , viral meristem invasion is transient and is followed by long-term meristem exclusion [19 , 20] . Hence , the mechanisms of long-term recovery from transient SAM invasion remain to be elucidated . CMV is the type virus of the genus Cucumovirus in the family Bromoviridae . The CMV genome is composed of three positive-stranded RNAs [24] . RNA1 and RNA2 encode 1a and 2a proteins , respectively , which comprise the viral RNA-dependent RNA polymerase subunits [24] . RNA2-derived subgenomic RNA4A encodes 2b protein , a well-known VSR and determinant factor in viral virulence and systemic infection [5 , 25 , 26] . RNA3 and its subgenomic RNA4 encode movement protein ( MP ) and coat protein ( CP ) that are required for viral cell-to-cell and systemic movements , respectively [24] . The CMV CP is a multifunctional factor that has roles in viral systemic movement , host range and aphid transmission [27–31] . For the Pepo and MY17 CMV strains , CP-mediated cell-to-cell movement is implicated in SAM invasion of host plants [32] . In addition , the CMV CP is an aphid transmission determinant that mediates viral spread between host plants [28 , 31] . These properties and studies of different CMV strains suggest that CMV CPs are key host range determinants [30] . Previous studies have also shown that some amino acid residues of CMV CP contribute to various symptoms induction . For instance , CMV pepper strain mutants , in which proline 129 is replaced by 19 other amino acids , induce various systemic symptoms in plants [33] . Moreover , CMV CP amino acid 129 also determines viral invasion of the SAM in tobacco plants [32] . Although CMV CPs have been extensively studied as factors involved in positive regulation of viral spread and symptom induction , their negative roles in SAM infections have not been described . In the current study , we found that the N-terminal R-rich region ( R13RRRPRR19 ) of the CMV CP has a negative role in persistent viral SAM invasion . Our data indicate that elevated expression levels of the CMV CP induces potent antiviral RNA silencing by down-regulating the accumulation level of 2b VSRs and inducing siRNA amplification . Thus , we propose a novel self-attenuation mechanism , in which the CMV CP antagonizes the suppression effects of 2b protein and plays a pivotal role in regulating compatible interactions between CMV and host plants to prevent viral over-accumulation and persistent viral invasion of the SAM .
Plant virus-encoded CPs mainly participate in encapsidation and movement [34–38] , and are increasingly appreciated as an important regulator of viral RNA replication and translation that are associated with CP RNA binding affinity [15 , 39 , 40] . Protein sequence analyses revealed that the N-terminal region ( R13RRRPRR19 ) of the CMV CP is enriched in basic and positively charged amino acid residues that contribute to functional RNA binding activities . To explore the requirements of this R-rich region in viral infection , the basic amino acid residues were substituted by alanine ( R13-19: A ) , and the resulted mutant was designated as CPRA ( Fig 1A ) . To compare the functions of CPWT and CPRA in the context of viral sequences , we first developed an Agrobacterium tumefaciens-mediated CMV infection system in N . benthamiana plants . The cDNAs of the three CMV genomic RNAs were engineered into pCass4-Rz to generate pCass-RNA1 , -RNA2 , and -RNA3 ( Fig 1A ) . The CPWT- and CPRA-containing viruses are named CMVWT and CMVRA , respectively . Infections were carried out by co-infiltration of equal concentrations of A . tumefaciens EHA105 mixtures harboring pCass-RNA1 , -RNA2 , and -RNA3 into N . benthamiana leaves . We first explored whether CMVRA could form normal viral particles during viral infections . The CMVWT and CMVRA-infected leaves were homogenized for virions purification , and the purified virions were observed by transmission electron microscopy , which showed that CMVWT and CMVRA formed viral particles with similar appearance ( S1A Fig ) . Nonetheless , the CMVRA particles were less stable than CMVWT virions in an RNase assay ( S1B Fig ) . These results indicate that the CP R-rich motif is not essential for virion assembly or systemic infection in N . benthamiana plants . The susceptibility of N . benthamiana was examined to determine the function of the R-rich region in systemic infections . At 7 dpi , CMVWT induced mosaic symptoms in the fully expanded leaves , but only elicited limited or recovered symptoms in shoot apices of infected N . benthamiana plants ( Fig 1B , right panel ) . In sharp contrast , CMVRA infections resulted in severely distorted newly emerging leaves ( Fig 1B , middle panel ) . Subsequently , CMVRA-infected plants exhibited extremely short internodes and petioles at shoot apices to produce a rosette appearance combined with substantial stunting between 14 and 21 dpi ( Fig 1C ) . At 42 dpi , CMVWT-infected plants developed mosaic symptoms in leaves present in the central parts of the stems , and exhibited substantial reductions in growth compared to mock-infected plants ( Fig 1D ) . However , symptoms in the shoot apices of CMVWT-infected plants were modulated and the apices maintained apical dominance ( Fig 1D ) . In contrast , all of the CMVRA-infected plants developed several lateral shoots without distinguishable primary stems ( Fig 1D ) . These symptoms suggested that the shoot apices were severely infected and that apical dominance was disturbed leading to production of lateral bud outgrowths [41] . Since the 2b protein is a strong silencing suppressor [5 , 25 , 26] , we next examined whether the persistent SAM invasion by CMVRA depends on the 2b protein . A 2b-deleted mutation ( CMV-Δ2b ) was engineered into pCass-RNA2 by point mutations as described previously [6] , and co-infiltrated into N . benthamiana with pCass-RNA1 and wild-type pCass-RNA3 ( CMVWT-Δ2b ) , or mutated pCass-RNA3-CPRA ( CMVRA-Δ2b ) . CMVWT-Δ2b caused mild mosaic symptoms in the systemic leaves , whereas CMVRA-Δ2b did not induce any obvious symptoms ( S2A Fig ) . To explore the replication of CMVWT-Δ2b and CMVRA-Δ2b at 7 dpi , the infiltrated leaves were sampled and the CP accumulation was detected by Western blotting at 7 dpi . Both CMVWT-Δ2b and CMVRA-Δ2b CPs had accumulated to similar levels ( S2B Fig ) , suggesting that the 2b deletion did not affect virus proliferation in infiltrated leaves . In contrast , CMVWT-Δ2b CP was present in the upper uninfiltrated leaves at 7 dpi , but CMVRA-Δ2b CP was not detected ( S2C Fig ) . RT-PCR was performed to detect viral RNA accumulation in the upper leaves with primers corresponding to the RNA3 CP region . CP specific bands were detected in plants inoculated with CMVWT-Δ2b , but not with CMVRA-Δ2b ( S2D Fig ) . Collectively , these data demonstrate that CMV CPWT exerts a negative role in shoot apex infections , and 2b protein is required for systemic infection of the CMVRA mutant . Because apical dominance was abolished in CMVRA infections , we proposed that CMVRA invaded meristems and altered the meristematic activity of the infected plants . To explore this possibility , longitudinal sections of the topmost flowers and shoot apices from mock- , CMVWT- , and CMVRA-infected plants were examined by in situ hybridization with digoxigenin-labeled CMV RNA3 probes . CMVWT invaded most shoot meristems and floral primordia by 7 dpi , but subsequently disappeared between 14 and 21 dpi , despite of some detectable signals below the SAM and floral primordia ( Fig 2 , middle panels ) . In contrast , CMVRA was abundant below the meristems at 7 dpi , and partially moved into the meristems by 14 dpi ( Fig 2 , right panels ) . CMVRA invaded the meristems of all the infected plants by 21 dpi ( Fig 2 , right panels ) , but as expected were absence in meristems from mock inoculated plants ( Fig 2 , left panels ) . These results demonstrate that CMVRA accumulated to high levels and resulted in severe stunting and abolished apical dominance , whereas CMVWT-infected plants recovered from meristem infection . Because RNA silencing is a key antiviral mechanism in meristems infections [11 , 13 , 14 , 19–21] , we analyzed levels of viral RNA and siRNAs by Northern blotting at 7 dpi . The newly grown tissues were collected for the Northern blotting analysis of viral RNA and siRNA at 7 dpi . A markedly increased accumulation of viral RNA of CMVRA was detected in the newly grown tissues compared with those of CMVWT ( Fig 3A , compare lanes 3 , 4 with 5 , 6 ) . However , CMVRA RNA3-derived siRNAs ( RNA3-vsiRNAs ) were present at a much lower level than those of CMVWT ( Fig 3B , compare lanes 3 , 4 with 5 , 6 ) . To determine the relative accumulation levels of viral genomic RNA3 and RNA3-vsiRNAs , the hybridization signal intensities in three independent experiments were evaluated . The values of RNA3 and RNA3-vsiRNAs in CMVRA-infected plants were set as one unit . CMVWT RNA3 accumulated to about 30% of CMVRA RNA3 ( Fig 3C , left panel , P-value < 0 . 001 ) . However , the levels of CMVWT RNA3-vsiRNAs were 2 . 3 times those of CMVRA RNA3-vsiRNAs ( Fig 3C , middle panel , P-value < 0 . 01 ) . The relative RNA3-vsiRNAs/RNA3 ratio in the CMVWT-infected newly emerging tissues was six-fold higher than that of CMVRA-infected plants ( Fig 3C , right panel , P-value < 0 . 01 ) , indicating that CMVWT induces more potent antiviral silencing than CMVRA in the emerging tissues . Collectively , CMVWT promotes the accumulation of vsiRNAs that inhibit virulence in the shoot apices of infected plants . In contrast , the lower amounts of CMVRA vsiRNAs reduced antiviral silencing and enabled CMVRA to persistently infect the newly emerging tissues . To provide more detailed genetic analyses of host effects on CMV infection , we investigated the impacts of the CMV CP derivatives during infection of Arabidopsis thaliana . CMVWT induced visible disease symptoms in the fully expanded leaves of Columbia-0 ( Col-0 ) wild-type plants at 21 dpi ( Fig 4A and S3A Fig ) . These plants had similar bolting times as the mock inoculated plants and maintained apical dominance during late infection stages between 35- and 56- dpi ( S3A Fig ) . Compared with CMVWT , CMVRA infection resulted increased numbers of distorted leaves in the newly emerging tissues ( Fig 4A and S3A Fig ) , late bolting , and reduced apical dominance in the late infection stages ( S3A Fig ) . Northern blotting results of viral RNA and vsiRNAs extracted from CMVWT- and CMVRA-infected systemic leaves at 21 dpi were consistent with those of infected N . benthamiana , showing that CMVRA infected plants accumulates higher levels of viral genomic RNA and less vsiRNAs compared with CMVWT infections in both plant species ( Fig 4B and S3 Fig ) . Thus , CMVWT elicits more potent antiviral silencing than CMVRA in emerging tissues of infected A . thaliana , and is similar in this regard to infected N . benthamiana plants . To explore whether CMVWT induction of potent antiviral silencing depends on host RDRs , we inoculated A . thaliana rdr6 and sgs3 mutants with CMVWT or CMVRA . The rdr6 and sgs3 mutants infected with CMVRA displayed severe symptoms similar to those of wild-type plants and whole plant development was restrained ( Fig 4A and Fig 4C ) . The rdr6 and sgs3 mutant plants infected with CMVWT exhibited more severe disease symptoms in the newly emerging tissues compared with wild-type plants ( Fig 4A and Fig 4C ) . In addition , Northern RNA hybridizations revealed that the accumulation level of viral RNA in CMVWT-infected wild-type plants was dramatically lower than those of infected rdr6 or sgs3 plants , whereas higher levels of vsiRNAs accumulated in infected wild-type plants than in rdr6 or sgs3 plants ( Fig 4B , compare lanes 3 , 5 , and 7 ) . In contrast , high levels of viral RNA and low levels of vsiRNAs were detected in the emerging tissues of Col-0 , rdr6 and sgs3 mutants infected with CMVRA ( Fig 4B , compare lanes 2 , 4 , and 6 ) . Collectively , these results demonstrate that RDR6 and SGS3-dependent amplification of vsiRNAs is required for the CMVWT-induced potent antiviral RNA silencing in emerging tissues of CMVWT infected plants . We next investigated the accumulation of the CP and 2b protein in emerging tissues of CMV infected Col-0 , rdr6 , and sgs3 mutants . Accumulation of CMVWT CP in the rdr6 and sgs3 mutants increased to level similar to those of CMVRA CP ( Fig 4D , top panel ) . Notably , accumulation of CMVWT 2b protein was lower level in the rdr6 and sgs3 mutants compared with CMVRA infection , indicating that the accumulation of 2b protein was significantly down-regulated in CMVWT-infected plants ( Fig 4D , middle panel ) . Collectively , these results demonstrate that the down-regulated 2b protein cannot efficiently inhibit RDR6/SGS3-dependent antiviral silencing that restricts elevated accumulation of CMVWT in emerging tissues . Our findings have demonstrated that CPWT exerts negative effects during viral SAMs infection . To independently verify the antagonistic roles of CP and 2b , we next examined the effects of CPWT and CPRA on VSR activities of 2b by co-infiltration assays in N . benthamiana plants [42] . Green fluorescence occurring early during transient co-expression of GFP disappeared completely at 5 dpi , indicating that potent silencing was induced ( Fig 5A ) . In contrast , high intensity of GFP fluorescence in the regions of leaves co-expressing 2b and GFP suggested that GFP silencing was efficiently suppressed by the 2b protein ( Fig 5A ) . We next compared local GFP silencing suppression by 2b co-expressed with the empty vector ( V ) , CPWT , or CPRA , respectively . The CPWT , unlike V or CPRA , significantly attenuated the 2b-mediated suppression of GFP silencing ( Fig 5A ) . The observations were further verified by Western blotting showing that co-expression of CPWT markedly reduced the expression levels of GFP protein expression compared to co-expression of V and CPRA ( Fig 5B , top panel ) . Simultaneously , the CP and 2b protein also accumulated to lower level during co-expression with CPWT than co-expression with V and CPRA ( Fig 5B , middle and bottom panels ) . To examine the silencing potency in infiltrated regions of N . benthamiana leaves , the accumulation of GFP mRNA and siRNAs was compared by Northern blotting analyses , and hybridization signal densities were measured to determine the relative accumulation of mRNA and siRNAs . The values from leaf samples co-expressing GFP , 2b and V were set as one unit . The GFP mRNA expression level in the sample co-expressing 2b protein and CPWT significantly decreased compared with those of V or CPRA ( Fig 5C , top panel and RA1 values , compare lane 4 with lanes 3 and 5 ) , but all patches accumulated similar levels of GFP-derived siRNAs ( Fig 5C , middle panel and RA2 values , compare lane 4 with lanes 3 and 5 ) . The relative accumulated levels of GFP siRNAs versus GFP mRNA ( RA2/RA1 ) were compared . The GFP siRNAs/mRNA ratios in the patches co-expressing 2b and CPWT were at least four-fold higher than those co-expressing 2b with V or CPRA ( Fig 5C , RA2/RA1 values , compare lane 4 with lanes 3 and 5 ) , demonstrating that CPWT enhances the potency of GFP silencing . During CMV infection , the accumulation of the CP increases gradually during viral propagation . Therefore , we wondered whether CP affected VSR activities of 2b in a dose-dependent manner . To answer this question , we compared the GFP fluorescence from the different patches infiltrated with a CP concentration gradient ( OD600 = 0 , 0 . 1 , 0 . 2 , 0 . 4 , and 0 . 8 ) and 2b ( OD600 = 0 . 2 ) ( Fig 5D , middle panel ) . A low concentration of infiltrated CP ( OD600 = 0 . 1 ) had negligible effects on 2b-mediated suppression of GFP silencing , whereas increasing CP concentrations gradually compromised the inhibitory effects as the increasing concentration of CP , and the highest CP concentration ( OD600 = 0 . 8 ) significantly down-regulated the 2b suppression ( Fig 5D , left panel ) . Additionally , Western blotting analyses revealed that the accumulation of GFP and 2b decreased in proportion to the increasing concentrations of co-expressed CP ( Fig 5D , right panel ) . Thus , our results indicate that low CMV CP levels do not affect 2b suppressive activities , whereas highly abundant CMV CP concentrations down-regulate 2b protein accumulation and attenuate 2b-mediated silencing suppression . The R-rich region of CP is highly conserved in many subgroup I and subgroup II CMV strains , as well as in Tomato aspermy virus ( TAV ) ( S4A Fig ) , and all the CPs exert negative effects on 2b-mediated suppression of RNA silencing ( S4 Fig ) . In the field , synergistic viral diseases are usually caused by interactions of different viruses that result in dramatically increased viruses titer and symptom induction , which are mainly dependent on VSRs activities [43] . With regard to the co-infection of CMV with other plant viruses in the field , we postulated that CMV CP compromises the suppression activity of other VSRs in co-infected plants . To test this hypothesis , we co-infiltrated CP with P19 , P38 , or HC-Pro in N . benthamiana leaves . In agreement with 2b , the suppressions mediated by the VSRs were attenuated by coinfiltration with CPWT , but this effect was not observed with the CPRA ( S5A Fig ) . The results were also confirmed by Western blotting which revealed a substantial reduction of GFP accumulation in the patches co-expressing of VSRs and CPWT compared with VSRs and CPRA ( S5B Fig ) . At the same time , the co-expression of CPWT rather than CPRA also decreased the accumulation of the VSRs ( S5B Fig ) . We further demonstrated that the compromising effect of the CMV CP on P19-mediated suppression was also dependent on CPWT concentrations ( S5C Fig ) . In conclusion , highly abundant CMV CP concentrations compromise various VSRs suppression activities in patch assays , implying that the CMV CP modulates the synergistic viral disease by regulating silencing interactions and VSRs in the co-infected plants . The decreased accumulation of VSRs by high-abundant CMV CP might be due to the strong RNA binding activity of CPWT and resulting in translation inhibition . To test this possibility , we carried out North-Western blotting assays to compare the RNA binding affinity of CPWT and CPRA . GST-tagged CPWT , CPRA , and untagged GST were expressed and purified from E . coli expression systems . Different amounts ( 1 μg , 2 μg , and 4 μg ) of GST-CPWT and GST-CPRA were separated in SDS-PAGE gels , transferred to nitrocellulose membranes , renatured , and exposed to digoxigenin-labeled CMV RNA4 or luciferase ( Luc ) mRNA . High amounts ( 4 μg ) of GST served as a negative control and failed to binding the RNAs ( Fig 6A , lane 7 ) , confirming that the GST tag does not have RNA-binding affinity . The GST-CPWT bound much higher levels of CMV RNA4 and Luc mRNA than the GST-CPRA protein ( Fig 6A , compare lanes 2 , 4 and 6 with lanes 1 , 3 and 5 ) . Together , these results clearly indicate that the R-rich region is important for the high RNA binding affinity of CMV CPWT . We further examined whether the high unspecific RNA binding activity of CPWT resulted in translation inhibition by comparing in vitro translation efficiency in the wheat germ system . The results showed that high concentrations of GST tag did not affect the translation of Luc mRNA ( Fig 6B , black line ) . The highest concentration ( 66 . 7 μg/ml ) of GST-CPRA had a~ 20% translation reduction compared with the empty control ( P-value < 0 . 01 ) ( Fig 6B , blue line ) , indicating that the weak RNA binding of the GST-CPRA protein partially affected mRNA translation . However , more dramatic reductions in the luciferase translation effects were observed as the GST-CPWT concentration was gradually increased , and finally the highest concentration ( 66 . 7 μg/ml ) , was ~1% of the empty control ( P-value < 0 . 001 ) ( Fig 6B , red line ) . These results are in agreement with previous studies , in which the Potato virus A ( PVA ) CP was found to be involved in inhibition of viral RNA translation [44 , 45] . Therefore , it appears that the strong RNA-binding affinity and resulting translation inhibition by the CMV CPWT contributes to reduced VSRs accumulation in CMVWT virus infections . In previous studies , high CP concentrations of several positive-sense RNA viruses repressed RNA translation and facilitated virion assembly [44 , 46 , 47] . Similarly , our results show that CMV CP also efficiently inhibits translation of Luc mRNA in wheat germ system ( Fig 6B ) . In addition to initiating viral encapsidation and/or CMV ribonucleoprotein ( RNP ) formation , translation inhibition also contributes to elevated accumulation of aborted mRNA transcripts without translation that might be recognized as aberrant RNA by host RDR to initiate siRNAs amplification . SGS3 , acting as co-factor of RDR6 , mainly binds to and stabilizes RNA substrates to amplify secondary siRNAs [48–50] . To visualize potential CP–RDR6/SGS3 protein associations in living cells and their subcellular occurrence , we conducted bimolecular fluorescence complementation ( BiFC ) assays with Agrobacterial-infiltrated leaves of N . benthamiana . For this assay , the SGS3 and RuBisco proteins ( Rub ) were fused with the N-terminal half of sYFP , and the tagged CPWT , CPRA and Rub proteins were fused with the C-terminal half of sYFP . YFPC-CPWT and YFPC-CPRA could be associated with YFPN-SGS3 in the cytoplasm , and formed punctate granules that co-localized with RFP-tagged RDR6 ( Fig 7A , top two panels ) . BiFC fluorescence was not detected in the Rub control samples that were co-expressed with CP or SGS3 ( Fig 7A , three bottom panels ) . To further evaluate the CP–SGS3 protein association in vivo , we performed co-immunoprecipitation ( co-IP ) assays in planta . In these experiments , Flag-CPWT , Flag-CPRA proteins were co-expressed with GFP-SGS3 or GFP proteins , and extracts from infiltrated leaves were used in co-IP assays with anti-Flag beads . Both the Flag-CPWT and Flag-CPRA , immunoprecipitated GFP-SGS3 efficiently , but the GFP protein did not ( Fig 7B ) . Nevertheless , we could not detect the interaction of CPs and SGS3 through a yeast-two-hybrid assay ( S6 Fig ) , hence the association of CPs and SGS3 in N . benthamiana might be indirect . Collectively , both CPWT and CPRA appear to colocalize with SGS3 in punctate granules in vivo and RDR6 also is present in these granules . However , in contrast with CPRA , the CPWT protein has a high affinity with RNA to result in general inhibition of host RNA and viral RNA translation . This inhibition results in production of aberrant RNAs that in turn appear to be associated with SGS3 and RDR6 for siRNA amplification . In summary , we conclude that the CMV CPWT protein RNA binding contributes indirectly to high potency RNA silencing that presumably results in reduced accumulation of the CMV 2b protein . Reductions in 2b VSRs activities in turn result in a cycle in which increased siRNAs production by RDR6/SGS3 dependent amplification leads to reductions in CMV RNAs , elevated virus attenuation , and protracted symptom recovery in newly emerging leaves of infected plants .
Symptom recovery represents an extreme virus attenuation effect , in which , infected plants initially develop sever leaf symptoms , but subsequently newly emerging leaves exhibit a drastically reduced virus accumulation due to induction of antiviral RNA silencing [15 , 51 , 52] . In previous studies , host RNA silencing effects and interactions with the CMV Pepo strain 2b protein were shown to be involved in transient appearance of CMV in meristems at 7 dpi , and decreases in virus concentration as new leaves emerged and disappearance in recovered tissues [19 , 21] . In agreement with these results , we found that CMV Fny strain also infected SAM transiently and then was excluded from the shoot apices leading to symptom recovery ( Fig 1 and Fig 4 ) . Our studies unexpectedly found that a mutant harboring an alanine substitution in the N-terminal R-rich region of the CMV CP could persistently invade meristems and block the growth of apical shoots , implying that CMV CPWT facilitates long-term SAM exclusion at late infection stages ( Fig 1 and Fig 2 ) . The previous studies have shown that the amino acid 129 of the Pepo CMV CP affects the cell-to-cell movement and determines successful SAM invasion in tobacco plants at 6–8 dpi , but the shoot meristems recovered from the Pepo infection at 21 dpi [21 , 32] . Combined with previous studies , our results indicate that CMV CP is not only required for successful SAM invasion at the early infection , but also modulates viral exclusion from SAM later in infection . Thus , we propose that the CMV CP plays a critical role in compatible interactions between CMV and host plants . Plant viral CPs have a primary function involving viral genome encapsidation , but also have been implicated in viral translation and/or replication . At low concentrations , viral CPs usually facilitate RNA replication and/or translation , whereas at higher concentrations , they may inhibit these processes in favor of virion assembly [39 , 40 , 44–46] . For instance , the Brome mosaic virus ( BMV ) CP binds to an RNA element within the 5´UTRs of the viral genome and suppresses the translation of RNA replication proteins [46] . In addition , the Hepatitis C virus ( HCV ) core protein binds specifically to the internal ribosome entry site ( IRES ) in the 5´UTR of the viral genome [47 , 53–55] . This binding requires positively-charged residues in the N-terminal portion of the core protein and results in suppression of translation of downstream genes . In addition , lysine-to-alanine mutations in the N-terminal region of Red clover necrotic mosaic dianthovirus ( RCNMV ) CP induced more severe symptoms than wild-type virus in N . benthamiana , indicating that the lysine-rich N-terminus of RCNMV CP modulates symptomatology , independently of its role in virion assembly [56] . In line with these examples , we have shown that the CMV CP has high unspecific RNA binding activities and inhibits the translation of Luc mRNA protein in the wheat germ system ( Fig 7 ) . The mutant CPRA harboring an alanine substitution in the N-terminal R-rich region was significantly compromised in translation inhibition , implying that the basic and positively charged amino acid residues are required for translation repression by the CMV CP ( Fig 7 ) . The high concentrations of CMV CP in the wheat germ system resemble CP concentrations at late stages of infection when the CP is among the most prominent proteins in the cell . Additional experiments are required to explore the binding regions of CMV CP in the mRNA and the viral genomic/subgenomic RNA interactions . Given the well-known functions of 2b as viral virulence determinants and silencing suppression [6 , 9 , 26 , 57–63] , we postulate that decreased accumulation of 2b protein by saturated CP directly or indirectly contributes to potent antiviral RNA silencing that resulting in virus exclusion from SAM and symptom recovery . However , we cannot exclude that other host and/or viral components possibly affected by the CP are involved in symptom recovery . For example , non-specifically binding of RNAs by the CPWT protein might disturb metabolic processes of the host and lead to low virus accumulation in the SAM . Another possibility is that unstable particles assembled by the CPRA mutant permit re-infections in the same cells and therefore lead to higher RNA accumulation . Future studies are anticipated to provide exciting insights into symptom recovery processes . In the process of RNA silencing , host RDRs are required to initiate or amplify RNA silencing via dsRNA synthesis , and the substrates for dsRNA synthesis in vivo are aberrant RNA lacking a cap structure or poly ( A ) tails . Arabidopsis ETHYLENE-INSENSITIVE5 ( EIN5 ) / EXORIBONUCLEASE4 ( XRN4 ) encodes a cytoplasmic 5′-3′ exoribonuclease that degrades RNA intermediates derived during mRNA decay and/or RISC slicing , and regulates RDR6-dependent production of siRNAs [64 , 65] . Arabidopsis Super-Killer2 ( SKI2 ) functions as a cytoplasmic SKI complex to unwind RNAs into the 3´to 5´exoribonuclease complex for decay , and acts as a repressor of endogenous PTGS [66] . Therefore , both 5´to 3´and 3´to 5´cytoplasmic RNA decay pathways function in RDR-dependent silencing . Here , we propose that the high binding activity of CPWT with RNA protects viral RNA intermediates from RNA decay , which increases the substrate concentration of RDR/SGS3 complex and subsequently improves host antiviral silencing . We consistently found that CMVWT-induced antiviral silencing is significantly compromised in emerging leaves of sgs3 and rdr6 Arabidopsis mutants compared with those of wild-type Arabidopsis plants ( Fig 4 ) . As the result of our comparative analyses of CPWT and CPRA in the virus infection and GFP co-infiltration assays , we propose the schematic model shown in Fig 8 . At early stages of infection , a low level of CP fails to efficiently inhibit 2b protein accumulation or to induce siRNA amplification , which facilitates high-speed replication of viral RNA . Later in infection , abundant CP binds to viral RNA to initiate virion packaging , inhibit RNA translation and facilitate ribonucleoprotein formation for viral movement , as is consistent with the CP functions of other plant viruses , such as BMV and PVA [44 , 46] . Simultaneously , saturated CP results in decreased accumulation of 2b protein , which interferes with inhibition of host antiviral RNA silencing . In our experiments , the CP bound RNAs failed to participate in translation and the mRNAs likely were recognized as aberrant RNAs by the RDR/SGS3 complex , and used as substrates for siRNA amplification . Accordingly , the high amounts of CP found at late infection stages reduces synthesis and accumulation of the 2b protein and/or induces siRNA amplification to culminate viral clearance from the shoot apices of infected plants ( Fig 8 ) . By comparison , the much lower affinity of CPRA with viral RNA fail to reduce the accumulation of 2b protein efficiently , or induce siRNA amplification . Therefore , we propose that the VSR activities of 2b facilitate continuous infection of CMVRA in SAM regions . Although , it must be noted here that the proposed model is based only on N . benthamiana and Arabidopsis as the host plants , but since CMV infects more than 1000 species , our proposed model provides the basis for a variety of other experiments in a diverse assay of host plants . Since a relative healthy host provides a better environment for multiplication within-host and between-host transmission , many plant viruses down-regulate viral virulence to avoid severe disease in order to promote the survival of their hosts [67] . For instance , the strong suppressor P0 suppressor encoded by poleroviruses does not accumulate to detectable levels because of suboptimal translation initiation , which leads to low suppressor activity and reduced viral pathogenicity [68] . The Tobacco mosaic virus movement protein also regulates the spread of RNA silencing to self-control viral propagation [69] . Here , our study has revealed a novel self-attenuation mechanism in which the suppression effects of the 2b protein are down-regulated by saturated CP . Collectively , the roles of the interactions between RNA silencing and virus-encoded suppressors in the co-evolution of hosts and pathogens have been extensively investigated . Our results demonstrate that the CMV CP serves as a central hub to facilitate regulation of dynamically integrated connections between antiviral silencing and VSR activities .
N . benthamiana plants were grown in a growth room with a controlled environmental climate programmed for 16 hours ( hrs ) of light at 24°C and 8 hrs in the dark at 21°C . Seedlings with six to eight fully expanded leaves were used for virus inoculations . For Agrobacterium tumefactions constructions , the cDNAs of RNA1 , RNA2 , and RNA3 , as well as RNA2-Δ2b and RNA3-CPRA were amplified , digested with Stu I and BamH I , introduced into pCass4-Rz , and transformed into A . tumefaciens strain EHA105 [70] . Equal amount of agrobacteria harboring CMV plasmid derivatives were mixed and infiltrated into N . benthamiana leaves as described previously [71] . All the experiments were repeated at least three times with reproducible results . Arabidopsis thaliana rdr6-15 ( SAIL_617_H07 ) and sgs3-1 in Columbia ( Col ) ecotype were described previously [6] . After vernalized in the dark at 4°C , the seeds were transferred into a growth room with the condition of 10 hrs in light and 14 hrs in dark at 22°C . CMVWT and CMVRA virions propagated in N . benthamiana leaves were purified and used as a inocula at 100 μg/mL . Shoot and floral apices of infected plants were collected from infected N . benthamiana , embedded in wax , sectioned , and in situ hybridized as described previously [14] . CMV RNA was detected with the digoxigenin ( Roche Diagnostics GmbH ) labelled vitro-transcribed RNA fragment corresponding to 3′ terminal 200 nucleotides of CMV RNA3 , and then detected with antibody anti-digoxigenin conjugated to alkaline phosphatase ( Roche Diagnostics GmbH ) with nitroblue tetrazolium ( NBT ) and 5-Bromo-4-Chloro-3-Indolyl Phosphate ( BCIP , Sigma ) . Stained samples were examined with a bright-field microscope ( DP72 , Olympus ) for visualization and photography . For transient protein expression in N . benthamiana leaves , CMV CPWT , CPRA and 2b cDNAs were introduced into the pGD binary vector [72] . Leaves of 4-week-old N . benthamiana plants were co-infiltrated with mixed Agrobacterium cultures harboring the positive sense GFP ( sGFP ) expression plasmid with different combinations of empty vector ( V ) , 2b , and CP plasmids . At 5 dpi , GFP fluorescence in infiltrated leaves was recorded under a long wavelength UV lamp ( UVP , California , USA ) using a 600D Cannon digital camera [62] . Local suppression assays were independently performed at least three times with reproducible results . The topmost infected leaves of 10 to 15 plants were pooled for RNA extraction with Trizol reagent according to the manufacturer instruction ( Invitrogen , USA ) . As described previously [6] , 5 μg and 10 μg total RNAs were used for detection of viral RNA and vsiRNAs , respectively . CMV genomic and sgRNAs cDNA detection probes from the 3′ terminal 240 nt of Fny-CMV RNA2 was randomly labeled with [α-32P] dCTP . VsiRNAs were detected by the labeled DNA oligonucleotides corresponding to CMV gRNA3 as described previously [6] . The upper uninoculated leaves inoculated with CMVWT-Δ2b or CMVRA-Δ2b were collected for RNA extraction and detection at 7 dpi . Total RNA was treated with RNase free-DNase I , and used as a template for first-strand cDNA synthesis with M-MLV reverse transcriptase ( Promega , USA ) as described previously [73] . Viral infections were monitored by RT-PCR using primers corresponding to the RNA3 CP region . Protein phosphatase 2A ( PP2A ) was used as an RT-PCR control . For local silencing suppression assay , 5 μg and 10 μg total RNA from infiltrated leaves were used for GFP mRNA and siRNA detections , respectively . The randomly-labeled cDNA probe corresponding to GF region ( nt 1–400 ) of GFP cDNA was used for GFP mRNA detection . The GFP-derived siRNA was detected by [α-32P] UTP-labeled RNA probe corresponding to GF region of GFP . Total proteins extracted from viral inoculated or agro-infiltrated leaf tissue with SDS buffer [100 mM Tris ( pH 6 . 8 ) , 20% glycerol , 4% SDS , and 0 . 2% bromophenol blue , 10% β-mercaptoethanol] were separated in SDS-PAGE gels , and transferred onto nitrocellulose membranes . Anti-GFP ( 1:1000 ) , -CP ( 1:4000 ) , and -2b ( 1:2000 ) polyclonal antibodies were used to detect accumulation of GFP , CP , and 2b , respectively . Then , goat anti-rabbit IgG horseradish peroxidase conjugate at a 1:3000 diluted was applied as the secondary antibody , and the membranes were incubated with Pierce ECL Plus chemiluminescent substrate before exposure to x-ray films . Firstly , amplified CPWT and CPRA cDNAs were introduced into pGEX-KG vectors respectively , and the recombinant plasmids were transformed into BL21 . After induction with 0 . 2mM isopropyl β-D-thiogalactoside ( IPTG ) at 18°C for 18 h , the resulting GST-tagged fusion proteins and GST tags were purified over Glutathione Sepharose 4B ( GE Healthcare ) affinity columns according to the manufacturer’s instructions . The RNA binding assays were performed via a described North-Western blot procedure as described [74] . Briefly , 5 μg of purified GST , GST-CPWT , or GST-CPRA were separated by 12 . 5% SDS-PAGE and transferred to nitrocellulose membranes . The membranes were incubated with renaturation buffer ( 50 mM Tris-HCl , pH 7 . 5 , 0 . 1% TritonX-100 , 10% glycerol , 0 . 1 mM ZnCl2 and 250 mM KCl ) overnight at 4°C . Then , the membranes were transferred into binding buffer ( 10 mM Tris–HCl , pH 7 . 5 , 1 mM EDTA , 100 mM NaCl , 0 . 05% Triton X-100 , and 1X Denhardt’s reagent ) containing digoxigenin-11-UTP-labelled ( Roche ) RNA probe corresponding to CMV RNA 4 or Luc RNA . The bound RNA was blotted with the anti-digoxigen conjugated alkaline phosphatase ( 1:3000 dilution , Roche ) in a NBT/BCIP solution . The in vitro translation assay were performed as described previously [75] . First , the full-length luciferase ( Luc ) cDNA was introduced into the pMD19-T vector . Then , the bacteriophage T7 promoter and a poly ( A ) tail were inserted at the Luc cDNA N- and C- termini , respectively , and the resulting Xba I-linearized plasmid was used as a template for in vitro transcription by the mMESSAGE T7 kit ( Ambion , USA ) . Two micrograms ( μg ) of Luc mRNA and different concentrations of purified GST-CPWT , GST-CPRA or GST were translated in the Wheat Germ Extract Plus kit ( Promega , USA ) for 2 hours at 25°C . Then , luciferase activity of translated products was determined with a 20/20 luminometer ( Promega , USA ) as described previously [76] . Coimmunoprecipition of CMV CP and AtSGS3 was performed as described previously [77] . N . benthamiana plants were agroinfiltrated for transient expression of Flag-CPWT , Flag-CPRA with GFP-SGS3 or GFP and leaf tissues were homogenized in coimmunoprecipition buffer [10% glycerol , 25 mM Tris-HCl ( PH7 . 5 ) , 200 mM NaCl , 1 mM EDTA , 0 . 1% Tritonx-100 , 2% PVP-40 , 50 μM MG132 , 10 mM DTT and cocktail] . After filtration and centrifugation , the supernatants were incubated with anti-Flag M2 affinity gel ( Sigma ) for 2 hrs followed by washing five times . The immunoprecipitated products were boiled with SDS buffer for Western blotting assays with corresponding antibodies . BiFC assays were performed with minor modifications as described previously [78] . AtSGS3 and CPWT/RA cDNA fragments , and were the RuBisco control protein ( Rub ) were cloned into the BiFC vectors pSPYNE-35S and pSPYCE-35S , respectively . A . tumefaciens EHA105 strains containing the recombinant BiFC plasmids and the tomato bushy stunt virus P19 plasmid were co-infiltrated into N . benthamiana leaves at a final ratio of 0 . 5:0 . 5:0 . 3 ( OD600 ) and epidermal cells of infiltrated leaves were observed for fluorescence analysis ( YFP ) at 2 dpi using confocal laser scanning microscopy ( CLSM ) ( Olympus FV1000 ) . | In many virus-infected plants , the development of viral symptoms on the upper leaves gradually decline , until finally the top leaves appear normal and become resistant to secondary infection . Many documented cases suggest that symptom recovery is accompanied with antiviral RNA silencing . Most plant viruses encode viral suppressors of RNA silencing ( VSRs ) that can facilitate transient viral entry into meristems by blocking host RNA silencing defenses . However , the mechanisms of the following longer-term viral exclusion and modulation of VSRs functions remain elusive . Our studies with a substitution mutation in the CP gene demonstrate that the CMV CP has a negative role in SAM invasion . The studies suggest that during late infections , increasing CP concentrations induce potent RDR6/SGS3-dependent antiviral silencing by down-regulating accumulation of the 2b protein and induction of efficient siRNA amplification . Here , we propose a new evolutionary strategy in which the CMV CP has a role mediating viral self-attenuation and long-term symptom recovery . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"plant",
"anatomy",
"rna-binding",
"proteins",
"rna",
"interference",
"gene",
"regulation",
"rna",
"extraction",
"brassica",
"plant",
"physiology",
"plant",
"science",
"model",
"organisms",
"experimental",
"organism",
"systems",
"epigenetics",
"plants",
"extraction",
"techniques",
"research",
"and",
"analysis",
"methods",
"arabidopsis",
"thaliana",
"small",
"interfering",
"rnas",
"genetic",
"interference",
"proteins",
"gene",
"expression",
"leaves",
"biochemistry",
"rna",
"plant",
"and",
"algal",
"models",
"nucleic",
"acids",
"protein",
"translation",
"genetics",
"biology",
"and",
"life",
"sciences",
"meristems",
"non-coding",
"rna",
"organisms"
]
| 2017 | Cucumber mosaic virus coat protein modulates the accumulation of 2b protein and antiviral silencing that causes symptom recovery in planta |
Recent reports have questioned the accepted dogma that mammalian mitochondrial DNA ( mtDNA ) is strictly maternally inherited . In humans , the argument hinges on detecting a signature of inter-molecular recombination in mtDNA sequences sampled at the population level , inferring a paternal source for the mixed haplotypes . However , interpreting these data is fraught with difficulty , and direct experimental evidence is lacking . Using extreme-high depth mtDNA re-sequencing up to ~1 . 2 million-fold coverage , we find no evidence that paternal mtDNA haplotypes are transmitted to offspring in humans , thus excluding a simple dilution mechanism for uniparental transmission of mtDNA present in all healthy individuals . Our findings indicate that an active mechanism eliminates paternal mtDNA which likely acts at the molecular level .
In eukaryotes , cytoplasmic genes are generally inherited from the mother . The mechanisms responsible for this appear to differ between organisms . The elimination of mitochondrial DNA ( mtDNA ) during spermatogenesis prevents paternal transmission in Drosophila melanogaster [1] . Conversely , in the Japanese medaka fish Oryzias latipes , sperm mtDNA is lost after fertilization [2] . In cows and humans , sperm mtDNA is eliminated from two or four cell embryos [3 , 4] , and sperm loss may also occur throughout embryogenesis in mice [5] . Although sperm mitochondria are tagged with ubiquitin and actively destroyed through P62 and LC3-mediated autophagy [6 , 7] , there is no direct evidence showing the destruction of sperm mitochondrial DNA ( mtDNA ) [8] , and there are several examples where paternal mtDNA has escaped this process . Extensive paternal transmission of mtDNA has been observed in the marine mussel ( Mytillus sp ) [9] , and occasionally in the fruit fly ( Drosophila melanogaster ) [10] , Lepidopteran insects and the honey bee ( Apis mellifera ) [11] . For the most part , the “leakage” of paternal mtDNA during transmission in mammals has only been observed in atypical situations such as inter-strain breeding in mice ( Mus musculus ) [12] , or following in vitro embryo manipulation in cattle ( Bos taurus ) [13] . However , the description of a paternally-derived 2 base pair pathogenic deletion in the mtDNA MTDN2 in a 28-year-old man with a mitochondrial myopathy [14] , the persistence of human sperm-derived mtDNA when introduced into somatic cells [15] and in abnormal fertilised human embryos [16] , coupled with evidence of paternally transmitted mtDNA in healthy sheep ( Ovis aries ) [17] , and the great tit ( Parus major ) in its natural habitat [18] , questions the accepted dogma of exclusive maternal transmission . In humans , the debate hinges on the analysis of mtDNA sequences at the population level . By studying partial or complete mtDNA sequences from individuals across the globe , some have argued that the co-occurrence of phylogenetically unrelated genetic variants indicates inter-molecular recombination between paternal and maternal mitochondrial genomes [19] , but others have argued that the high mtDNA mutation rate confounds this analysis through the generation of homoplasy [20] , which can reach ~20% . Recent experimental data showed no evidence of the active removal of sperm mtDNA from developing mammalian embryos [8] , pointing towards a passive dilution process based on differences in the amount of mtDNA between human gametes . Here we test this hypothesis directly .
First we determined the proportion of paternal haplotypes transmitted to the offspring assuming no preferential destruction of sperm-derived mtDNA . We measured the amount of mtDNA in healthy human sperm and pre-fertilization oocytes on the same assay plate . This gave a mean ratio of 1:15 , 860 , in keeping with previous reports [8 , 21–23] . Based on these observations , the 95% prediction interval for the proportion of paternal haplotypes at fertilization is 10–5 to 1 . 8x10-4 . Next we developed an extreme-depth mtDNA re-sequencing assay to detect very low levels of paternal haplotypes . Given previous work showing a background level of ~1% heteroplasmy for single variants using deep mtDNA re-sequencing [24] , we set out to identify trios where the father and the child had two or more variant differences within a <200bp stretch of mtDNA , thus allowing the detection of extremely rare paternal haplotypes at a much lower heteroplasmy level captured within the same sequencing read . Sanger sequencing of the mtDNA from 99 mother-father-child trios from the Avon Longitudinal Survey of Parents and Children identified four different trios with >2 discordant alleles ( subsequently referred to as discordant variants , S1 & S2 Tables , Fig 1A & S1 Fig ) . Parent-offspring trios were confirmed with >99 . 9% accuracy using 16 short microsatellites in all members of each trio . Next we designed PCR amplicons spanning the four discordant regions . Each amplicon was >2Kb to eliminate nuclear pseudogene amplification , and BLAST analysis predicted exclusive annealing to mtDNA ( Fig 1A ) , which was confirmed by the failure to generate a template from rho-0 cells devoid of mtDNA . All four amplicons were amplified in all four trios using ultra-high fidelity polymerase ( PrimeSTAR GXL , TaKaRa Bio Europe , fidelity = 1 error/~16 , 230 base pairs ) , allowing each trio to act as a control for the others . Extreme high-depth re-sequencing ( Illumina MiSeq , 250bp paired-end reads ) yielded stable coverage spanning the region of interest in the offspring ( Fig 1B ) . We then compared the frequency of minor alleles and haplotypes both within and between the four trios . The frequency of isolated single variants was similar to that observed previously at lower depth at ~ 0 . 5% [24] . As expected , the number of minor haplotypes containing two or three variants was substantially less ( Table 1 ) . Overall , the frequency of minor haplotypes containing two variants approximated the square root of the single variant frequency , and the frequency of the three-allele haplotypes was approximately equal to the cube root of the single variant frequency . These observations were in keeping with a random background mutation frequency affecting single base-pairs of ~0 . 5% [24] . The observed alleles contributing to the ultra-rare haplotypes did not correspond to commonly co-occurring population variants , making external sample contamination unlikely ( Fig 1D and S2 Table ) . The frequency of unexpected maternal or paternal haplotypes was greater in other members of the same family trio than in other trios . Given that all of the trios were analysed in the laboratory simultaneously , these rare shared haplotypes probably arose through very low-level contamination at the time the original samples were taken , in the order of <10–5 molecules ( Table 1 ) . We incorporated these observations in a significance test of our findings , and determined whether we had adequate statistical power to detect paternally inherited mtDNA . The power to detect paternally inherited mtDNA was estimated by determining the difference between mismatched haplotype frequencies in mothers and children . These are simply modelled by Poisson distributions with a without paternal contributions , and the difference is described by a Skellam distribution [25] . The power was investigated for a range of theoretical paternal mtDNA contributions , and the observed background heteroplasmy levels in mothers ( Fig 2 ) . For the observed background heteroplasmy levels in the mothers ( 4 . 3 x 10–5 , 6 . 0 x 10–5 , 5 . 7 x 10–5 , and 7 . 6 x 10–5 , Table 1 ) , sequencing at >300 , 000-fold depth in the relevant offspring had >80% power to detect the predicted level of transmitted paternal mtDNA . A bootstrap test was used to determine whether the observed discordant haplotypes in the children were consistent with the predicted paternal contribution , incorporating the 95% confidence intervals for proportion of sperm haplotypes in the oocytes ( 10–5 to 1 . 8x10-4 ) . A paternal contribution was not compatible for trios A and B ( p = 0 . 004 , p = 0 . 016 ) , but there was insufficient evidence to reject the other two trios ( p = 0 . 21 , p = 0 . 12 ) that had a lower total coverage for the full range of possible transmitted levels of sperm mtDNA . The power calculation suggests that the uncertainty in two trios likely reflects the confidence interval for our estimate of sperm haplotypes in the fertilized oocyte , and not directly reflect the likelihood of any paternal transmission of mtDNA . The full data set and relevant code is available from the authors and at: https://www . staff . ncl . ac . uk/i . j . wilson/PaternalTransmission
When taken together these findings indicate that the extremely rare variant haplotypes seen in the offspring are highly unlikely to have arisen through the passive ‘leakage’ of paternal mtDNA within the trio . Lower levels of paternal mtDNA transmission are also unlikely because we did not see common mtDNA population haplotypes in any of the individuals studied ( Fig 1D ) , which would be expected if there were frequent paternal leakage in the human population . The lack of common population haplotypes at extreme high depth also shows that our experimental approach was not subject to significant cross-contamination . Finally , we show that there is little to be gained by sequencing at >5000-fold if single variants are of interest , because the ‘noise’ level will prevent further resolution of low-level heteroplasmy . The ‘background noise’ level that we observed has several potential origins . First , the near exponential decrease in frequency of single variants , two variant , and three variant haplotypes is consistent with a random background mutation frequency introduced by DNA replication errors , either within the biological system or through PCR amplification . We did not observe patterns of nucleotide changes that would imply a direct insult to the DNA templates ( such as C>U deamination damage ) . Finally , even at extreme high depth ( >1 million reads in some trios ) we saw negligible or no haplotypes observed in other trios . Given that the trios were analysed together , this means that our laboratory procedures were robust with negligible cross contamination . It is therefore likely that the rare paternal haplotypes seen in maternal samples ( Table 1 ) were introduced when the samples were collected . Our observations were made on DNA extracted from buccal swabs . It is thus theoretically possible that paternal mtDNA was originally transmitted to the offspring and subsequently lost from buccal cells before the DNA samples were acquired . The loss of mutated mtDNA has been observed in patients with mtDNA diseases , most typically m . 3243A>G , probably through selection against a deleterious allele at the stem cell level [26] . However , for most inherited heteroplasmic mtDNA mutations ( eg m . 8993T>C ) , the percentage level of the mutation is the same in a wide range of different tissues [27] , including buccal swabs . It is therefore likely that if there were paternal transmission , this would be detected in all tissues . Moreover , since variants we studied here are haplogroup markers and are unlikely to have significant biochemical consequences [28] , it is highly unlikely that a paternal haplotype allele would be lost from a particular tissue through selection . Absolute reassurance would be provided by studying other tissues , but this approach has its own difficulties because somatic point mutations accumulate in non-dividing tissues , increasing the background ‘noise’ levels to a point that would preclude the detection of passively transmitted paternal mtDNA . Given that human paternal mtDNA can be detected in very early embryos [4] , what is the mechanism underpinning exclusive maternal inheritance of mtDNA in humans ? Recent observations that ubiquitination and P63/LC3 tagging of sperm mitochondria does not lead to mitophagy in mice , the mechanism is likely to act at the level of the mitochondrial genome [8] . This is plausible , given recent observations in heteroplasmic mice where a similar mechanism was proposed for heteroplasmy segregation during inheritance [29 , 30] , and within different tissues and organs during life [31] , where selection against specific mtDNA molecules occurs at levels well below the level required to cause a biochemical defect affecting oxidative phosphorylation within the cell . A similar mechanism was proposed following the introduction of sperm mtDNA into somatic cells [15] . Although it remains to be determined how this selection process occurs at the molecular level , it is of fundamental importance in preventing the accumulation of deleterious mutations in the human population , effectively taking Muller’s ratchet [32] ‘down a gear’ .
Ethical approval for the study was obtained from the ALSPAC Ethics and Law Committee and the Local Research Ethics Committees . The ALSPAC study web site contains details of all the data that is available on the cohort through a fully searchable dictionary ( http:www . bris . ac . uk/alspac/researchers/data-access/data-dictionary/ ) . Excess oocytes were collected after follicular reduction from healthy donors , lysed for 16 hours in 50mM Tris-HCl , pH 8 . 5 , with 0 . 5% Tween 20 and 200ng/ml proteinase K , at 55°C , followed by heat inactivation at 95°C for 10 minutes . Three single oocytes and 43 sperm DNA samples were analysed using quantitative real-time PCR ( qPCR ) . A multiplex qPCR assay , using probes and primers targeting a region of MT-ND1 and the nuclear housekeeping gene β-2 microglobulin ( β2M ) , were used to measure mtDNA copy number on a CFX96 Touch™ Real-Time PCR Detection System ( BioRad , USA ) . Buccal DNA samples were extracted in a different laboratory from 100 mother-father-child trios obtained from the Avon Longitudinal Study of Parents and Children Cohort ( ALSPAC ) [33 , 34] . Mitochondrial DNA ( mtDNA ) haplogroup defining single nucleotide polymorphisms ( SNPs , www . phylotree . org ) were determined in each mother and father by Sanger sequencing specific regions of the mitochondrial DNA genome . ( Big Dye v3 . 1 kit and capillary electrophoresed on an ABI3130xl Genetic Analyzer , Life Technologies , Warrington , UK ) . Alignment and variant calling was performed using SeqScape software ( v2 . 1 . 1 , Applied Biosystems ) reference to the GenBank sequence NC_012920 . 1 . Four trios were identified with >2 discordant variants falling within a ~250 bp read-length ( S1 Table ) . Parent-offspring trios were confirmed using the Promega Powerplex 16 HS system ( Promega , Southampton , UK , performed by NorthGene Ltd . ) . 2x 2Kb mtDNA amplicons were designed to span the discordant regions in each trio ( Fig 1A ) and to avoid nuclear pseudogene co-amplification with a high-fidelity polymerase ( PrimeSTAR GXL DNA Polymerase , TaKaRa Bio Europe , France ) . Primer sequences were: set 1 = TATCCAGTGAACCACTATCAC-F ( m . 11010-11030 ) and GGGAGGTTGAAGTGAGAGG-R ( m . 13453-13435 ) ; set 2 = ATTCATCGACCTCCCCACC-F ( m . 14797-14815 ) and CTGGTTAGGCTGGTGTTAGG-R ( m . 389-370 ) . Initially , primer efficiency and specificity was assessed as successful after no amplification of DNA from rho0 cell lines , which contain no mtDNA . Amplified products were assessed by gel electrophoresis against DNA+ve and DNA-ve controls , and quantified using a Qubit 2 . 0 fluorimeter ( Life Technologies , Paisley , UK ) . Each amplicon was individually purified using Agencourt AMPure XP beads ( Beckman-Coulter , USA ) , pooled in equimolar concentrations and re-quantified . Amplicons were ‘tagmented’ , amplified , cleaned , normalised and pooled into a multiplex using the Illumina Nextera XT DNA sample preparation kit ( Illumina , USA ) . Multiplex plex pools were sequenced using MiSeq Reagent Kit v3 . 0 ( Illumina , USA ) in paired-end , 250 bp reads on the same flow cell . Post-run FASTQ files were analysed using an in-house developed bioinformatic pipeline . Reads were aligned to the rCRS ( NC_012920 ) using BWA v0 . 7 . 10 [35] , invoking—mem [35] . Aligned reads were sorted and indexed using Samtools v0 . 1 . 18 [36] . Variant calling was performed in tandem using VarScan v2 . 3 . 13 [37 , 38] , ( minimum depth = 1500 , supporting reads = 10 and variant threshold = 1 . 0% ) and LoFreq v0 . 6 . 1 [39] . Concordance calling between VarScan and LoFreq was >99 . 5% . Concordant variants were annotated using ANNOVAR v529 [40] . In-house Perl scripts were used to extract base/read quality data and coverage data . Potential paternal haplotypes were identified from the pool of analysed reads ( * . sam files ) using command line scripting , generating motif-specific counts for each trio . For example , [grep—c 'CCTCACTGCCCAAGAACTATCAAACTCCTGAGCCAACAACTTAATATGACTAGCTTACACAATAGCTTTTATAGTAAAGATACCTCTTTACGGACTCCACTTATGACTCCCTAAAGCCCATGTCGAAGCCCCCATCGCTGGGTCAATAGTACTTGCCGCAGTACTCTTG\|CAAGAGTACTGCGGCAAGTACTATTGACCCAGCGATGGGGGCTTCGACATGGGCTTTAGGGAGTCATAAGTGGAGTCCGTAAAGAGGTATCTTTACTATAAAAGCTATTGTGTAAGCTAGTCATATTAAGTTGTTGGCTCAGGAGTTTGATAGTTCTTGGGCAGTGAGG' . /* . sam] was used to identify whole reads containing m . 11299C and m . 11467G in forward and complement orientations ( corresponding bases underlined ) . Counts were generated for all possible permutations of motifs , in all available samples . The mtDNA counts for healthy human sperm had mean 77 . 2 ( SD 53 . 9 , n = 43 ) and oocytes had mean 1220000 ( SD 183000 , n = 3 ) . An empirical distribution for the ratio of sperm to oocyte mtDNA levels immediately after fertilization was estimated by bootstrapping . 100 , 000 individual mtDNA counts were sampled from the raw sperm mtDNA counts and 105 corresponding oocyte mtDNA counts were drawn from a normal distribution with mean and standard deviation as above for the oocytes . A 95% prediction interval was then calculated from the 2 . 5th and 97 . 5th percentiles of this ratio of sperm mtDNA to oocyte count . A hypothesis test for paternal transmission was performed by bootstrapping under the null hypothesis that the discordant haplotypes were due to paternal inheritance at the ratio predicted from the direct measurements on individual gametes . For each of the four discordant haplotypes , 105 sampled ratios were simulated in the same way as for the prediction interval , and from this 105 replicate discordant child haplotypes were generated by binomial sampling , with p equal to these ratios and n equal to the total coverage over all haplotypes for this motif and trio . Observed background noise was added to this count at a Poisson rate equal to the discordant maternal haplotype rate . The simulated maternal count was pure Poisson noise , generated at the same rate as the child noise . The bootstrapped differences in proportion were then compared to the observed data , and extreme values are evidence to reject the null hypothesis . The p-value is estimated percentile rank of the observed data in the bootstrapped distribution . | Emerging evidence raises the possibility that human mitochondrial DNA ( mtDNA ) is not strictly maternally inherited , but it has not been technically possible to test this hypothesis directly . We identified trios with discordant mtDNA haplotypes , parent-offspring trios were validated using polymorphic microsatellites , and then used extreme-high depth mtDNA re-sequencing to look for paternally transmitted mtDNA . Despite having up to ~1 . 2 million-fold coverage of mtDNA , we find no evidence that paternal mtDNA haplotypes are transmitted to offspring in humans . Our findings exclude a simple dilution mechanism for uniparental transmission of mtDNA present in all healthy individuals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | Extreme-Depth Re-sequencing of Mitochondrial DNA Finds No Evidence of Paternal Transmission in Humans |
As cells proceed along their developmental pathways they make a series of sequential cell fate decisions . Each of those decisions needs to be made in a robust manner so there is no ambiguity in the state of the cell as it proceeds to the next stage . Here we examine the decision made by the Drosophila R7 precursor cell to become a photoreceptor and ask how the robustness of that decision is achieved . The transcription factor Tramtrack ( Ttk ) inhibits photoreceptor assignment , and previous studies found that the RTK-induced degradation of Ttk was critically required for R7 specification . Here we find that the transcription factor Deadpan ( Dpn ) is also required; it is needed to silence ttk transcription , and only when Ttk protein degradation and transcriptional silencing occur together is the photoreceptor fate robustly achieved . Dpn expression needs to be tightly restricted to R7 precursors , and we describe the role played by Ttk in repressing dpn transcription . Thus , Dpn and Ttk act as mutually repressive transcription factors , with Dpn acting to ensure that Ttk is effectively removed from R7 , and Ttk acting to prevent Dpn expression in other cells . Furthermore , we find that N activity is required to promote dpn transcription , and only in R7 precursors does the removal of Ttk coincide with high N activity , and only in this cell does Dpn expression result .
Development proceeds in a stepwise manner in which cells pass through a series of interim states in the progression from a pluripotent precursor to a fully-specified and final cell fate . Positional signals play a prominent role in this mechanism , with the receipt of the signals promoting the cells to the next state in the series . Each interim state needs to be established robustly so that upon receipt of the subsequent round of signals , the advance to the next state is unambiguous . In this paper we use the R7 photoreceptor of the developing Drosophila eye as a model cell with which to study the molecular mechanisms that occur in the acquisition of interim cell states . Specifically we examine the choice made by the R7 precursor to become a photoreceptor and the mechanisms that cooperate and consolidate that choice . The R7 precursor belongs to a group of cells known as the R7 equivalence group , which gives rise ( in addition to R7 ) to the two R1/6 photoreceptors and the four ( non-photoreceptor ) cone cells . The fates of the equivalence group cells are determined by the positions they occupy in the growing ommatidial cluster ( Fig 1A ) , and each initially expresses the transcription factor Tramtrack ( Ttk88—hereafter referred to as Ttk ) which represses the photoreceptor fate . If a cell eliminates Ttk then it becomes a photoreceptor; if not it becomes a cone cell [1 , 2] . Removal of Ttk is achieved by the activation of the RTK pathway which induces the transcription of phyllopod ( phyl ) [3 , 4] , a gene encoding an adaptor protein that promotes the Ubiquitination and degradation of Ttk [5] ( Fig 1G ) . The degradation of Ttk in R1/6 precursors is achieved in a relatively simple manner; activation of the RTK pathway by the Drosophila EGF Receptor ( DER ) removes Ttk ( Fig 1G ) . But the process in the R7 precursor is more complicated because this cell also experiences high N activity which antagonizes the RTK function ( Fig 1H ) , and a much stronger RTK signal is needed in this cell for Ttk removal . This is supplied by the Sevenless ( Sev ) RTK [7] ( Fig 1I ) which is expressed at high levels in R7 precursors , and is activated by its ligand Bride of Sevenless ( Boss ) presented on the adjacent R8 cell [8] . Previously , it was assumed that the Sev-boosted RTK signal simply overcame the N antagonism and promoted the levels of Phyl required for Ttk degradation . In this paper we find that , in addition to the expression of Phyl , a concurrent suppression of ttk transcription is also required for R7 photoreceptor specification , and this is achieved by the action of the Deadpan ( Dpn ) protein . Dpn is a transcription factor of the helix-loop-helix class [9] , closely related to the E ( Spl ) proteins , and behaves as a direct N response gene in neuroblasts [10] and in the eye [11] . Dpn is expressed in R7 precursors , but not in any other cells of the R7 equivalence group [12] ( Fig 1B ) . This feature led us to investigate the roles it plays in the specification of this cell . In this work , we not only determine that Dpn is required to repress ttk transcription specifically in R7 precursors , but we also establish that ectopic Dpn disturbs the fates of the other cells of the R7 equivalence group , and we infer that Dpn expression needs to be tightly restricted to the R7 precursor . This restriction is established by a two-tier regulation of its gene transcription . First , Ttk represses dpn transcription ( in the nascent R7 precursor Ttk levels are high and dpn transcription is repressed ) . However , as the action of the RTK pathway degrades Ttk , that repression is lost . Second , high N activity is required to promote dpn transcription , and the only cell of the R7 equivalence group that degrades Ttk and has high N activity is R7 . Dpn and Ttk act here as mutually repressive transcription factors ( Fig 1C ) , and this relationship lies at the heart of the mechanisms used to silence ttk transcription and to restrict Dpn expression to the R7 precursor . In the nascent R7 , Ttk is dominant in the relationship with active transcription and high protein levels , while dpn transcription is repressed and no protein is present ( Fig 1D ) . But once the RTK pathway becomes active this relationship inverts . First Ttk protein is degraded , removing the repression on dpn transcription , and the high N activity then drives dpn gene expression with the resulting Dpn protein acting to inhibit ttk transcription . Thus at the end of the mechanism , Dpn is dominant; with active transcription and high protein levels , whilst ttk transcription is silenced and no protein is present ( Fig 1E ) . Dpn action is required in R7 precursors because the high N activity opposes Ttk degradation . Even with the use of Sev to provide a potent RTK signal in the cell , the degradation mechanism is not able to robustly specify R7 precursors; Dpn repression of ttk transcription is also required . The degradation mechanism removes extant protein , and the transcriptional regulation prevents protein resupply , and the two combine to specify the photoreceptor fate in the R7 precursor .
To determine the roles Dpn plays in R7 specification , we undertook an investigation of its loss-of-function phenotypes . When Dpn is ectopically expressed in R1/6 or cone cell precursors , these cells are re-specified as R7 types ( Fig 4A–4C ) , which highlights the need to restrict Dpn expression in the R7 equivalence group to the R7 precursor . We therefore investigated the mechanisms that restricts Dpn expression , and began with a fine-detailed analysis of its expression pattern . Dpn is first expressed in unidentified cells of very early ommatidial clusters , and then becomes expressed in the R3/4 precluster cells ( top bracket Fig 5A' ) . As the second-wave-cells are added to the cluster , the expression decays in R3/4 , and Dpn is then expressed in R7 precursors ( lower bracket Fig 5A' ) . This R7 expression is striking in that it is early and ephemeral , and is lost from the cell before any neuronal or R7-specific markers are expressed . This suggested that Dpn expression is likely controlled by the signals that act on the early R7 precursor , and the RTK and N pathways were the most likely candidates . In this section we examine how the RTK pathway and the removal of Ttk control Dpn expression , and in later sections we address the role played by N . When we examined Dpn expression in GMR-ttkRNAi eye discs we saw the normal expression pattern ( including R7s ) but additionally observed ectopic staining in the cone cell precursors in the more mature ommatidial clusters ( cc , Fig 6A ) . This argues that removal of ttk leads to ectopic expression of Dpn , and we next asked whether Ttk regulates dpn transcription . We began by examining the expression pattern of the dpn . lacZ transgene ( Fig 6B ) , and observed that expression began in the R3/4 and R7 precursors normally , but did not extinguish and persisted into the more mature ommatidial clusters . This reporter appears as a faithful indicator of when dpn transcription is initiated , but whether the later staining is indicative of normal dpn transcription ( which we doubt ) or whether it is artifactual ( perdurance of the lacZ gene product for example ) remains unclear . We introduced GMR-ttkRNAi into this background and observed the additional ß-galactosidase staining in cone cell precursors ( Fig 6C ) . The anterior ( AC ) and posterior ( PC ) cone cells are the first two cone cells recruited , and the presence of GMR-ttkRNAi re-specifies them both as R7s ( as evidenced by Runt staining—red ) . Interestingly , the ectopic ß-galactosidase staining occurred in PC and not AC ( Fig 6C' ) . In the next section of the paper we define a critical role for N function in dpn transcription , and note here that N activity terminates first in AC and we suspect that it is the low N activity in this cell that prevents dpn . lacZ expression , and we revisit and expand upon this in the Discussion . The gain of dpn . lacZ expression in cone cells argues that Ttk regulates Dpn expression at the level of transcription . Indeed , if Ttk functions to repress dpn transcription , then the forced expression of Ttk would be expected to prevent the expression of dpn . lacZ in R7 precursors . To this end , we engineered and transformed a UAS . ttk88 transgene and expressed it with GMR . Gal4 in the dpn . lacZ background ( Fig 6D ) . Here we observed the loss of ß-galactosidase staining from the R7 regions of the eye discs ( lower bracket ) but saw a normal expression at the R3/4 level ( upper bracket ) . Collectively , from the effects of loss and gain of Ttk activity , we infer that Ttk functions to repress dpn transcription . The results above suggest that Ttk degradation is required for the expression of Dpn . But if removal of Ttk were all that was needed for Dpn expression , then following Ttk removal we would expect chronic expression of Dpn in the R7 precursor . This does not happen; the Dpn expression in R7 precursors is short lived . Furthermore , Ttk is degraded in R1/6 precursors and these cells do not express Dpn . We therefore inferred the action of another agency , and since the N pathway is active in the R7 precursor , and ectopic N activation in the eye had already been shown to induce ectopic Dpn expression [11] , we began a detailed investigation of the role of N in Dpn regulation .
The Drosophila eye is an exquisitely structured photoreceptive organ that requires a precisely orchestrated developmental program to specify the appropriate cells in the correct position , and in this study we examine the mechanisms that ensures that an R7 photoreceptor is generated in every ommatidium . Within the R7 developmental program , the Boss ligand and Sev RTK are engaged to provide a requisite high-level activation of the RTK pathway which promotes the expression of Phyl and elicits the Ubiquitination and degradation of Ttk . Since Ttk is the transcription factor that represses the photoreceptor fate , the RTK pathway plays an indispensible role in specifying the R7 precursor as a photoreceptor . Indispensible as it is , expression of Phyl alone does not appear to be sufficient for the specification of all R7s , since abrogation of dpn gene function leads to ommatidia in which R7s are absent . We find here that Dpn functions to repress transcription from the ttk locus , and cooperates with the degradation mechanism in Ttk removal . We note that Dpn expression ceases early in the life of the R7 precursor and does not therefore function as a long-term mechanism to prevent the resupply of Ttk . Rather , we see it as acting coincidentally with the degradation mechanism , with both required to achieve effective Ttk elimination , and the resulting unambiguous specification of the R7 photoreceptor fate . Both the N and RTK pathways that converge in this mechanism are also short lived , and it remains unclear how the clearing of Ttk from the R7 precursor locks the cell into the photoreceptor pathway . Dpn is both expressed and required in the R7 precursor , and an elegant mechanism utilizing the combined action of the RTK and N pathways ensures that it is selectively expressed in this cell . First we will address the part played by Ttk and argue that it performs two different roles . The primary role of Ttk is to repress the photoreceptor fate ( through the regulation of unknown target genes ) . The secondary role is to prevent the expression of Dpn , and thereby prevent a reduction in its own transcription . This is the function that it plays in cone cell precursors , where in silencing Dpn it prevents the reduction of its own levels that would make it vulnerable to the low-level RTK activity in those cells . Cone cells and nascent R7s are identical in this regard , but as the high level RTK pathway operates in the R7 precursor , a dramatic change ensues resulting in the expression of Dpn and the shutting down of ttk transcription . However , in the cone cell precursors nothing changes , and Ttk persists in repressing Dpn . The other key aspect of Dpn regulation is the role played by N . Consider the R1/6 precursors; they degrade Ttk ( and thereby become photoreceptors ) but do not express Dpn . Hence , the simple removal of Ttk is insufficient for the expression of Dpn; high-level N activity is also required . N activity is low in R1/6 precursors but high in the R7 ( and cone cell ) precursors , where it provides the promotion of Dpn expression . Only in the R7 precursor are the requisite conditions for Dpn expression achieved; the repressor ( Ttk ) is removed , and the promoter ( high N activity ) is present . Mutually repressive transcription factor interactions normally resolve to one of two bi-stable states in which one dominates the other . In the R7 precursor we can see how a switch between two such states can be regulated by signals entering the cell . In the nascent R7s , Ttk levels are high and the dpn locus is silent . But this situation is altered once the RTK pathway begins Ttk degradation . As the level of Ttk drops below a certain threshold , dpn transcription initiates , and the resulting Dpn protein then feeds back to silence transcription from the ttk locus . Thus the RTK pathway acts to switch the Ttk/Dpn relationship from Ttk-dominant to Dpn-dominant , and in doing this , the clearance of Ttk protein required for photoreceptor specification is achieved . The Enhancer of Split genes are targets of the N pathway , and a number of them encode transcription factors of the helix-loop-helix class . These transcription factors mediate many N activities , but their expression patterns are poorly understood in the developing eye , and which ones mediate N functions in R7 specification has been difficult to determine . However , an enhancer element from the Mdelta gene is expressed selectively in the R7 precursor ( not in the R1/6 or cone cell precursors ) and ectopic expression of this gene in R1/6 precursors re-specifies them as R7 types ( albeit at a modest frequency ) [15 , 16] . Since Dpn is closely related to the E ( Spl ) family of proteins , we suggest that it likely has the same ability to re-specify R1/6 precursors as R7s . This begs the question of whether Dpn normally performs this function in R7 , or whether it mimics the activity of one or more of the E ( Spl ) proteins . We cannot currently answer this question , but we note that Dpn is expressed in R7 and not R1/6 and so has the spatial restriction needed to perform that function . We also note that dpn loss of function does not switch the fate of R7s to R1/6s ( although this may simply result from redundancy with other factors ) . Although we remain agnostic on the question of whether Dpn functions to specify the R7 photoreceptor type , the fact that ectopic Dpn expression in R1/6 precursors changes their fate , highlights the fact that are now two reasons why Dpn expression is exclusive to R7 precursors in second wave cells . First , as described above , if it were expressed in cone cell precursors , their lowered Ttk levels might lead them to be specified as photoreceptors . Second , we now see that although Dpn expression in R1/6 precursors would not disturb their specification as photoreceptors ( their Ttk is already degraded ) , it would disturb which type of photoreceptor they would be specified as . When Ttk activity is reduced in cone cell precursors ( GMRG4 . ttkRNAi ) , although both the anterior and posterior cone cells ( AC and PC ) were ( as expected ) re-specified as R7s ( as evidenced by Runt expression ) there was a striking asymmetry in the de-repression of dpn . lacZ activity; it occurred in PC and not AC . We infer that this relates to the different N activities in these two cells . Sev expression is a simple meter of N activity in the R7 equivalence group [6] , and we noticed many years ago that Sev staining decayed in AC before PC [17] . From this we infer that N levels fall first in this cell , and since dpn . lacZ critically requires high N activity for its expression , the failure to de-repress dpn . lacZ in AC simply results from the absence of the requisite high N activity . In addition to the roles played by Ttk and N in regulating Dpn , we also infer an additional ( Phyl-independent ) role played by the RTK pathway . Consider the expression of Phyl in a sev mutant background . In one experiment we show that this is not sufficient to restore all R7s ( Fig 2C ) , requiring the co-expression of Dpn ( Fig 2E ) for full rescue . Yet , in another experiment we show that expression of Phyl in a sev mutant background promotes that requisite Dpn expression ( Fig 5D ) . The simplest explanation here is that the RTK pathway promotes dpn transcription through two independent mechanisms , and only when both are active is a sufficient level of Dpn generated . One mechanism is the process we have described in which RTK activity promotes phyl expression that results in the removal of Ttk . The other mechanism is unknown , but one possibility is that the dpn gene is a direct target of the RTK pathway . In this scenario we envisage the RTK pathway to directly promote dpn transcription , and to indirectly de-repress its expression through promotion of phyl transcription . There is an enigma at the heart of N activity in the R7 precursor; it both promotes and opposes the removal of Ttk from the cell . This enigma likely reflects the evolutionary history of the ommatidium , and suggests that the role of N in opposing Ttk removal represents an ancient function designed to prevent the promiscuous specification of photoreceptors [18] , and that in repurposing the cell in the R7 position to become a photoreceptor , the high N activity was used to activate new mechanisms to counteract the primary function . We previously described the transcription of sev as one of these new mechanisms , and now recognize dpn transcription as another . Indeed , we now list four distinct N functions in the R7 precursor; the ancient role that opposes Ttk removal , the two functions that promote Ttk elimination ( sev and dpn transcription ) , and the role that specifies the R7 versus the R1/6 fate once Ttk is degraded . All four functions occur roughly simultaneously , highlighting the impressive , and counterintuitive ways that N is employed in R7 specification , and the R7 photoreceptor is emerging not just as a classic model for studying cell specification , but also becoming as an excellent system to probe a cell’s evolutionary history .
Protocols for adult eye sectioning and antibody staining have been described previously [6 , 19] . In all data analyses the R8 photoreceptors were individually identified to ensure that they were not mistaken as supernumerary R7s . Primary antibodies: rabbit anti-β galactosidase ( Cappel ) ; rabbit anti-GFP , mouse anti-GFP IgG2a ( Molecular Probes ) ; guinea pig anti-Runt ( gift of J . Reinitz , Stony Brook University , NY , USA ) ; rabbit anti-Runt ( gift of Andrea Brand , Gurdon Institute , Cambridge , UK ) ; guinea pig anti-Dpn ( gift of James Skeath , Washington University School of Medicine , Washington , USA ) ; mouse anti-Svp ( gift of Y . Hiromi , National Institute of Genetics , Japan ) ; rat anti-Elav , mouse anti-Cut ( Developmental Studies Hybridoma Bank ) . Alexa Fluor 488 , 555 , 647 conjugated secondary antibodies were used ( Molecular Probes ) . For amplification of the β-galactosidase staining , Donkey anti-rabbit HRP conjugated secondary antibodies were used . Amplification was conducted using Tyramide Signal Amplification ( TSA ) systems ( PerkinElmer ) following manufacturer’s protocol . sev[d2] , sev . N[act] , sev . sev[act] , Nts , sev . Su ( H ) enR , GMR . sev , sev . phyl , ttk . lacZ , sev . Gal4 , GMR . Gal4 , ( see [6 , 20] . dpn1 , UAS . dpn ( gifts of Sarah Bray ) , UAS ttkRNAi [TRiP stock HMS03008] , UAS dpnRNAi [TRiP stock HMC03154] ( Bloomington Stock Centre ) . UAS phylRNAi [GD12579 v35469] VDRC , >tub . Gal80> ( gift of Gary Struhl ) . Sev . Su ( H ) EnR clones were induced as previously described[6] . Clones expressing sevG4-dpn were induced in a hs . flip; >tub . Gal80>/dpn . lacZ; Sev . Gal4/UAS . dpn background by 37°C heat shocks given to larvae at 24-48hrs AEL . dpn1 clones were generated in w , hs . flip; FRT42D , dpn1/FRT42D , w+ , M ( 2 ) 53C flies by 37°C heat shocks delivered 24-72hrs AEL . Nts flies were reared at 18°C and shifted to 30°C in the third instar for 24 h followed by dissection and immunohistochemistry . The coding sequence for ttk88 was amplified from the cDNA provided by Craig Montell ( John Hopkins University School of Medicine ) using appropriate PCR primers and cloned into the attB UAS vector using KpnI-EcoRI restriction enzymes , and transformed using standard protocols . | Animals are made from a vast diversity of different cell types , and understanding how they are specified is a major goal of developmental biology . In this study we use the Drosophila R7 photoreceptor as a model system for understanding how cell fate specification occurs . We examine the step when the R7 precursor cell adopts the photoreceptor fate , and ask how the signaling pathways active in the cell are integrated to provide an unambiguous directive to become a photoreceptor . The transcription factor Tramtrack ( Ttk ) represses the ability of the cell to become a photoreceptor , and how it is removed is the focus of this study . Previous work identified a protein degradation mechanism , and here we describe the role of the transcription factor Deadpan ( Dpn ) in repressing ttk transcription . We find that both the protein degradation mechanism and transcriptional silencing are required for efficient Ttk removal . Dpn expression needs to be restricted to the R7 precursor and we describe how the mutual antagonism between Ttk and Dpn and the action of the Notch signaling pathway are integrated to ensure that Dpn is selectively expressed in the cell . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"gene",
"regulation",
"regulatory",
"proteins",
"precursor",
"cells",
"social",
"sciences",
"dna-binding",
"proteins",
"cloning",
"neuroscience",
"dna",
"transcription",
"gene",
"function",
"transcription",
"factors",
"molecular",
"biology",
"techniques",
"eyes",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"animal",
"cells",
"proteins",
"gene",
"expression",
"sensory",
"receptors",
"head",
"molecular",
"biology",
"biochemistry",
"signal",
"transduction",
"cellular",
"neuroscience",
"psychology",
"cell",
"staining",
"cell",
"biology",
"anatomy",
"neurons",
"genetics",
"photoreceptors",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"ocular",
"system",
"afferent",
"neurons",
"sensory",
"perception"
]
| 2016 | R7 Photoreceptor Specification in the Developing Drosophila Eye: The Role of the Transcription Factor Deadpan |
Inhibition of nitric oxide ( NO ) signaling may contribute to pathological activation of the vascular endothelium during severe malaria infection . Dimethylarginine dimethylaminohydrolase ( DDAH ) regulates endothelial NO synthesis by maintaining homeostasis between asymmetric dimethylarginine ( ADMA ) , an endogenous NO synthase ( NOS ) inhibitor , and arginine , the NOS substrate . We carried out a community-based case-control study of Gambian children to determine whether ADMA and arginine homeostasis is disrupted during severe or uncomplicated malaria infections . Circulating plasma levels of ADMA and arginine were determined at initial presentation and 28 days later . Plasma ADMA/arginine ratios were elevated in children with acute severe malaria compared to 28-day follow-up values and compared to children with uncomplicated malaria or healthy children ( p<0 . 0001 for each comparison ) . To test the hypothesis that DDAH1 is inactivated during Plasmodium infection , we examined DDAH1 in a mouse model of severe malaria . Plasmodium berghei ANKA infection inactivated hepatic DDAH1 via a post-transcriptional mechanism as evidenced by stable mRNA transcript number , decreased DDAH1 protein concentration , decreased enzyme activity , elevated tissue ADMA , elevated ADMA/arginine ratio in plasma , and decreased whole blood nitrite concentration . Loss of hepatic DDAH1 activity and disruption of ADMA/arginine homeostasis may contribute to severe malaria pathogenesis by inhibiting NO synthesis .
Current estimates of world-wide mortality due to malaria range from 367 , 000 to 755 , 000 deaths per year , mostly in African children [1 , 2] . Prompt treatment with parenteral artesunate improves survival in children with severe malaria but mortality remains high in those presenting with complications such as coma or acidosis [3] . Development of effective therapies for these patients will require improved understanding of the pathophysiology of severe malaria . Patients with severe malaria exhibit impaired endothelium-dependent vasodilation [4] and reduced nitrite and nitrate concentrations in plasma and urine [5] , indicating decreased nitric oxide synthesis . Impaired NO signalling has been implicated in microcirculatory dysfunction [6] , loss of blood-brain barrier integrity [7 , 8] and cytoadherence of infected erythrocytes to the vascular endothelium in mice [9] . Similar pathology has been directly observed in human malaria , but the importance of NO signalling in these processes is less certain [10–13] . NO production by nitric oxide synthase ( NOS ) is dependent in part on the relative bioavailability of arginine , the NOS substrate , and asymmetric dimethylarginine ( ADMA ) , an endogenous NOS inhibitor released during hydrolysis of proteins that have been methylated by protein arginine methyltransferase [14–16] ( Fig 1 ) . By inhibiting NOS , ADMA not only causes vasoconstriction , increased blood pressure , increased systemic vascular resistance and decreased forearm blood flow in vivo [17–19] , but also affects adhesion , inflammation , thrombosis , barrier integrity , motility , growth and repair in vitro [20–31]–endothelial functions that are relevant to the pathophysiology of malaria . Dimethylarginine dimethylaminohydrolase 1 ( DDAH1 ) metabolizes ADMA at a rate inversely proportional to arginine concentration [32] and thus stabilizes the ratio of ADMA to arginine when arginine levels vary [33 , 34] . In Gambian children , an intronic DDAH1 polymorphism is associated with susceptibility to severe malaria [35] , raising the possibility that DDAH1 might be functionally linked to disrupted ADMA/arginine homeostasis and impaired NO synthesis in severe malaria . In this study , we identify dysregulation of ADMA/arginine homeostasis in Gambian children with severe malaria and hypothesize that ADMA clearance is impaired by hepatic DDAH1 inactivation . To test this hypothesis , we infected mice with P . berghei ANKA and assessed changes in DDAH1 expression , protein levels and activity in hepatic tissue , a major site of ADMA metabolism [29 , 36–38] .
We determined ADMA and arginine concentrations in blood plasma obtained from Gambian children with severe or uncomplicated malaria at initial presentation and 28 days later . Healthy afebrile aparasitemic Gambian children served as an additional control group . Baseline characteristics of the study populations are presented in Table 1 . Plasma ADMA was lower in children with severe malaria ( median [IQR]: 0 . 40 [0 . 30–0 . 51] μmol/L ) or uncomplicated malaria ( 0 . 40 [0 . 33–0 . 47] μmol/L ) compared to healthy children ( 0 . 61 [0 . 56–0 . 69] μmol/L , p < 0 . 0001 vs . uncomplicated; p < 0 . 0001 vs . severe; Fig 2 and Table 2 ) . ADMA remained low at the 28-day follow-up visit for patients recovered from malaria . Plasma arginine was profoundly depleted in children with severe malaria compared to children with uncomplicated malaria or healthy children ( severe malaria: 31 . 7 [23 . 0–40 . 6] μmol/L; uncomplicated malaria: 45 . 0 [35 . 4–55 . 7] μmol/L , p < 0 . 0001 vs severe; healthy: 88 . 7 [79 . 3–102 . 5] μmol/L , p < 0 . 0001 vs severe; Fig 2 and Table 2 ) . By the 28-day follow up visit , plasma arginine concentration increased to 56 . 7 [42 . 1–78 . 9] μmol/L among children who recovered from severe malaria and to 70 . 8 [58 . 6–85 . 1] μmol/L among children who recovered from uncomplicated malaria ( p < 0 . 0001 vs acute , Fig 2 ) , but remained lower than the arginine concentration observed in healthy Gambian children ( 88 . 7 μmol/L; p < 0 . 0001 for either comparison ) . ADMA is a competitive inhibitor of NOS , and the ratio of ADMA to arginine determines NOS activity [39] . The ratio of ADMA to arginine was elevated among children with severe malaria ( 13 . 5 [11 . 2–17 . 1] ×10−3 ) compared to children with uncomplicated malaria ( 8 . 8 [7 . 2–11 . 6] ×10−3 , p < 0 . 0001 ) or healthy children ( 6 . 7 [5 . 8–8 . 2] ×10−3 , p < 0 . 0001 , Fig 2 ) . After recovery from severe malaria , the ADMA/arginine ratio returned to the level observed in healthy Gambian children ( recovered from severe malaria: 7 . 4 [5 . 9–10 . 1] ×10−3 , p < 0 . 0001 vs . acute; p = 0 . 25 vs . healthy children; Fig 2 and Table 2 ) . Thus elevation of the ADMA/arginine ratio appears to be an acute metabolic disturbance associated with a symptomatic episode of malaria ( modeled in S2 Fig ) . In healthy Gambian children , plasma ADMA concentration was correlated with plasma arginine concentration ( r = 0 . 43 , p < 0 . 05 ) ; children with lower arginine tended to have lower ADMA ( S1 Fig ) . In children with uncomplicated malaria , the correlation was stronger ( r = 0 . 59 , p < 0 . 0001 ) and the slope of the linear regression was steeper ( p < 0 . 0001 vs healthy; S1 Fig ) . In children with severe malaria , the correlation was stronger ( r = 0 . 77 , p < 0 . 0001 ) and the slope of the linear regression was steeper still ( p < 0 . 0001 vs uncomplicated; S1 Fig ) . At the day 28 follow up visit , the relationship between ADMA and arginine ( ie , the slopes of the linear regressions ) had returned to normal ( S1 Fig ) , though the absolute levels of ADMA and arginine remained lower than in health children . Lactate is a biomarker of impaired tissue perfusion that is associated with mortality from severe malaria in children [40–43] . In our study , lactate ( median [IQR] ) was elevated in children with severe malaria ( 5 . 0 [3 . 2–7 . 0] mmol/L ) compared to children with uncomplicated malaria ( 2 . 8 [2 . 2–4 . 1] mmol/L , p < 0 . 0001 , Table 1 ) . Lactate correlated positively with ADMA among children with severe malaria ( r = 0 . 34 , p = 0 . 004 , S3 Fig ) , implying that tissue perfusion was impaired among those with higher ADMA levels . Lactate did not correlate with arginine ( r = 0 . 16 , p = 0 . 20 , S3 Fig ) but did correlate with the ADMA/arginine ratio ( r = 0 . 28 , p = 0 . 02 , Table 3 ) . In multiple linear regression analysis using ADMA and arginine as explanatory variables , ADMA was positively related to lactate ( β = 0 . 758 , p = 0 . 002 ) while arginine was negatively and non-significantly related to lactate ( β = -0 . 393 , p = 0 . 09; Table 4 ) . We measured soluble vascular cell adhesion molecule ( sVCAM ) as a biomarker of endothelial activation . Plasma sVCAM was elevated in children with severe malaria at the time of admission compared to children with uncomplicated malaria or healthy children ( severe malaria admission: 1266 [828–1798] ng/mL; acute uncomplicated: 905 [726–1313] ng/mL , p < 0 . 05 vs severe , healthy children: 905 [773–1078] ng/mL , p < 0 . 05 vs severe , Table 1 ) . sVCAM returned to normal at day 28 among children who had severe malaria ( 841 [655–1147] ng/mL , p < 0 . 001 vs admission ) . In contrast , the sVCAM level of children with uncomplicated malaria was similar to the level measured in healthy children ( p = 0 . 84 ) . Endothelial activation appears to be a distinctive feature of acute severe malaria . Soluble VCAM was positively correlated with plasma ADMA in severe malaria patients ( r = 0 . 60 , p < 0 . 0001 , Table 3 and S3 Fig ) . This observation suggests that ADMA , a NOS inhibitor , is associated with endothelial activation and release of sVCAM into circulation . sVCAM was also positively correlated with arginine ( r = 0 . 59 , p < 0 . 0001 , Table 3 and S3 Fig ) . In multiple linear regression analysis , ADMA ( β = +0 . 413 , p = 0 . 02 ) was more significantly related to sVCAM levels than was arginine ( β = +0 . 332 , p = 0 . 06; Table 4 ) . Haptoglobin becomes depleted from plasma during acute intravascular hemolysis [44] . Haptoglobin was low at the time of admission in children with severe malaria or uncomplicated malaria ( severe: 0 [0–4 . 1] mg/dL; uncomplicated: 1 . 3 [0–49 . 3] mg/dL ) , and increased by the 28-day follow up visit ( severe day 28: 13 . 8 [1 . 2–44 . 3] md/dL , p < 0 . 0001 vs admission; uncomplicated day 28: 18 . 7 [0 . 2–59 . 4] mg/dL , p < 0 . 07 vs admission ) . Admission haptoglobin values , but not day 28 values , were significantly lower than in healthy children ( 44 . 5 [15 . 6–79 . 9] mg/dL ) . Because haptoglobin was undetectable in many children with severe malaria , we analyzed the correlation with ADMA and arginine using Spearman’s method . The correlation between haptoglobin and ADMA was weak ( r = -0 . 29 , p = 0 . 02 , Table 3 and S3 Fig ) , and weaker still with arginine ( r = -0 . 24 , p = 0 . 06 , Table 3 and S3 Fig ) . There were however , moderate negative correlations with hemoglobin , a measure of anemia that may be partially reflective of hemolysis ( correlation with Hb and ADMA: r = -0 . 44 , p <0 . 0001; Hb and Arg r = -0 . 32 , p = 0 . 004; Table 3 and S3 Fig ) . Multiple linear regression analysis again revealed that this correlation was primarily due to the association of hemoglobin with ADMA ( β = -3 . 039 , p = 0 . 003 ) and not with arginine ( β = +0 . 404 , p = 0 . 66; Table 4 ) . P . falciparum histidine-rich protein 2 ( PfHRP2 ) , a circulating marker of parasite biomass , was higher in children with severe malaria compared to uncomplicated malaria ( severe: 249 [133–605] ng/mL vs uncomplicated: 118[54–226] ng/mL , p = 0 . 0001 , Table 1 ) . However , PfHRP2 was not correlated with ADMA or arginine ( Table 3; S3 Fig ) . To determine whether P . berghei infection was associated with systemic changes in ADMA and arginine , we analyzed plasma from P . berghei ANKA-infected mice . Similar to Gambian children with malaria , plasma ADMA concentrations were lower in infected animals compared to uninfected controls on day 6 post-inoculation ( 0 . 44 [0 . 40–0 . 49] vs . 0 . 59 [0 . 54–0 . 63] μmol/L , p < 0 . 0001 , Fig 3A ) . Arginine concentrations were also lower in infected mice compared to uninfected controls ( 46 . 7 [39 . 2–53 . 3] vs . 78 . 1 [66 . 5–101 . 9] μmol/L , p<0 . 0001 , Fig 3B ) . Arginine decreased to a greater extent than ADMA , resulting in an increased ratio of ADMA to arginine among infected mice ( 9 . 83 [8 . 65–12 . 49] ×10−3 vs . 7 . 10 [5 . 94–8 . 81] ×10−3 , p<0 . 0001 , Fig 3C ) . The murine findings recapitulated our observations in Gambian children with malaria . Whole blood nitrite is reflective of NOS activity [45] . We determined nitrite concentrations in whole blood samples from P . berghei ANKA-infected mice and uninfected controls using a gas-phase chemiluminescent assay . Whole blood nitrite was decreased in infected mice ( 0 . 36 [0 . 28–0 . 45] μmol/L ) compared with uninfected controls ( 0 . 49 [0 . 41–0 . 71] μmol/L , p = 0 . 0001 , Fig 3D ) , suggesting that P berghei ANKA infection causes a decrease in systemic NO production in mice . DDAH1 is highly active in the liver and plays a key role in regulating circulating levels of ADMA [29 , 33 , 36–38] . To determine whether severe malaria affects hepatic DDAH1 function , we assessed hepatic Ddah1 gene expression and protein levels in liver tissue from C57BL/6 mice 6 days after inoculation with P . berghei ANKA . Using quantitative RT-PCR to assess hepatic Ddah1 mRNA transcript number relative to Gapdh , we found that Ddah1 gene expression was not changed by P . berghei ANKA infection ( median [IQR] fold change: 1 . 1 [1 . 0–1 . 1] , p = 0 . 07 , Fig 3E ) . In contrast , Western blot analysis revealed a decrease in hepatic DDAH1 protein from P . berghei ANKA-infected mice compared to uninfected control mice ( median [IQR] fold change: 0 . 33 [0 . 28–0 . 46] , p < 0 . 0001 , Fig 3F ) . These data demonstrate that P . berghei ANKA infection decreases DDAH1 protein abundance by a post-transcriptional mechanism . To determine the functional impact of hepatic DDAH1 inactivation by Plasmodium infection , we quantified ADMA clearance in liver homogenates by measuring the rate of de novo citrulline production in the presence of saturating concentrations of ADMA substrate ( assay validation presented in S5 Fig ) . Hepatic ADMA clearance was lower in mice infected with P . berghei ANKA compared with uninfected controls ( infected: 3 . 83 [3 . 22–4 . 19] nmol citrulline × mg protein-1 × hr-1 vs uninfected control: 6 . 48 [5 . 23–7 . 49] nmol citrulline × mg protein-1 × hr-1 , p < 0 . 0001 , Fig 3G ) . To assess the impact of P . berghei infection on hepatic ADMA metabolism , we determined intracellular ADMA concentrations in PBS-perfused liver samples from infected and control mice . ADMA was increased in liver tissue from infected mice compared with uninfected controls ( infected: 126 . 5 [88 . 9–198 . 2] pmol/mg protein vs uninfected control: 49 . 7 [35 . 3–77 . 8] pmol/mg protein , p < 0 . 0001 , Fig 3H ) . We calculated the correlation between hepatic DDAH activity and hepatic ADMA concentration in healthy mice and found a positive correlation ( r = 0 . 46 , p = 0 . 01 ) , i . e . , hepatic DDAH activities were greater in mice that had higher tissue levels of ADMA . This may reflect induction of DDAH activity in response to tissue levels of ADMA . Among P . berghei-infected mice , the correlation was similar ( r = 0 . 42 , p = 0 . 04 ) though the tissue levels of ADMA were higher , and the DDAH activities were lower than in uninfected mice ( S4 Fig and S1 Table ) . The partial correlation coefficient between hepatic DDAH activity and hepatic ADMA concentration , accounting for infection status , remained positive ( rpart = 0 . 34 , p = 0 . 01 ) . We also calculated the correlation between hepatic DDAH activity and plasma ADMA/Arginine ratio in mice , and found a negative trend , i . e . , mice with lower DDAH activity tended to have higher ADMA/Arginine ratios in plasma ( rpart = -0 . 23 , p = 0 . 09; S4 Fig and S1 Table ) .
We have analyzed ADMA and arginine concentrations in plasma from children with severe malaria to determine whether malaria is associated with disruption of ADMA/arginine homeostasis . We found that children with acute severe malaria have uncompensated hypoargininemia , i . e . , low arginine with an elevated ADMA/arginine ratio . The hypoargininemia persisted over the 28 days of follow-up , while the ratio of ADMA to arginine returned to normal . We interpret this as a transient inability to metabolize ADMA at a sufficient rate to compensate for low arginine during acute infection . Although plasma ADMA levels were below normal in patients with severe malaria , ADMA was positively correlated with lactate , a biomarker of severity , and sVCAM , a biomarker of endothelial activation , suggesting that higher ADMA levels are associated with adverse pathophysiologic changes . DDAH1 metabolizes ADMA , so we examined changes in DDAH1 activity in mice infected with a Plasmodium berghei ANKA , a model of severe malaria . P berghei infection caused inactivation of hepatic DDAH1 , accumulation of intracellular ADMA in liver tissue , elevation of the ADMA to arginine ratio in plasma , and decreased levels of nitrite in blood . Although these findings in the mouse model cannot be directly extrapolated to human malaria , it raises the possibility that the elevated ADMA/arginine ratio observed in children with severe malaria could be due in part to inactivation of DDAH1 . Our results extend upon a previous report of ADMA and arginine levels in Tanzanian children . Weinberg et al . observed ADMA/arginine ratios of 13 . 2 [11 . 1–16 . 4] ×10−3 in cerebral malaria , 12 . 3 [10 . 0–15 . 1] ×10−3 in non-cerebral severe malaria , 12 . 6 [10 . 7–15 . 1] ×10−3 in moderately severe malaria and 7 . 1 [5 . 8–9 . 0] ×10−3 in healthy children [46] . These values are consistent with the ADMA/arginine ratios observed in our study . We found the ADMA/arginine ratio to be significantly greater in children with severe malaria compared to children with uncomplicated malaria , while Weinberg et al found no differences among the ADMA/arginine ratios of cerebral malaria , non-cerebral severe malaria and moderately severe malaria groups . This discrepancy may be explained by the increased severity of the moderately severe malaria group in Weinberg , et al . that differed from our uncomplicated group by the inclusion of patients who could not tolerate oral medication [46] . As a result , these children may have had greater dietary insufficiency of arginine and arginine precursors than our group of uncomplicated malaria patients . Although the plasma arginine concentration was lower in Gambian children with severe malaria compared to the Tanzanian children with cerebral malaria , the rise from admission to day 28 in the Gambian children ( 31 . 7 to 56 . 7 umol/L , an increase of 25 umol/L ) was similar to the rise from admission to day 7 in the Tanzanian children ( 45 umol/l to 70 umol/L , an increase of 25 umol/L ) . In both Gambian and Tanzanian children with severe malaria , ADMA and the ADMA/Arg ratio were each correlated with lactate . This could be mediated through the vasoconstrictive or pro-adhesive effects of ADMA on vascular endothelium especially in the setting of hypoarginemia , with subsequent impairment of tissue perfusion leading to anaerobic glycolysis and lactate generation . Inter-individual differences in hepatic blood flow could also be responsible for the strong correlation between ADMA and lactate , since the clearance of each is dependent on hepatic perfusion . Impaired perfusion of liver tissue has been observed in adults with severe malaria [47] and could limit hepatic clearance of plasma ADMA [33 , 36–38] . In both Gambian and Tanzanian children with severe malaria , ADMA was correlated with biomarkers of endothelial activation ( sVCAM and Angiopoietin-2 , respectively ) . This could be through direct effects of ADMA on endothelial cells [48] or via the pro-adhesive effects of ADMA on circulating immune cells that interact with endothelium [49 , 50] . P . falciparum histidine-rich protein 2 ( PfHRP2 ) has been previously assessed as a quantitative marker of parasite biomass [51 , 52] . PfHRP2 did not correlate significantly with ADMA , arginine or the ADMA/arginine ratio among children with severe or uncomplicated malaria ( Table 3 ) . Our findings are in agreement with the prior study [46] and together suggest that in children host ADMA metabolism is not determined by parasite biomass . Arginine depletion and disruption of ADMA/arginine homeostasis have also been observed in adults with moderately severe and severe malaria [53] . In contrast to African children , Indonesian adults with severe malaria demonstrated elevated ADMA [53] . The apparent discrepancy in plasma ADMA in children and adults with severe malaria might be explained by differences in severe malaria pathophysiology observed in older versus younger patients [54] . Changes in plasma ADMA also differ between children and adults with acute sepsis; compared to age-matched healthy controls , ADMA was decreased in pediatric sepsis , but studies of adult sepsis found ADMA to be either unchanged or increased [55–58] . Disruption of ADMA/arginine homeostasis in children with severe malaria could be due to increased protein methylation , accelerated proteolysis of methylated proteins or impaired clearance of free ADMA . One might expect plasma ADMA to be directly elevated during a severe malaria infection , due to the combination of increased release of ADMA from erythrocytes undergoing hemolysis [59] and the impaired activity of hepatic DDAH that we present here . Instead , we observed lower plasma ADMA concentrations in both human and mouse malaria , consistent with prior measurements in children with malaria [46] . This could be due to increased uptake of ADMA from plasma into cellular compartments as has been observed in vitro after LPS , TNF or IL-1 stimulation [60] . In addition , plasma ADMA was strongly correlated with plasma arginine in our study , suggesting that arginine deficiency might lead to lower plasma ADMA . Mechanisms that could potentially link plasma ADMA to plasma arginine are inadequately understood , but might include the requirement for protein-incorporated arginine as the substrate for PRMTs that generate ADMA [16] , upregulation of the cationic transporters that allow both ADMA and arginine to cross cell membranes [61 , 62] , and negative feedback of arginine on DDAH activity [32] . DDAH1 is known to regulate ADMA/arginine homeostasis: heterozygous knock-out of DDAH1 in mice increased plasma ADMA , decreased NO-dependent vasodilation , and elevated blood pressure [63] . Conversely , transgenic over-expression of DDAH1 decreased plasma ADMA concentrations , increased urinary nitrites/nitrates and decreased blood pressure [64] . A second DDAH isoform ( DDAH2 ) has been identified [65] , but in contrast to DDAH1 , suppression of DDAH2 expression did not result in altered plasma ADMA concentrations [34] . The liver expresses DDAH1 [65] and metabolic tracer studies identified the liver as a major site for clearance of circulating ADMA [37] . Induction of DDAH1 expression in liver significantly lowered plasma ADMA [33] . Endothelial cell-specific knock-out of DDAH1 revealed hepatic DDAH1 expression not only in hepatic endothelial cells but also in hepatocytes [29 , 66] , which appear to be primarily responsible for systemic ADMA metabolism . Moreover , DDAH1 may regulate the release of ADMA from non-endothelial cell sources that affect local ADMA levels and vascular function . Patients with hepatic failure had elevated plasma levels of ADMA [38 , 67 , 68] that decreased after liver transplantation [67] . Conversely , patients with acute rejection of their liver graft had elevated ADMA compared to patients without episodes of rejection [67] . Taken together , results from human patients and animal studies implicate hepatic DDAH1 as a key regulator of circulating ADMA . In severe malaria , renal insufficiency [17] , in addition to the hepatic DDAH1 dysfunction we present here , could contribute to dysregulation of ADMA/arginine homeostasis . Plasmodium infection appears to accelerate the degradation of DDAH1 protein in hepatic tissue . Oxidative stress is a potential trigger of DDAH degradation that is present during malaria infection [69 , 70] . Overexpression of the p22phox subunit of NADPH oxidase in smooth muscle cells increased oxidative stress , decreased DDAH protein levels , decreased DDAH activity and caused accumulation of both intracellular and extracellular ADMA [71] . Treatment of p22phox-transfected smooth muscle cells with the proteasome inhibitor epoxomicin raised DDAH protein concentrations and reduced intracellular ADMA , demonstrating that oxidative modification of DDAH protein may target it for degradation by the proteasome . Plasmodium infection causes oxidative stress in liver tissue of mice [72] , which could be sufficient to accelerate the degradation of DDAH1 in liver endothelium . The liver may be exposed to reactive oxygen species generated by the increased populations of neutrophils and pigment-laden monocytes found in the hepatic vasculature during malaria infection [72 , 73] . Increased cell-free heme due to hemolysis promotes neutrophil infiltration and resulting liver damage [72] , raising the possibility that hemolysis may contribute to hepatic DDAH dysfunction . Hemolysis may also result in direct release of ADMA into circulation . Human erythrocytes contain total ( free plus protein-incorporated ) ADMA concentrations in the range of 47 . 85 ± 1 . 68 μmol/L , extrapolated from a concentration of 15 . 95 ± 0 . 56 μmol/L reported for hydrolysates of erythrocyte samples diluted 1:3 in water [59] . Rat erythrocytes contain a similar concentration of total ADMA , estimated to be 40 . 6 ± 7 . 2 μmol/L [74] . Following hemolysis , methylated erythrocyte proteins are exposed to proteases that disproportionately release ADMA relative to arginine [74 , 75] . Thus hemolysis could both increase ADMA release and inhibit ADMA clearance by promoting DDAH degradation . Elevation of ADMA relative to arginine favors NOS inhibition because ADMA is a competitive inhibitor of NOS [14] . ADMA also competes with arginine for cellular uptake , which could limit arginine availability for NO synthesis [61] . The intracellular concentration of ADMA in endothelial cells is approximately 10-fold higher than extracellular levels , reaching a concentration of 3–5 μmol/L which is near the Ki of eNOS [76 , 77] . Even small elevations in extracellular ADMA concentration to 2 umol/L had profound effects on brain NOS activity and gene expression profiles of endothelial cells in culture or in mice [78 , 79] . Impaired NO synthesis has been implicated in impaired vasoregulation , loss of blood-brain barrier integrity and cytoadherence of parasitized erythrocytes to the vascular endothelium during severe malaria . In a mouse model , treatment with an NO-donor improved cerebral microcirculation , reduced cerebral hemorrhages and prevented blood-brain barrier break-down [6 , 7] . NO synthase inhibition by ADMA downregulates tight junction protein expression [24] , which may explain the beneficial effect of NO on endothelial barrier integrity . Impaired NO synthesis is associated with increased adhesion molecule expression [80] and L-NAME ( a synthetic NOS inhibitor ) increased cytoadherence of parasitized red blood cells to vascular endothelial cells in vitro [9] . Thus , impaired NO signaling may contribute to microhemorrhage , vascular leak and sequestration of parasitized red blood cells observed in children with fatal cerebral malaria [13] . Taken together , these findings suggest that disruption of ADMA/arginine homeostasis could contribute to severe malaria pathogenesis by inhibiting NO synthesis . Therapeutic strategies that preserve or enhance DDAH activity during Plasmodium infection are needed to establish a causal relationship between DDAH degradation and disruption of ADMA/arginine homeostasis . While restoring ADMA/Arginine homeostasis might be necessary to improve endothelial NO synthesis , it might not be sufficient: impaired endothelial NO synthesis is likely to be limited by arginine deficiency and oxidation of tetrahydrobiopterin . NO that is produced will have a limited half-life due to reactions with cell free hemoglobin , superoxide , and other radicals that have increased abundance during malaria infection . In summary , through clinical observational studies of Gambian children and controlled experiments mice , we have identified hepatic DDAH dysfunction as a potential mechanism disturbing ADMA/Arginine homeostasis and limiting nitric oxide synthesis in severe malaria .
Patient enrollment and sample collection were conducted following ethical review and approval by the Gambian Government/MRC Joint Ethics Committee and the Ethics Committee of the London School of Hygiene & Tropical Medicine ( SCC 670 , SCC 1002 , SCC 1003 , SCC 1077 & SCC 1113 [healthy children] ) . Analysis of plasma samples and de-identified clinical data at the National Institutes of Health was exempted from further ethical review by the NIH Office of Subjects Research ( Exemption #5161 ) . Written informed consent was provided by a parent or guardian on behalf of children enrolled in the study . Families who declined to participate were provided standard medical care . Animal studies ( described in detail in the supplementary information , S1 Methods ) were specifically approved by the National Institutes of Allergy and Infectious Diseases ( NIAID ) Animal Care and Use Committee ( ACUC ) under the protocol identification LMVR 18E . The NIAID ACUC complies with the U . S . Government Principles for the Utilization and Care of Vertebrate Animals , the Public Health Service ( PHS ) Policy on Humane Care and Use of Laboratory Animals , and the Animal Welfare Act . Children with severe malaria or uncomplicated malaria were enrolled at health centers in a peri-urban area around Fajara , The Gambia as previously described [81] . Enrollment sites included the Royal Victoria Teaching Hospital , the Brikama Health Centre , the MRC Fajara Gate Clinic and the Jammeh Foundation for Peace Hospital in Serekunda . Acute uncomplicated malaria was defined as asexual P . falciparum parasitemia of >5000 parasites/μl detected by slide microscopy with an episode of fever ( temperature >37 . 5°C ) within the previous 48 hrs and the absence of severe criteria . Acute severe malaria was defined as parasitemia of >5000 parasites/μl , a history of fever and one or more of the following: severe anemia ( Hb < 6g/dl ) , severe acidosis ( serum lactate >7 mmol/L ) , cerebral malaria ( Blantyre coma score 2 or less in the absence of hypoglycemia or hypovolemia with the coma lasting for at least 2 hrs ) , and severe prostration ( inability to sit unsupported in children >6 months or inability to suck in children <6 months ) . Patients with severe malaria were admitted and treated with quinine , and patients with uncomplicated malaria were treated with chloroquine plus sulfadoxine-pyrimethamine according to Gambian Government Treatment Guidelines [82] . Ninety-six children with severe malaria and 102 children with uncomplicated malaria were enrolled during the 2005–2008 malaria seasons . Four milliliters of blood were collected in heparinized vacutainers ( BD ) at the time of initial presentation and at a follow-up visit 28 days later . Blood samples were immediately refrigerated , placed on ice for transport , and processed within 2 hours of collection . Plasma was frozen at -80°C . Three patients with severe malaria died and 65 completed follow-up visits; plasma sample were available from 47 of them . Eighty-five patients with uncomplicated malaria completed follow-up visits; plasma samples were available from 65 of them . Thirty-one healthy , afebrile , aparasitemic Gambian children of similar ages were also studied at a single visit . P . falciparum parasitemia was determined by bright-field microscopy of giemsa-stained blood smears . 50 fields were counted at high power . Full blood counts were obtained with an automated instrument ( Clinical Diagnostics solutions , Inc . , Fort Lauderdale , FL , USA ) . Lactate was measured with a handheld Lactate Pro device ( Arkray , Edina , MN , USA ) . Soluble VCAM ( sVCAM ) and plasma haptoglobin concentrations were determined by ELISA ( sVCAM: R&D Systems , Minneapolis , MN , USA; haptoglobin: Alpco , Salem , NH , USA ) according to the manufacturer’s instructions . P . falciparum HRP2 was measured in duplicate in plasma by ELISA ( Cellabs ) according to the manufacturer’s instructions . A standard curve was constructed using serial dilutions of the PfHRP2 standard and run with every plate . Laboratory staff were unaware of clinical status of the subjects . Each plasma sample was diluted in PBS containing NG-monoethyl-L-arginine ( MEA ) as an internal standard before undergoing solid-phase extraction ( Oasis MCX 96-well μElution Plate , Waters Corporation , Milford , MA , USA ) . The eluted cationic amino acids were dried , resuspended in water , and derivatized with ortho-phthalaldehyde ( OPA ) in 3-mercaptopropionic acid . Derivatized samples were separated by reverse-phase liquid chromatography over a 1×100 mm C18 ( 2 ) column ( Phenomenex , Torrance , CA , USA ) and fluorescence detected at excitation and emission wavelengths of 340nm and 455nm , respectively . Concentrations were determined by integrating peak area with reference to the internal standard ( MEA ) and daily external standards ( Arg , ADMA , and MEA ) . Data are expressed as median and inter-quartile range ( IQR ) . Statistical analyses were performed with GraphPad Prism 6 . 02 software and the R computing environment . P-values of less than 0 . 05 were considered significant . Mann-Whitney test was used to compare median values between healthy children and children with uncomplicated or severe malaria . Wilcoxon matched-pairs signed rank test was used to compare acute and recovery values in children with uncomplicated or severe malaria . Correlations were calculated on transformed data using Pearson’s correlation test , except in the case of correlation with haptoglobin in which Spearman’s method was used . When two separate groups were combined , partial correlations were calculated . Multiple linear regression analysis was used to assess the independent contributions of ADMA or arginine to correlations with hemoglobin , HRP2 , sVCAM , or lactate . In these linear models , arginine and ADMA were the explanatory variables , and the direction , strength and significance of the association was assessed by the beta value and p-value . Mann-Whitney test was used to compare values between infected mice and uninfected controls . Data are available in supplemental files S1 Data and S2 Data . | During a malaria infection , the vascular endothelium becomes more adhesive , permeable , and prone to trigger blood clotting . These changes help the parasite adhere to blood vessels , but endanger the host by obstructing blood flow through small vessels . Endothelial nitric oxide ( NO ) would normally counteract these pathological changes , but NO signalling is diminished malaria . NO synthesis is inhibited by asymmetric dimethylarginine ( ADMA ) , a methylated derivative of arginine that is released during normal protein turnover . We found the ratio of ADMA to arginine to be elevated in Gambian children with severe malaria , a metabolic disturbance known to inhibit NO synthesis . ADMA was associated with markers of endothelial activation and impaired tissue perfusion . In parallel experiments using mice , the enzyme responsible for metabolizing ADMA , dimethylarginine dimethylaminohydrolase ( DDAH ) , was inactivated after infection with a rodent malaria . Based on these studies , we propose that decreased metabolism of ADMA by DDAH might contribute to the elevated ADMA/arginine ratio observed during an acute episode of malaria . Strategies to preserve or increase DDAH activity might improve NO synthesis and help to prevent the vascular manifestations of severe malaria . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| []
| 2015 | Plasmodium Infection Is Associated with Impaired Hepatic Dimethylarginine Dimethylaminohydrolase Activity and Disruption of Nitric Oxide Synthase Inhibitor/Substrate Homeostasis |
Atrial fibrillation , a common cardiac arrhythmia , often progresses unfavourably: in patients with long-term atrial fibrillation , fibrillatory episodes are typically of increased duration and frequency of occurrence relative to healthy controls . This is due to electrical , structural , and contractile remodeling processes . We investigated mechanisms of how electrical and structural remodeling contribute to perpetuation of simulated atrial fibrillation , using a mathematical model of the human atrial action potential incorporated into an anatomically realistic three-dimensional structural model of the human atria . Electrical and structural remodeling both shortened the atrial wavelength - electrical remodeling primarily through a decrease in action potential duration , while structural remodeling primarily slowed conduction . The decrease in wavelength correlates with an increase in the average duration of atrial fibrillation/flutter episodes . The dependence of reentry duration on wavelength was the same for electrical vs . structural remodeling . However , the dynamics during atrial reentry varied between electrical , structural , and combined electrical and structural remodeling in several ways , including: ( i ) with structural remodeling there were more occurrences of fragmented wavefronts and hence more filaments than during electrical remodeling; ( ii ) dominant waves anchored around different anatomical obstacles in electrical vs . structural remodeling; ( iii ) dominant waves were often not anchored in combined electrical and structural remodeling . We conclude that , in simulated atrial fibrillation , the wavelength dependence of reentry duration is similar for electrical and structural remodeling , despite major differences in overall dynamics , including maximal number of filaments , wave fragmentation , restitution properties , and whether dominant waves are anchored to anatomical obstacles or spiralling freely .
Atrial fibrillation ( AF ) is a cardiac arrhythmia characterized by rapid and irregular atrial activation . Such desynchronized activation may occur when multiple waves circulate the atria . Unlike ventricular fibrillation , where unsynchronized activation of the ventricles ( the main pumping chambers of the heart ) causes an immediate and typically fatal loss of blood pressure , atrial fibrillation may be a repetitive , even chronic , disease . In fact , AF is the most common sustained cardiac arrhythmia in the United States and the rest of the developed world [1] , [2] , with more than 2 . 3 million sufferers in the U . S . [3] . AF becomes increasingly common with age [2] and is associated with significant mortality and morbidity , such as heart failure and stroke [1] . AF is more prominent in the context of alterations in atrial tissue properties – due to disease , arrhythmias , or age – known as remodeling . In fact , AF itself leads to remodeling , causing electrophysiological ( “electrical” ) , contractile , and structural changes [4] . Although AF can typically be reversed in its early stages , it becomes more difficult to eliminate over time due to such remodeling – hence the expression “AF begets AF” [5] . A central hypothesis for why AF begets AF is that electrical and structural remodeling due to chronic or persistent AF shorten the action potential wavelength , which measures the spatial extent of the action potential . Such wavelength shortening allows more waves to fit in the atria and maintain the arrhythmia [6] . Electrical remodeling primarily shortens the refractory period and the action potential duration ( APD ) of the atrial action potential , while structural remodeling impedes propagation and hence decreases conduction velocity ( CV ) . Since the wavelength is given as the product of APD and CV ( or , alternatively , the product of the effective refractory period and CV ) , electrical remodeling and structural remodeling both decrease the wavelength , thus potentially perpetuating AF . Additionally , the stability of reentrant waves may be affected by remodeling . Prior modeling work has shown that flattening APD restitution ( the dependence of APD on the previous resting interval or Diastolic Interval , DI ) , which typically occurs as a consequence of electrical remodelling [7] , may stabilize reentry [8] . Likewise , diffuse fibrosis , which may occur during structural remodelling [9] may stabilize reentrant waves [10] . Clinically , because electrical and structural remodeling typically present jointly in patients with chronic AF , their effects are difficult to separate . Animal models of primarily electrical remodeling ( due to rapid atrial pacing ) and predominantly structural remodeling ( induced heart failure or mitral regurgitation ) exist , however the rapid atrial pacing models also typically develop some degree of structural remodeling while the heart failure animals undergo some concomitant electrophysiological changes [9] , [11] . We therefore decided to use computer modeling as a means to investigate the mechanisms of how APD shortening due to electrical remodelling , and CV slowing due to structural remodelling , influence the duration and spatiotemporal dynamics of simulated AF in a computational multiscale model of human electrophysiological dynamics and substrates . Because structural aspects of the complex atrial anatomy are important for , e . g . , anchoring waves to anatomical obstacles and thus influencing the duration of reentrant activity , we use an anatomically detailed structural model of the human atria .
We model the atria using the Courtemanche et al . cellular model of human atrial cell electrophysiology [12] , with computational cells diffusively coupled to their nearest neighbors in an anatomically derived , three-dimensional structural model of the human atria [13] . As in previous work from our group [14] , we increase the conductance of the inward rectifier current , IK1 , here by 75% , in order to get the baseline action potential duration and resting membrane potential closer to experimentally observed values . Further , as in our previous work [14] , we fix the intracellular concentrations of K+ and Na+ ( at 139 . 0 mM and 11 . 2 mM , respectively ) to avoid long-term drift . The anatomical model incorporates heterogeneous coupling , resulting in different conduction velocities in different anatomical regions , in agreement with human data [13] , [15] . Specifically , the model exhibits fast conduction in Bachmann's bundle , the pectinate muscle network , the crista terminalis , and the limbus of the fossa ovalis ( 120 cm/s during sinus pacing ) ; slow conduction in the isthmus and the fossa ovalis ( 36 cm/s ) , while the remaining ( bulk ) atrial tissue has intermediate conduction velocity ( 65 cm/s ) . The different regions are shown in Fig . S1 in Text S1 ( online Supporting Information ) . The model does not include anisotropy . In human atria , the bulk atrial muscle has a more random fiber orientation than the fast conduction pathways ( and also more random than the ventricular myocardium ) , which have well-organized orientations along the bundles [13] . However , due to the strip-like anatomy of the fast tissues ( Fig . S1 in Text S1 ) , anisotropy is predicted to play a minor role there . This mathematical representation of the atria reproduces basic features such as depolarization time and spatial profile during normal ( sinus ) pacing [13] . The three-dimensional anatomical model is discretized in a 300×285×210 grid of spatial nodes , with a spacing of Δx = 0 . 025 cm , and no-flux boundary conditions . The equations were solved using an operator-splitting method [16] with forward Euler integration of both operators . We used a fixed time step of Δt = 0 . 01 ms for the partial differential equation describing the diffusion of voltage . The cellular model was integrated using an adaptive time step [16] . The code was parallelized using OpenMP , and run on multi-core machines . Simulating 60 s of reentrant activity took 5 days on a 24-core machine ( 2 . 66 GHz Intel® Xeon® X7460 processors , 128 GB memory ) . Because the simulations are this computationally costly , simulations were stopped when reentry terminated or at 60 s ( in which case , the arrhythmia was classified as sustained ) , whichever came first . Electrical remodeling due to chronic AF was simulated as in previous work , incorporating a 70% decrease in the conductance of the L-type calcium current ( ICaL ) , a 50% decrease in the conductance of the transient outward current ( Ito ) , and a 50% decrease in the conductance of the atrial-specific , ultra-rapid potassium current ( IKur ) [7] . These values are based on current recordings in cells isolated from human atrial appendages . We refer to this set of values as 100% , or full , electrical remodeling . In order to simulate different degrees of electrical remodeling ( 10–90% ) , in some simulations the percentage changes in the three affected conductances were downscaled by the same factor . We simulate structural remodeling by decreasing the diffusion coefficients ( i . e . , the coupling strengths between computational cells ) , which reduces conduction velocity . The three different nominal diffusion constant values ( assigned to fast , bulk , and slow conducting tissue ) were scaled by the same factor . Maximal structural remodeling was set to a 50% decrease in diffusion , causing the time for full activation of the atria with sinus pacing to increase from 108 ms to 149 ms . However , as experimental and clinical data show a large range of conduction impairment with structural remodelling [9] , we investigate two more levels of structural remodeling , using downscaling in diffusion of 70% and 83% ( increasing activation time to 119 ms and 130 ms , respectively ) . Because the duration of reentrant episodes may depend on where the reentry is initiated , we simulated reentry initiated at three different locations: the left atrial free wall , the left atrium near the left pulmonary veins , and the right atrial free wall . At each of these locations , reentry was initiated using a cross-gradient protocol , using a stimulus current of 80 nA/µF for 1 ms . Because the vulnerable window for reentry initiation is very small in the non-remodeled virtual tissue , we applied a brief hyperpolarizing clamp ( −80 mV for 1 ms ) to the region of the second wave excitation , 30 ms after its initiation . This allows for earlier reentry into this region and increases the vulnerable window . The coupling interval ( i . e . , the time between the first and the second excitation ) was varied systematically between simulations in steps of 10 ms within the vulnerable window . The size of the vulnerable window varies with variation in electrical and structural remodeling parameters , but was in the range of 30–60 ms , such that 4–7 reentry simulations were initiated at each location . Applied to three different locations , this means that for a given set of parameters describing the degree of electrical and/or structural remodeling , 12–21 simulations were run with different initiations . DI and APD were recorded from 16 different locations , spread evenly throughout the atria ( see Fig . S2 in Text S1 ) . The APD was measured as the time from the crossing of −70 mV on the upstroke to the crossing of −70 mV during repolarization . Inversely , DI was measured as the time between the crossing of −70 mV during repolarization to the crossing of −70 mV on the next upstroke . The wavelength ( WL ) is difficult to measure accurately during reentrant activity , even in computational studies [17] , due to wave collisions and irregular wave propagation . We use a method similar to that employed by Graux et al . ( Ref . [18] ) and determine CV during periodic pacing in the left atrial free wall . However , rather than obtaining CV for very few pacing rates as necessitated in the clinic , we systematically varied the pacing rate to establish the dependence of the CV on diastolic interval . Such restitution curves were obtained for all the combinations of the different levels of electrical and structural remodeling simulated . We later used these restitution curves to estimate local CV for DI values measured during reentry and , finally , to compute the local wavelength as WL = APD×CV . Note that this definition of the wavelength is more practical than WL = ERP×CV , where ERP is the effective refractory period , since ERP measurement requires a series of stimuli and cannot be measured directly during simulated AF . In paced tissue simulations , we found that APD underestimates ERP by 9–16 ms ( 4–10% ) depending on pacing rate and level of remodeling . Hence , our calculations of the wavelength using APD are presumably 4–10% larger than estimates based on ERP . Atrial fibrillation and flutter can be characterized by the number of wavelets present in the tissue . As in our group's previous work [14] , we compute the location of wave tips ( filaments ) from the crossing of two isopotential curves ( the crossing of −30 mV on the upstroke ) , separated in time by 2 ms . The number of separate filaments is determined by applying a k-means clustering analysis to the filament location data . To characterize individual simulations we use the maximal number of filaments present in that run . Dominant waves were defined as waves existing for at least five rotations . Their locations were determined directly from the filament location data or ( in the case of anchored waves without filaments ) based on periodicities in the transmembrane potential from the 16 recording sites , as well as visual inspection of isopotential surface maps .
As shown previously , electrical remodeling leads to shortening of the APD [7] . In particular , in our simulations of tissue strands , full electrical remodeling reduces the APD at 1 Hz pacing from 228 ms to 135 ms , while at 5 Hz the APD is shortened from 134 ms to 103 ms . This reduction in the amount of APD shortening with faster pacing demonstrates the flatter APD restitution occurring with electrical remodeling ( see Fig . S3A in Text S1 ) . The CV is unchanged with electrical remodeling ( Fig . S3C in Text S1 ) . In contrast , structural remodeling decreases CV ( CV restitution slope remains largely unchanged; Fig . S3D in Text S1 ) , while the APD is unchanged ( Fig . S3B in Text S1 ) . As a measure of the effects of electrical and structural remodeling , we focus primarily on the duration of reentrant activity . In our three-dimensional model , in the absence of electrical and structural remodeling , reentrant activity is not sustained: in all simulations of normal tissue , with varying initiation time and location ( see Methods ) , reentry ended 1–3 seconds after initiation . This is consistent with clinical findings in the normal human atria , where AF episodes typically self-terminate soon after initiation . Fig . 1A shows an example of non-sustained reentrant activity in normal tissue . The wavelength in this case is sufficiently long that the reentrant wave eventually runs into refractory tissue and dies out . In contrast , when simulating full electrical plus structural remodeling , reentry was sustained for 60 s in 18 of 21 simulations . With such electrical plus structural remodeling the wavelength is much shorter than in normal tissue ( Fig . 1B ) , and the reentrant wave in the left atrial free wall does not self-terminate . Videos showing these dynamics with and without remodeling are available as Supporting Information ( Videos S1 and S2 ) . To investigate whether this maintained reentrant activity results from more waves being present versus those present being more stable , we measured the maximal number of filaments in a simulation , as well as the mean duration of reentrant activity , while varying the level of electrical and structural remodeling . Electrical and structural remodeling both lead to increases in the duration of reentrant activity , and the combination of electrical plus structural remodeling gives even longer sustained reentry ( Fig . 2A ) . Structural remodeling also causes an increased number of filaments , while the level of electrical remodeling does not exhibit a clear correlation with the maximal number of filaments ( Fig . 2B ) . Note that the number of filaments present at baseline is consistent with previous experimental and modeling studies [14] , [19] , [20] . To test the hypothesis that a smaller wavelength perpetuates AF , we determined the dependence of the duration of reentrant activity on the estimated wavelength . For both electrical and structural remodeling , a decrease in WL is associated with longer reentry duration when WL is below a threshold value of around 7 cm ( Fig . 3A ) . Interestingly , the dependence of reentry duration on WL is similar for both electrical and structural remodeling , suggesting that WL is a more important determinant of duration than other factors , such as APD restitution slope , that vary between electrical and structural remodeling . In contrast , the dependence of the maximal number of filaments on WL is very different for electrical vs . structural remodeling , with more filaments present during structural than electrical remodeling for the same WL ( Fig . 3B ) . This indicates that more conduction block and wavebreaks occur during structural remodeling , but might also be the result of fewer waves being anchored to anatomical obstacles , since an anchored wave does not necessarily have a filament . From analyzing the dynamics of the number of filaments , we found that with structural remodeling , frequent occurrences of conduction block and wavebreaks cause the larger maximal number of filaments relative to electrical remodeling . These wavebreaks often heal , such that the increase in filaments is transient ( Figs . S4 and S5 in Text S1 ) . Taken together , these results demonstrate that the wavelength determines the duration of simulated AF , despite differences in dynamics such as APD restitution , conduction block , and number of filaments . As mentioned above , in our simulations of periodically paced tissue strands , electrical remodeling shortens APD without changing CV , and structural remodeling decreases CV without affecting APD . If these dependencies hold during reentry in the three-dimensional atrial anatomy , then APD itself should be a marker for reentry duration during electrical remodeling , while CV should correlate with reentry duration during structural remodeling . Such markers might be valuable given the methodological difficulties in determining WL ( see Methods ) . However , during reentry in the anatomical model with simulated electrical remodeling , there is both a decrease in APD and a concomitant fall in CV ( Fig . 4A , C ) . For structural remodeling , there is a primary decrease in CV ( Fig . 4D ) and a secondary increase in APD ( Fig . 4B ) . These results show first of all that APD alone ( for electrical remodeling ) and CV alone ( for structural remodeling ) are not accurate surrogates for WL during reentry . A similar finding was reported for AF/AFL inductance in a canine model [20] . Importantly , the secondary changes also suggest that the dynamically induced differences in the wave characteristics ( APD and CV ) may be due to differences in preferred pathways during reentry in the anatomical model , since differences in pathway lengths would affect the size of the excitable gap , and hence cause dynamical changes in DI , APD , and CV . Our different reentry initiation protocols allow a range of different spatiotemporal dynamics to occur . In all simulations that ran the full 60 s , the reentrant activity settled into a relatively periodic rhythm , with dominant waves remaining in a particular location ( often circulating an anatomical obstacle ) . However , as shown in Table 1 , the location of the dominant wave ( s ) varied significantly . In general , with only electrical remodeling dominant waves tended to be located in the left atrium , while dominant waves were found in the right atrium when we simulated structural remodeling only . With combined electrical and structural remodeling , dominant waves were in either or both atria . More specifically , for electrical remodeling , 4 of 6 simulations resulted in reentry around the pulmonary veins ( Table 1 ) . Fig . 5A shows an example of such dynamics ( see also Video S3 in the Supporting Information ) . A wave is anchored to the left pulmonary veins during the entire rotation ( Fig . 5A , left ) . Another wave front is circulating the right pulmonary veins , but does not remain completely anchored for the entire rotation ( Fig . 5A , right ) . Excitation spreads from the left atrium to the right . During structural remodeling , all simulations of sustained reentry result in dominant waves rotating around the tricuspid annulus and the inferior vena cava ( Table 1 ) . In most cases ( 5 of 7 ) , the rotation is counter-clockwise around the tricuspid annulus . An example is shown in Fig . 5B ( and Video S4 ) , with the wave anchored to the tricuspid annulus on the left , and the wave circulating the inferior vena cava on the right . With electrical plus structural remodeling , the outcome is more varied . In some cases waves are anchored ( to the pulmonary veins , the superior vena cava , or the tricuspid annulus ) . However , in some simulations , the dominant waves are un-anchored but spiral around the left or the right atrial free walls ( Table 1 ) . An example of such a scroll wave is shown in Fig . 5C , left ( and Video S5 ) . In this example , there is also a wave anchored around the superior vena cava in the right atrium ( Fig . 5C , right , and Video S5 ) , while in other simulations the excitation of the right atrium by the scroll wave in the left atrium is more irregular . Note that the dominant periods vary with the different interventions . In the baseline ( no remodeling ) model , the mean period of ( non-sustained ) reentry is 176 ms . With full electrical remodeling , the shortening of the APD allows a faster mean rhythm of 137 ms . For full structural remodeling , the conduction slowing increases the main period to 205 ms . Interestingly , for full electrical plus structural remodeling , the period is the same as for only electrical remodeling ( 138 ms ) , suggesting that the preferred pathways are shorter on average for electrical plus structural remodeling . These values fall within the range seen in patients with paroxysmal and persistent AF [21] , [22] . The different rhythms and their occurrence patterns ( Table 1 ) correspond to clinical and experimental observations . The pulmonary veins frequently act as triggers of AF [23] , and reentrant waves have been mapped in the pulmonary vein region [24] , [25] . In canine models of structural remodeling , and in typical atrial flutter , excitation often occurs around the tricuspid annulus [26] , sometimes in concert with reentry around the inferior vena cava [27] .
Recent years have seen a large increase in modeling atrial-specific aspects of arrhythmogenesis . In particular , there has been increased development of anatomical models , powered by increasing computational speed and data handling ( see , e . g . , [28] for a review on modeling and [29] for technical details on computational and visualization aspects ) . At the cellular level , multiple mathematical models describe the same ionic currents in different representations of human atrial myocytes . The Courtemanche et al . model [12] and the Nygren et al . model [30] are both well-established and have been compared in great detail [31] , [32] . Although their simulated behavior can be quite different , we believe that neither is empirically better . Rather they may represent intrinsic variability . Given that neither is obviously better , we have opted to use the Courtemanche et al . model , largely because it is more widely used . The Nygren et al . model was recently updated in terms of some of its potassium currents [33] and its intracellular calcium handling system [34] . The effects of electrical remodeling on action potential morphology , in particular the role of the individual currents involved in remodeling [7] , [35] , have been studied at the cellular level . In two-dimensional tissue , electrical remodeling accelerates spiral waves generated with both the Courtemanche et al . model [32] , [36] and the Nygren et al . model [32] . With electrical remodeling there is also a decrease in spiral wave meandering in the Courtemanche et al . model [32] , [36] , but not with the Nygren et al . model [32] . Such a decrease in spiral meandering with the Courtemanche et al . model is enhanced with increased IK1 [37] and can indeed occur in simulated atrial tissue with increased IK1 in the absence of electrical remodeling [38] . Increased IK1 also accelerates spiral waves [37] , as does another inwardly rectifying current IK , ACh , which is triggered in cholinergic AF [39] , [40] . In simulated tissue ( Ref . [36] , as well as in our simulations ( Fig . 4 ) ) , electrical remodeling decreases the wavelength in tissue strands and causes arrhythmias in anatomical models to be of longer duration ( Fig . 2; [32] , [36] ) . Decreasing the calcium current conductance , which is a main component of electrical remodeling , has similar effects in a human anatomical structure of virtual guinea pig ventricular cells [17] , [19] . Combining electrical remodeling and left atrial dilation leads to increased vulnerability to reentry in an anatomical model [41] , while combining electrical remodeling and decreased intercellular coupling causes shortened wavelength and sustained spiral wave activity in two-dimensional tissue simulations [42] , consistent with our results . Other lines of study , pursued with complex atrial models , include the effects of myocardial stretch on conduction [43] , [44] , incorporation of intrinsic APD heterogeneity [45] , [46] , [47] , and simulated ablation [48] , [49] . The mechanism underlying AF maintenance is not entirely clear . There are two predominant theories: ( i ) the multiple wavelet hypothesis and ( ii ) the “mother rotor” hypothesis . The multiple wavelet theory [50] hypothesizes that AF is composed of multiple interacting electrical wavelets and is maintained by the processes of wavebreak and reentry . The mother rotor theory [51] hypothesizes that , rather than the multiple , equally important wandering wavelets of the multiple wavelet hypothesis , there is one dominant “mother” reentrant wave that sheds and initiates daughter waves as conduction block occurs at multiple sites away from its core . The dynamics in our simulations are characterized by one or two reentrant waves , rotating fairly periodically around an anatomical obstacle or un-anchored in the ( left or right ) atrial free wall . Hence , our simulations are more consistent with the mother rotor theory . Further , in the presence of simulated structural remodeling , waves emanating from these dominant waves often exhibit local conduction block and transient fragmentation . Recent high-density mapping studies of patients with persistent AF and structural heart disease also found increased occurrence of conduction block in the right atrium compared to patients without persistent AF and structural heart disease [52] . However , the AF dynamics in the former patient group is characterized by more epicardial breakthroughs and more waves than our simulations , possibly due to increased decoupling of neighboring myocyte bundles [52] , [53] . Using non-invasive imaging techniques , Cuculich et al . determined the number of wavelets in AF patients as 1–5 , with more wavelets in patients with long-term persistent AF ( average 2 . 6 ) than in patients with paroxysmal AF ( average 1 . 1 ) [25] , consistent with our findings . However , these patients also had a considerable number of focal sites ( possibly due to spontaneous triggering , microreentry , or epicardial breakthroughs ) [25] , not seen in our simulations . Atrial flutter ( AFl ) is usually associated with a single macroreentrant circuit and may be very regular . However , AF can also exhibit a large degree of both temporal periodicity and spatial organization [54] , [55] . Hence , our simulated arrhythmias demonstrate characteristics of both AF and AFl . Clinically , AF and AFl may occur in the same patient , and AFl often converts to AF . By initiating reentry at different locations and at different times , we were able to explore a large range of spatiotemporal dynamics in our simulations . A large variety of ensuing dynamics is also observed clinically and experimentally . While much of this variability may stem from differences among experimental AF models and patient-to-patient variation in underlying heart disease , age , and stage of remodeling , there is considerable variability in activation patterns even among patients with the same underlying condition [22] . Importantly , many features observed in our simulations correlate directly with experimental and clinical characteristics . One example is reentry around the pulmonary veins , seen in our model with electrical remodeling . Another example is reentry around the tricuspid annulus , with 5 of 7 simulations of sustained activity with structural remodeling resulting in the direction of rotation being counter-clockwise . Such right-atrial reentry , involving slow conduction through the isthmus , is typical in patients with AFl [26] , [56] often with the majority ( 19/26 ) of cases having counter-clockwise rotation around the tricuspid annulus [57] . Further , with structural remodeling in our simulations , waves were also anchored around the inferior vena cava . Such dual-loop reentry involving the tricuspid annulus and the inferior vena cava is often observed in typical atrial flutter [27] . In addition to anatomical reentry , we also observed several examples of meandering scroll waves , which have been mapped experimentally and simulated computationally [58] . Finally we observed incidences of double-wave reentry , also observed clinically [56] and experimentally [59] . Rate gradients , with higher activation frequencies in the left atrium , exist in some animal models of AF [11] , [54] , [55] and have been observed in patients with both paroxysmal and persistent AF [22] , [60] , [61] , [62]; however , other chronic AF patients do not exhibit such inter-atrial variability [21] . In our simulations of remodeled tissue , there are no significant left-to-right gradients in activation frequency . However , our simulated AF with electrical remodeling alone was primarily a left atrial phenomenon in the sense that most dominant waves were found in the left atrium . This suggests that under some circumstances , the left atrium may be the driver of AF even in the absence of electrophysiological left-to-right heterogeneity ( see below ) . The extent to which perpetuation of AF depends on the wavelength varies considerably among different studies using different AF models and different methods for wavelength estimation . While several studies , both computational and clinical , have demonstrated a facilitation of AF maintenance with a decrease in wavelength [17] , [18] , others have not found such a dependence [63] . We found a clear functional dependence of reentry duration on wavelength in our simulations , and found that the wavelength must be below a value of around 7 cm for reentry perpetuation . However , the wavelength has to be considerably shorter for the majority of simulations to be sustained; this critical value for sustenance is about 5 cm . This value is consistent with a previous report of 5 cm using a direct measurement of the wavelength during simulated AF [17] . In contrast , clinical estimates for this threshold tend to be larger , with reported values of 12–13 cm [20] , [64] . Importantly , as detailed in Ref . [17] , these values are obtained through indirect methods , due to the inherent problem of insufficient spatial mapping , and tend to overestimate the wavelength . AF and AFl become increasingly common with age , and are also particularly frequent in patients with existing structural heart disease , valvular heart disease , coronary artery disease , ischemic heart disease , hypertension , or a history of heart attacks . Indeed , the majority of AF patients have one or more cardiovascular diseases in addition to AF . Electrical and structural remodeling due to chronic AF occur in concert with substrate changes due to any existing condition ( s ) . Changes in ionic currents due to chronic AF have been observed consistently in experimental models and in patients [9] . Recent studies have shown differences in several of the outward potassium currents between the left and the right atrial appendages , with some current differences being present in sinus rhythm or paroxysmal AF and others in chronic AF [65] , [66] . As it is presently unclear what type of inter-atrial gradients these appendage differences represent , we did not incorporate any left vs . right electrical heterogeneity in our model at this point . Structural remodeling in chronic AF shows more variability among patients and animal models than does electrical remodeling . Structural remodeling processes also occur on a much slower time scale than electrical remodeling , which may contribute to the substrate variability . The processes include myocyte hypertrophy , fibrosis , and changes in the expression levels of connexin , the protein comprising the gap junctions that couple cells . Fibrosis can manifest in both patchy patterns and diffuse morphologies . Fibrosis and decreases in connexin levels both impede propagation , while atrial dilation and cell hypertrophy increase atrial activation times . Our simulations using decreased coupling between virtual cells , which decrease conduction velocity and increase activation times , represent an unspecified propagation impairment that may be thought of as a decrease in connexin levels or a diffuse fibrosis . Although contractile remodeling is associated with chronic AF , in addition to electrical and structural remodeling , we do not include such remodeling here . The main reason is that the current data on contractile remodeling in the human atria is insufficient to incorporate into a quantitative model . Further , attempts at including contractile remodeling into the cellular model would almost certainly require a considerable expansion of the intracellular calcium handling system such as that recently developed by Koivumäki et al . [34] . Such an expansion would significantly increase the computational load of the simulations , and as we are already operating on the cusp of computational tractability , we did not incorporate those changes in this study . The pulmonary vein plays an important role in triggering AF and electrical isolation ( ablation ) of this area can be an efficient treatment for AF [23] . However , the pro-arrhythmic role of the pulmonary vein region is significantly lessened in chronic AF [61] , [67] , [68] , which is also exemplified by the decreased success of ablation therapy in this patient group . The pulmonary vein region has electrophysiological properties that are different from other regions of the atria , typically showing shorter refractory periods [67] , [69] . The details of which ionic currents are responsible for this heterogeneity and how the electrophysiology changes with progression of AF are not known , and hence we did not attempt to include special heterogeneity of the pulmonary vein region in this work . Further , this study focused on the maintenance , rather than the initiation of AF; further studies will be required to help elucidate triggering of AF due to mechanisms such as repolarization alternans and ectopy [14] , [70] in remodeled and healthy tissue . As described in the Methods , the three-dimensional atrial structure includes a gross coupling heterogeneity , with fast , normal , and slow tissues . However , the model does not include anisotropy . More prominent anisotropy during some types of structural remodelling [9] may play a role in AF perpetuation . We simulated structural remodeling as a decrease in coupling . A first step towards more realistic simulations of fibrosis could include implementing structural remodeling by randomly assigning a variable fraction of cells to be electrically inactive and surrounded by no-flux boundary conditions [10] . Test simulations show that such implementation of structural remodeling does not alter our conclusions . Recent studies have shown how fibrosis may affect repolarization [71] in addition to slowing down conduction - we did not include such effects in this work . Electrical and structural remodeling both shorten the atrial wavelength; electrical remodeling primarily through a decrease in APD , while structural remodeling primarily slows down conduction . The wavelength determines the mean reentry duration in the same manner for electrical and structural remodeling despite major differences in overall dynamics ( maximal number of filaments , conduction block , wave fragmentation , restitution properties , preferred dominant wave pathways , and whether dominant waves are anchored to anatomical obstacles or spiralling freely ) . As such , these findings have implications for our understanding of the mechanisms by which AF remodeling processes perpetuate AF . | Atrial fibrillation is an abnormal heart rhythm characterized by rapid and irregular activation of the upper chambers of the heart . Atrial fibrillation often shows a natural progression towards longer and more frequently occurring episodes and often occurs in patients with existing heart disease ( s ) . Because atrial fibrillation has several variants , is complex in nature , and evolves over time , it is very difficult and expensive to study comprehensively in large-animal models , in part due to the inherent technical difficulties of imaging whole-atria electrophysiology in vivo . Predictive multiscale computational modeling has the potential to fill this research void . We have incorporated aspects of chronic atrial fibrillation to model some of its various disease states . As such , this study represents the first comprehensive computational study of chronic atrial fibrillation maintenance in a biophysically detailed cell model in a realistic three-dimensional anatomy . Our simulations show that disease-like modifications to cellular processes , as well as to the coupling between cells , perpetuate simulated atrial fibrillation by accelerating the rhythm and/or increasing the number of circulating activation waves . Given the model's ability to reproduce a number of clinically and experimentally important features , we believe that it presents a useful framework for future studies of atrial electrodynamics in response to , e . g . , ion channel mutations and various drugs . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"medicine",
"physiology",
"biology",
"anatomy",
"and",
"physiology",
"biophysics",
"computational",
"biology",
"cardiovascular"
]
| 2012 | Effects of Electrical and Structural Remodeling on Atrial Fibrillation Maintenance: A Simulation Study |
The architecture of iso-orientation domains in the primary visual cortex ( V1 ) of placental carnivores and primates apparently follows species invariant quantitative laws . Dynamical optimization models assuming that neurons coordinate their stimulus preferences throughout cortical circuits linking millions of cells specifically predict these invariants . This might indicate that V1’s intrinsic connectome and its functional architecture adhere to a single optimization principle with high precision and robustness . To validate this hypothesis , it is critical to closely examine the quantitative predictions of alternative candidate theories . Random feedforward wiring within the retino-cortical pathway represents a conceptually appealing alternative to dynamical circuit optimization because random dimension-expanding projections are believed to generically exhibit computationally favorable properties for stimulus representations . Here , we ask whether the quantitative invariants of V1 architecture can be explained as a generic emergent property of random wiring . We generalize and examine the stochastic wiring model proposed by Ringach and coworkers , in which iso-orientation domains in the visual cortex arise through random feedforward connections between semi-regular mosaics of retinal ganglion cells ( RGCs ) and visual cortical neurons . We derive closed-form expressions for cortical receptive fields and domain layouts predicted by the model for perfectly hexagonal RGC mosaics . Including spatial disorder in the RGC positions considerably changes the domain layout properties as a function of disorder parameters such as position scatter and its correlations across the retina . However , independent of parameter choice , we find that the model predictions substantially deviate from the layout laws of iso-orientation domains observed experimentally . Considering random wiring with the currently most realistic model of RGC mosaic layouts , a pairwise interacting point process , the predicted layouts remain distinct from experimental observations and resemble Gaussian random fields . We conclude that V1 layout invariants are specific quantitative signatures of visual cortical optimization , which cannot be explained by generic random feedforward-wiring models .
Processing high-dimensional external stimuli and efficiently communicating their essential features to higher brain areas is a fundamental function of any sensory system . For many sensory modalities , this task is implemented via convergent and divergent neural pathways in which information from a large number of sensors is compressed into a smaller layer of neurons , transmitted , and then re-expanded into a larger neuronal layer . When sensory inputs are sparse , compression of the inputs through random convergent feedforward projections has been shown to retain much of the information present in the stimuli [1–3] . On the other hand , random expanding projections can lead to computationally powerful high-dimensional representations of such compressed signals , which combine separability of the inputs with high signal-to-noise ratio to facilitate downstream readouts [4] . Given these computational benefits , one might expect randomness to be a fundamental wiring principle employed by different sensory systems . The most striking example of a random expansion so far has been observed in the olfactory system of Drosophila melanogaster . Kenyon cells in the fly brain’s mushroom body were shown to integrate input from various olfactory glomeruli in combinations that are consistent with purely random choices from the overall distribution of glomerular projections to the mushroom body [5] . What is the role of random projections between neural layers in mammalian sensory systems ? Sompolinsky and others have argued that the human visual system , for instance , implements a compression-transmission-expansion strategy [3 , 4] . In fact , visual stimulus information is transmitted from about 5 million cone photoreceptors [6 , 7] to 1 million retinal ganglion cells ( RGCs ) [7] and then via the optic nerve to about 1 million lateral geniculate relay cells [8] to on the order of 100 million neurons in the primary visual cortex ( V1 ) [9 , 10] . We note that , while the overall connectivity indeed suggests compression for peripheral retinal regions [11] , close to the fovea RGC density is higher than the density of photoreceptors [12 , 13] . How much does randomness contribute to shaping the functional architecture of early visual cortical areas ? Projections between individual layers of the early mammalian visual pathway are clearly not entirely random . Visual information is mapped visuotopically from the retina to V1 such that neighboring groups of V1 neurons process information from neighboring regions in visual space . Yet , it has long been realized that many features of the spatial progression of receptive fields across V1 layer IV naturally emerge if random feedforward connections from groups of RGC cells to layer IV neurons ( via the lateral geniculate nucleus ( LGN ) ) are assumed ( see [14] for an early example ) . The most important of such features is orientation selectivity , i . e . the selective response to edge-like stimuli of a particular orientation . In carnivores , primates and their close relatives , orientation selectivity is arranged in patterns of iso-orientation domains . Iso-orientation domains ( orientation domains for short ) in V1 exhibit a continuous , roughly repetitive arrangement . A distance in the millimeter range , called the column spacing , separates close-by domains preferring the same orientation . The continuous progression of preferred orientations is interrupted by a system of topological defects , called pinwheel centers , at which neurons selective to the whole complement of stimulus orientations are located in close vicinity [15–20] . These topological defects exhibit two distinct topological charges , indicating that preferred orientations change clockwise or counterclockwise around the defect center [15 , 18 , 21–23] . More than 25 years ago , Soodak [24 , 25] ( see also [26] ) proposed random wiring between irregularly positioned retinal ganglion cells ( RGCs ) and layer IV neurons in V1 via the thalamus as a candidate mechanism defining the pattern of iso-orientation domains . According to this statistical wiring hypothesis , a V1 neuron randomly samples feedforward inputs from geniculate projections in the immediate vicinity of its receptive field center ( see e . g . [27] ) . The neuron then is likely to receive the strongest inputs from a central pair of ON/OFF RGCs , forming a so-called RGC dipole [28–30] . In this scheme , one ON and one OFF subregion dominate the receptive field ( RF ) of the V1 neuron and its response is tuned to the orientation perpendicular to the dipole axis . Thus , the preferred orientation of the neuron in this case is determined by the orientation of the RGC dipole . Consequently , the key prediction of the statistical wiring hypothesis is that the spatial arrangement of ON/OFF RGC cells in the retina essentially determines the spatial layout of orientation preference domains in V1 . Recently , Paik & Ringach showed that the statistical wiring hypothesis—when constructed with a hexagonal grid of RGCs—predicts a periodic orientation domain layout with a hexagonal autocorrelation function [28] . Moreover , it predicts that orientation preference is differently linked to the visuotopic map around pinwheels of positive or negative topological charge [29] . Qualitative signatures of both predictions were reported to be present in experimentally measured patterns [28 , 29] . Thus the statistical wiring model has conceptual appeal and is a mechanistically particularly transparent candidate explanation for V1 functional architecture ( see however [31 , 32] ) . Does the predictive power of the random wiring hypothesis for the early visual pathway reach beyond this qualitative agreement ? The recent discovery of species-invariant quantitative layout laws for the arrangement of pinwheel centers in tree shrews , galagos and ferrets [23] provides a unique opportunity to address this question . Kaschube et al . demonstrated that in these species , the statistics of pinwheel defect layouts is quantitatively invariant , with potential deviations in geometrical layout parameters of at most a few percent [23] . Specifically , the overall pinwheel density , defined as the average number of defects within the area of one square column spacing Λ2 was found to be virtually identical . Subsequently , orientation domain layouts from cat V1 were shown to exhibit pinwheel densities very close to those of the three species previously studied [33] . Additionally , Kaschube et al . found an entire set of local and non-local quantitative pinwheel layout features to be species-invariant ( see below ) . Following [23] , we refer to this overall layout of orientation domains as the common design . During mammalian evolution , the common design most likely arose independently in carnivores and euarchontans and potentially even in scandentia [23 , 34] . This is suggested by two lines of evidence: ( i ) The four species in which the common design has been observed so far are widely separated in terms of evolutionary descent , belonging to distinct supra-ordinal clades that split already during basal radiation of placentals [35–42] ( Fig 1A , see also [23 , 33] ) . Their last common ancestor was a small shrew-like mammal [40–42] that is unlikely to have possessed a columnar V1 architecture [23 , 34] . ( ii ) Distinct neuronal circuits underlie the generation of orientation selectivity in galago , ferret , tree shrew , and cat ( Fig 1B ) . Tree shrews , for instance , lack orientation selectivity in the input layer IV of V1 [43 , 44] and use intracortical circuits to compute contour orientation . In contrast , cats exhibit both , orientation selectivity and organization of selectivity into orientation domains already in layer IV and thus first generate orientation selectivity by thalamo-cortical circuits [45 , 46] ( see Fig 1B for further differences ) . Kaschube et al . used a dynamical self-organization model with long-range suppressive interactions , the long-range interaction model , to explain all features of the common design [23] . The hypothesis that randomness of feedforward connections between the retina/LGN and V1 could explain the common design is conceptually diametrically opposed to large-scale self-organization . In the long-range interaction model , the orientation preference of a neuron is chosen from an , in principle , unlimited afferent repertoire of potential receptive fields . Single neurons dynamically select a particular preferred orientation as a result of large-scale circuit interactions involving millions of other cortical neurons . In the statistical connectivity model , to the contrary , the preferred orientation of a cortical neuron is essentially imposed by the alignment of only one pair of neighboring ON-OFF RGCs , a local process involving in principle not more than 5 cells . Can the invariant layout laws of iso-orientation domains and pinwheels be explained as the generic outcome of a locally stochastic feedforward wiring of the early visual pathway ? More generally , do iso-orientation domains and pinwheels in different species adhere to identical layout laws because any mechanism that generates a retinotopic random feedforward circuit will automatically set up a layout that adheres to the common design ? Here , we systemically investigate the arrangements of iso-orientation domains generated by the statistical connectivity model and assess their consistency with the experimentally observed common design invariants . First , we consider the statistical wiring model with perfectly hexagonal mosaics of RGCs , its most tractable form . We derive closed-form expressions for cortical neuron receptive fields and orientation domain layouts resulting from the Moiré interference effect of hexagonal ON and OFF ganglion cell mosaics [28 , 29] . The pinwheel density of these pinwheel layouts is ρ = 2 3 ≈ 3 . 46 , substantially larger than experimentally observed . We then characterize the orientation domain layouts resulting from spatially disordered hexagonal mosaics . We find that parameters of RGC position disorder can not be tuned such that the statistical wiring model’s layouts match the quantitative invariants of the common design . Next , we examine a generalized class of noisy hexagonal mosaics that allows for spatially correlated disorder of RGC positions . This correlated retinal disorder induces local variations in column spacing , mimicking column spacing heterogeneity in the visual cortex [47 , 48] . With these mosaics , Moiré interference persists to larger disorder strength . Pinwheel densities , however , are unaffected by low and intermediate levels of disorder and increase from a lower bound of 3 . 5 for stronger disorder . Finally , we characterize the statistical connectivity model with RGC mosaics generated by Eglen’s pairwise interaction point process ( PIPP ) , the most realistic model for RGC mosaics currently available [31 , 32 , 49] . The resulting arrangements of iso-orientation domains and pinwheels are identical to those predicted by Gaussian random field models [22 , 50 , 51] . Their pinwheel densities can be tuned by applying band pass filters of different bandwidths . However , for all plausible filter shapes , pinwheel densities are substantially larger than experimentally observed . Our findings demonstrate that the mechanism for seeding patterns of iso-orientation domains described by the stochastic wiring model predicts column arrangements substantially different from the long-range interaction model and distinct from the experimentally observed invariant common design .
Our overall goal was to assess whether the layout of orientation domains predicted by the statistical wiring model are consistent with the observed common design invariants . To achieve this , we first sought to establish a benchmark for models of orientation domain layouts in general , to which predicted layouts can then be compared . To this end , we re-analyzed the data set used in [23] using the fully automated method described in the same study . The data set contains optical imaging of intrinsic signal experiments from tree shrew ( N = 26 ) , ferrets ( N = 82 ) , dark-reared ferrets ( N = 21 ) and galagos ( N = 9 ) . Because many previous studies used the statistical wiring model with parameters optimized to mimic the early visual pathway of the cat , e . g . [82] , we additionally analyzed data from 13 cat V1 hemispheres . Following [23] , we first computed the average pinwheel densities ( Fig 1C ) . Pinwheel densities of all four species , including cat were statistically indistinguishable from each other and statistically indistinguishable from π ( dark-reared ferrets excluded ) —the value predicted for the average pinwheel density by the long-range interaction model [23] . As a measure of pinwheel position variability , spanning all scales from single hypercolumn to the entire imaged region , we calculated the standard deviation , SD , of pinwheel density estimates in circular subregions of area A ( see Fig 1D for an illustration ) . For all species , the function SD ( A ) was well described by SD ( A ) = c ρ A γ ( 1 ) ( Fig 1E ) with ρ denoting the average pinwheel density . The variability exponents γ and variability coefficients c were similar in all four species ( Fig 1F ) . As a measure of relative pinwheel positioning on the hypercolumn scale , we computed the nearest neighbor ( NN ) distance statistics for pinwheels of same or opposite topological charge as well as independent of their topological charge ( see Fig 1D for an illustration ) . Distance distributions were unimodal and very similar ( Fig 1G–1J ) . Importantly , the distributions obtained from cat V1 were indistinguishable from the other three species . Mean NN distances , when measured in units of hypercolumns , were statistically indistinguishable ( Fig 1G–1J , insets ) . These findings confirm the results of [23 , 33] and show that cat primary visual cortex follows the same quantitative layout laws as in tree shrew , galago and ferret . From the above results , we extracted two types of consistency ranges that can be used as a benchmark for models of orientation domains in V1 . To be consistent with an observed layout of orientation domains , a model’s predictions should not be significantly different from experimental observations in at least one species . We thus defined one species consistency ranges spanned by the minimal lower and maximal upper margin of the single species confidence intervals for each parameter . If a model’s predicted layout parameters are located outside one or more of the one species consistency ranges , data from every species rejects this model at 5% significance level . This criterion is thus conservative in nature and does not assume that there is in fact one species invariant common design . If such a truly universal common design for orientation domains in fact exists , it would be appropriate to pool data from different species and to consider the more precisely defined confidence intervals of the grand average statistics as the relevant benchmark . To perform this more demanding test of model viability , we also defined common design consistency ranges as the 95% bootstrap confidence intervals obtained from the whole data set . If the layout parameters predicted by a model are within all of the common design consistency ranges , the model offers a quantitative account of the bona fide universal common design . If one or more layout parameters have predicted values outside the common design consistency ranges the model is inconsistent with the common design . With the current data set , if a model is common design consistent , it is also one species consistent . One species ( common design ) consistency ranges are shaded in light ( dark ) green in Fig 1C , 1E–1J and summarized in Table 1 . The tests of model viability defined above are most simply performed if the parameter values predicted by a model are determined exactly or with a numerical error that is much smaller than the empirical uncertainties . For models that can be solved accurately numerically , this can in principle always be achieved by a sufficiently large sample size of simulations . In the following , through analytical and numerical calculations , we will perform a comprehensive search through the statistical wiring model’s parameter space to identify regimes in which the model is one species consistent or common design consistent . The statistical wiring model formalizes the hypothesis that the spatial progression of orientation preference domains arises from the spatial distribution of RGC receptive fields on the retina via feedforward wiring . Fig 2A shows a simplified schematics of the early visual pathway in the cat [27 , 45 , 46 , 54–65] , from the retina to layer IV of V1 . A stimulus is focussed onto the retina through the cornea and lens , is sampled by RGC RFs and transmitted to the LGN . LGN neurons project to stellate cells in layer IV of V1 , whose responses are orientation tuned . Orientation tuning varies smoothly across the cortical surface . In the model , RGCs are assumed to be mono-synaptically connected one-to-one to relay cells in the LGN . Thus , the receptive fields of LGN neurons are similar to those of RGCs and the spatial arrangement of ON/OFF receptive fields of relay cells in the LGN mirrors the RGC receptive field mosaic . Neurons in the model visual cortex linearly sum inputs of LGN neurons ( Fig 2B ) . Their spatial receptive fields and orientation preferences are assumed to solely depend on the spatial arrangement of their afferent inputs . V1 neurons are assumed to receive dominant inputs from a small number of geniculo-cortical axons . Most of them sample from a single pair of ON/OFF RGCs , a so-called RGC dipole ( Fig 2B ) . The neuron’s receptive field then consists of one ON and one OFF subregion and its response to edge-like stimuli is tuned to an edge orientation orthogonal to the RGC-dipole vector ( Fig 2B ) . Within a mosaic of ON and OFF center RGCs , many such dipoles are present and the spatial arrangement of dipoles on the retina determines how tuning properties , e . g . the preferred orientation , change along a two-dimensional sheet parallel to the layers of the visual cortex . If ON and OFF RGCs are positioned on hexagonal lattices , the model predicts that a hexagonal pattern of orientation preference can arise through Moiré interference ( MI ) between the two lattices ( Fig 2C ) . Following [28 , 83] , we model RGC receptive fields using a Gaussian function GRFj ( x ) of width σr localized at the center position xj: GRF j ( x ) = ± exp - ( x j - x ) 2 2 σ r 2 , ( 2 ) where x indicates position in retinal space . All subsequent results remain qualitatively unchanged if a biologically more realistic difference-of-Gaussians ( see [84] ) is used . A plus or minus sign in Eq ( 2 ) indicates an ON or OFF center cell , respectively . The receptive field RFy of a visual cortical neuron at position y in the two-dimensional cortical sheet is obtained by summing several ganglion cell receptive fields with positive synaptic weights wj: RF y ( x ) = ∑ j w j ( y ) GRF j ( x ) . ( 3 ) The synaptic weights are chosen as w j ( y ) = exp - ( x j - y ) 2 2 σ s 2 . ( 4 ) The parameter σs sets the range from which a V1 neuron receives retino-thalamic inputs , xj denotes the center of an RGC receptive field . According to Eq ( 3 ) the spatial distribution of RGC locations determines how response properties change across cortex . For σs smaller than the lattice spacing , each cortical cell receives substantial input only from a very small number of ganglion cells . Inputs received by most cortical cells are dominated by one ON and one OFF center RGCs ( see inset in Fig 2A ) , forming an RGC dipole . The small σs regime is thus generally referred to as the dipole approximation of the model . While the dipole approximation leads to the robust emergence of simple-cell receptive fields with one ( ON , OFF ) or two ( ON-OFF ) subfields in the model V1 layer , it is worth mentioning that simple cells in cat and macaque monkey sometimes have more than two aligned , regularly spaced subfields ( e . g . ON-OFF-ON or OFF-ON-OFF ) ( see [85 , 86] ) . In the dipole approximation of the statistical wiring model , such simple-cell RFs almost never occur . While the model as defined above implements a deterministic wiring scheme , it represents a simplification of a more detailed formulation of the statistical connectivity model proposed in [83] . In the more detailed formulation , the synaptic weights between the cortical units and the retina/LGN are chosen at random from a Gaussian distribution with the shape given in Eq ( 4 ) . Ringach established in [83] that the spatial structure of the resulting domain layouts for the detailed and simplified model are nearly identical . We therefore refer to the model as statistical connectivity model . We used the linear response assumption [87 , 88] to determine cortical stimulus responses . A response R of a cortical neuron is modeled by the inner product between its receptive field RFy ( x ) and the stimulus , in our case an illumination pattern L ( x ) : R y = ∫ d 2 x RF y ( x ) L ( x ) . ( 5 ) Because Ry can become negative , a firing rate f of the cortical neuron is then defined through a static nonlinearity , e . g . half-wave rectification [87] . For the purposes of the present study , this nonlinearity can be neglected assuming that it does not alter core properties of the receptive field such as orientation preference and spatial frequency preference [82 , 83] . We derived close-form expressions for the pattern of cortical receptive fields across V1 that arises through Moiré interference in the case that ON and OFF center cells are localized on perfectly hexagonal lattices with different lattice constants r and r′ and relative angle α between the lattices ( see Fig 2C ) . Detailed derivations are provided in Methods , along with closed-form expressions for receptive fields , the frequency response of orientation selective neurons , and their spatial organization . Fig 3A depicts the analytically calculated orientation preference pattern generated through Moiré interference between two hexagonal ON/OFF RGC mosaics . Iso-orientation domains are organized in fine-grained parcellations on small scales and repeat in a hexagonal pattern on a larger scale Λ . The larger scale is the predicted column spacing of the orientation domain layout ( see Methods ) , and model parameters are chosen such that Λ ≈ 1mm as experimentally measured ( see Methods and [28 , 29] for details ) . The scale of the small parcels therefore is < 200μm . A magnified view of a small region of the domain layout is provided in Fig 3B along with three analytically determined cortical receptive fields at closely spaced locations roughly 100μm apart from each other . These receptive fields highlight that individual parcels contain highly tuned units with vastly different preferred orientations . This means that orientation preference changes abruptly on scales < 200μm in the predicted patterns . Clearly , these features distinguish the obtained pattern of orientation preferences from the experimentally observed domain layouts . While orientation selectivity in V1 exhibits some small scale scatter within orientation domains [89 , 90] , two-photon imaging suggests that orientation preferences progresses rather smoothly across the cortical surface [65 , 78] . Paik & Ringach implicitly assumed that random feedforward wiring from the retina/LGN to V1 effectively results in a smoothed version of the dipole layout ( see Fig 2C ) . To extract this smooth pattern of orientation preferences from the statistical connectivity model , they adopted a two-step procedure to suppress the small-scale variation in the Moiré interference pattern: First , locations with orientation selectivity index ( OSI ) larger than a threshold value are determined [28 , 29 , 82 , 83] . Second , the orientation selectivity of all other location is set to zero . The resulting layout is then filtered with a Gaussian lowpass filter resulting in continuous and smooth array of iso-orientation domains [28 , 29 , 82] . We find that the thresholding/smoothing procedure effectively extracts the dominant lowest spatial frequency Fourier components of the Moiré interference pattern . As derived in Methods , the lowest spatial frequency contribution to the Moiré interference pattern consists of six Fourier modes with identical amplitude and wave number k c = 4 π 3 r r ′ r 2 + r ′ 2 - 2 r r ′ cos ( Δ α ) , ( 6 ) Here , r , r′ denote to the lattice constants and Δα the angle between the hexagonal ON/OFF lattices ( Fig 2C ) . The smooth orientation domain layout resulting from Moiré interference can therefore be summarized in a complex-valued field z ( y ) composed of six planar waves with wave numbers kj and fixed phase factors uj , z ( y ) = ∑ j = 1 6 exp ( i k j y ) · u j . ( 7 ) The pattern of preferred orientations across the cortical coordinate y is given by the phase of this complex-valued field as , ϑ pref ( y ) = 1 2 arg z ( y ) . ( 8 ) Fig 3A ( bottom ) depicts ϑpref ( y ) as analytically determined . The pattern of pinwheels and iso-orientation domains is organized into a smooth hexagonal crystalline array . Interestingly , an identical layout of iso-orientation domains was constructed by Braitenberg et al . [91] based on an the idea that orientation preference is generated by discrete centers of inhibition in V1 . It was also found by Reich et al . to solve a symmetry defined class of models for the self-organization of iso-orientation domains [92 , 93] . Fig 3C shows the differences in preferred angle between the unfiltered domain layout of the Moiré interference pattern and its low frequency contribution , together with a histogram of the differences . With Δ ( x ) = ϑ1 ( x ) − ϑpref ( x ) , the difference d ( x ) between the two preferred angles is defined as d ( x ) = 1 2 abs ( arg ( e 2 i Δ ( x ) ) ) . The bimodal shape of the histogram indicates that the orientation preference of a large fraction of cortical locations differs substantially between unfiltered and smoothed layout . Roughly one fifth of all locations exhibit differences of orientation preferences of more than 45° . To compare our mathematical expression for the column spacing of the orientation domain layout to previous results , Eq ( 6 ) can be rewritten by introducing a parameter β representing the detuning between the two lattice constants in units of the lattice constant r′ → ( 1 + β ) r . The expression for the column spacing becomes Λ c ≡ 2 π k c = 3 2 · S · r , ( 9 ) where S is the distance between two vertices of the Moiré pattern in units of r , called the scaling factor [94–96] S = 1 + β β 2 + 2 ( 1 - cos ( Δ α ) ) ( 1 + β ) . ( 10 ) The difference between S ⋅ r and Λc is displayed in Fig 3A ( bottom ) . Eqs ( 9 ) and ( 10 ) are identical to previous results for the spacing of hexagonal Moiré patterns derived via geometrical considerations [95 , 96] . Using these explicit expressions for the iso-orientation domain layout and its column spacing Λc , we first evaluated the central quantity of the common design—the pinwheel density , i . e . the number of pinwheels per unit area Λ c 2 . Within each unit cell of area A = 3 2 ( S · r ) 2 , there is one “double pinwheel” of topological charge 1 , around which each orientation is represented twice , and two pinwheels of topological charge ± 1 2 . With Λ c 2 = 3 4 ( S · r ) 2 and counting the pinwheel with charge 1 as two pinwheels ( see below ) , the pinwheel density is ρ = 2 + 2 · 1 · Λ c 2 A = 2 3 ≈ 3 . 46 . ( 11 ) Notably , this value is outside of both , the common design consistency range and the single species consistency range for the experimentally measured pinwheel densities ( cf . Table 1 ) . Since the statistical connectivity model for perfectly hexagonal RGC mosaics results in a periodic array of pinwheels , all three nearest neighbor distance distributions of pinwheels are sharply peaked ( see also Supplementary Material of [23] ) and , thus , in disagreement with the distributions experimentally observed ( cf . Fig 1 ) . We compared these analytical results to numerically evaluated Moiré interference patterns ( Fig 4 ) . The fine-grained layouts of numerically and analytically obtained unfiltered layouts are almost indistinguishable ( cross-correlation coeff . 0 . 9 , Fig 4A top ) . This confirms the analytical treatment and indicates accuracy of the numerical implementation . A hierarchy of discrete spatial frequency contributions is apparent in amplitude spectra of both domain layouts ( Fig 4A ( bottom ) ) . The peaks at larger spatial frequencies in Fig 4A and 4B are localized at 3 k c as analytically predicted ( see Methods ) . To numerically generate the smoothed array of orientation domains , the layout in Fig 4A ( top ) was thresholded ( OSI > 0 . 25 , see Methods ) and subsequently smoothed with a Gaussian lowpass filter ( Fig 4B top ) [28 , 82] . In general , strongly tuned locations are those exactly between ON-OFF RGC pairs ( Fig 4B , inset ) . Fig 4C depicts the crystalline pinwheel arrangement of the analytically calculated smoothed layout ( Eq ( 8 ) ) as well as the six dominant low frequency Moiré modes in the amplitude spectrum . While the numerically obtained layouts and its analytical approximation are similar ( cross-correlation coeff . 0 . 6 ) , one major difference can be observed: the pinwheel of topological charge 1 is replaced by two pinwheels of topological charge 1 2 in the numerically obtained layouts , along with subtle deformations of adjacent orientation domains ( compare Fig 4B and 4C ) . To see why this is the case , we note that the pinwheel with charge 1 in the analytically calculated pattern ( Eq ( 8 ) ) arises from a zero of the field z ( x ) ( Eq ( 7 ) ) with multiplicity two . A phase singularity of a complex-valued field arising from a zero with multiplicity N > 1 is structurally unstable and unfolds upon generic infinitesimal perturbations into N closely spaced singularities of multiplicity one [97] . The numerical procedure of discretizing V1 unit positions on a numerical grid , OSI thresholding , and smoothing realizes such a perturbation and this explains why in the numerical solutions the pinwheel of charge 1 unfolds into two adjacent pinwheels of charge 1 2 . So far , we have studied the idealized situation of iso-orientation domains induced by perfectly ordered hexagonal RGC mosaics . RGC mosaics in the eye , however , are not perfectly hexagonal but exhibit substantial spatial irregularity [26 , 98] . Therefore , we next turned to numerically investigate the statistical connectivity model with hexagonal RGC mosaics subject to Gaussian disorder in RGC position as previously described [28 , 29] . The effect of ganglion cells displaced by Gaussian distributed offsets with standard deviation σ = η ⋅ r is illustrated in Fig 5 . The parameter r is the lattice constant and η is the disorder strength . Fig 5 shows the unfiltered orientation domain layout ( far left ) , the layout thresholded for cells with an OSI > 0 . 25 ( left ) , the smoothed thresholded layout ( right ) as well as its amplitude spectrum ( far right ) , numerically obtained for η = 0 . 12 . As in the perfectly ordered case , the unfiltered layout of the noisy Moiré interference model exhibits a substantial scatter of orientation preferences across small scales . For a disorder strength of η = 0 . 12 ( Fig 5A ) , the domain layout is still dominated by the six lowest spatial frequency Moiré modes also present in the perfectly ordered system . For a disorder strength of η = 0 . 3 ( Fig 5B ) , the amplitude spectrum ( Fig 5B , far right ) lacks any indication of theses Moiré modes indicating that Moiré interference no longer takes place . As a consequence the resulting layouts of iso-orientation domains lack a typical column spacing . To characterize the model orientation domain arrangements , we first calculate amplitude spectra for both , unfiltered and smoothed layouts ( Fig 5A and 5B ) , | R ( k ) | = ∫ d 2 x z ( x ) e i k x where z ( x ) = e 2 i ϑ pref ( x ) . ( 12 ) Normalizing and radially averaging yields the so-called marginal amplitude spectrum ( Fig 5C and 5D ) , f ( k ) = ∫ 0 2 π d ϑ | R ( k cos ( ϑ ) , k sin ( ϑ ) ) | max k ∫ 0 2 π d ϑ | R ( k cos ( ϑ ) , k sin ( ϑ ) ) | . ( 13 ) The sharp peak at kc corresponds to the dominant Moiré mode indicating that orientation domain layouts exhibit a typical column spacing . For increasing disorder , the relative levels of peak height to background decreases while the peak width remains small . As expected , marginal amplitude spectra of unfiltered and the smoothed layout mainly differ in the strength of background components . The flat amplitude spectrum of the unfiltered iso-orientation domain layouts for large disorder strength is transformed into a Gaussian amplitude spectrum by the lowpass filtering . Based on this assessment , the disorder strength η has to be smaller than 0 . 3 to ensure that layouts exhibit a typical spacing between adjacent iso-orientation domains . We next systematically evaluated the core layout parameters of the common design—pinwheel density and pinwheel nearest neighbor distance distributions for the statistical wiring model with disordered hexagonal RGC mosaics . To compare the model predictions with experiments , we estimated the column spacing of the model orientation domain layouts as well as pinwheel layout parameters using the exact same methods that we applied to the experimental data ( see Methods ) . For weak disorder , column spacing estimates closely match the theoretical prediction Λc ( Fig 6A ) , confirming the accuracy of the wavelet method . For disorder strengths larger then 0 . 12 , Moiré modes are no longer the dominant spatial frequency contribution in the model layouts and the estimated column spacing increases with disorder strength . Having estimated the column spacing , we analyzed model orientation domain layouts with respect to the common design parameters ( Fig 6B–6D ) . As expected , pinwheel densities approach the analytical predicted value of 2 3 for weak disorder ( Fig 6B ) and increase with increasing disorder strength . This increase is largely caused by the increase in the estimated column spacing ( Fig 6A ) and does not involve a massive generation of additional pinwheels for larger disorder strength . We next calculated the standard deviation of pinwheel densities as a function of the area A of randomly selected subregions of the iso-orientation domain layouts . Generally , the standard deviation’s decay with subregion size followed a power law with increasing area size , with larger exponents for weak disorder ( Fig 6D ) . Fig 6E and 6F show a complete characterization of pinwheel nearest neighbor ( NN ) distance distributions of the noisy Moiré interference model . Histograms for NN distances for arbitrary charge are bimodal for weak disorder ( Fig 6E ) . The peak at smaller NN distances results from the unfolding of pinwheels with topological charge 1 into two adjacent pinwheels of topological charge 1/2 for finite disorder strength ( see above and Figs 3 and 4 ) . For the same reason , the NN distance histogram for pinwheels of identical topological charge is also bimodal ( Fig 6F ) . With increasing disorder strength , both distributions become unimodal ( Fig 6E and 6F left ) . The NN distance distribution for pinwheels of opposite sign is unimodal for all parameter values , indicating that only very few additional piwheel pairs are added to the pinwheels of the Moiré layout for small and intermediate disorder strengths ( Fig 6G ) . The overall decay in mean NN distance for strong disorder in all three histograms mostly reflects the increase in measured column spacing ( see Fig 6A ) . Based on these results , we attempted to identify a disorder strength for which all NN pinwheel distance distributions resembled the experimental data . To this end , we calculated the squared error between the calculated histograms and the experimental data as a function of disorder strength ( Fig 6H ) . Smallest deviations from experimental data were obtained around η ≈ 0 . 11 for all three NN distance distributions . Fig 7 summarizes all common design features determined for disordered Moiré interference model layouts as a function of disorder parameter and compares them to the experimentally observed values in tree shrew , galago , ferret , dark-reared ferrets , and cats . Light ( dark ) green shaded areas indicate the single species ( common design ) consistency ranges ( see Fig 1 , cf . Table 1 ) . With increasing disorder , pinwheel density of model layouts steadily increases from ρ = 2 3 ≈ 3 . 46 ( Fig 7A ) and always lies above the single species consistency range . Thus , the pinwheel density of model orientation domain layouts is inconsistent with pinwheel densities observed in all species . Next , we fitted the empirically observed power law , Eq ( 1 ) , to the standard deviation of the pinwheel density estimate in increasing subregions of area A ( see Fig 6D ) [23] . The variability exponent γ is consistent with experiments for disorder levels exceeding η = 0 . 15 . The variability constant c is monotonically increasing up to η ≈ 0 . 15 at which point the model domain layouts lose their typical column spacing ( cf . Fig 5 ) and the increasing pinwheel density ρ causes a drop ( Fig 7C ) . Fig 7D–7F displays the mean pinwheel NN distances as function of the disorder strength , all of which substantially decrease with increasing disorder strength . This can be attributed to the increasing mean column spacing of the domain layouts under increasing disorder ( see Fig 6A ) . Mean NN distances for weak disorder strength are close to the experimental data , but NN distance distributions for pinwheels of different topological charge and independent of topological charge are bimodal for weak disorder ( Fig 6E and 6F ) . The latter is clearly distinct from the experimental data ( cf . Fig 1 , [23] ) . Fig 7G shows an overview of the consistency of model orientation domain layout parameters with the data for various disorder strengths . As can be seen , no strength of disorder results in layouts that are consistent with the common design for all layout parameters . Perhaps , even more surprising , pinwheel density and NN distance for pinwheels of the same sign are inconsistent with the individual values obtained for each species , no matter how the strength of disorder is chosen . The above results show that the statistical connectivity model with hexagonal mosaics is unable to reproduce all features of the common design , even if spatially uncorrelated position disorder is imposed on the RGC positions . Whatever the source of disorder that causes the irregularity in the RGCs’ positions , it is plausible to assume that it is correlated on scales spanning several RGCs . Such spatial correlations would preserve the Moiré effect locally , yet generate spatial irregularity in orientation domain layouts . To test whether correlated positional disorder can produce model arrangements of orientation domains that match experimental observations , we generalized the noisy hexagonal mosaics proposed in [28 , 83] to include spatial correlations . To obtain noisy hexagonal RGC mosaics with spatial correlations , we started with a hexagonal array of RGC positions . The position of each lattice point xi was then shifted depending on its position according to xi → xi + η y ( xi ) . The shift η y ( x ) with amplitude η for y ( x ) = ( y1 ( x ) , y2 ( x ) ) was chosen from a Gaussian random field with vanishing mean 〈y1 ( x ) 〉 = 〈y2 ( x ) 〉 = 0 , fixed standard deviation std ( y1 ( x ) ) = std ( y2 ( x ) ) = 1 and correlation function 〈 y ( x 1 ) y ( x 2 ) 〉 = 2 exp ( - | x 1 - x 2 | 2 2 σ 2 ) with correlation length σ ( see Methods ) where y1 and y2 are statistically independent . The two parameters , correlation length σ and amplitude η were expressed in units of the lattice constant r . Fig 8A and 8B illustrates this procedure . RGCs are shifted in a coordinated manner across the plane , correlated in both direction and magnitude of the shift . The determinant of the Jacobian det J ( x ) = det ∂ y 1 ( x ) ∂ x 1 ∂ y 1 ( x ) ∂ x 2 ∂ y 2 ( x ) ∂ x 1 ∂ y 2 ( x ) ∂ x 2 , ( 14 ) measures the local change of RGC lattice constant . In regions of negative det J , RGCs are closer together than average , in regions of positive det J , RGCs are further apart ( contour lines in Fig 8B ) . In primates , regions of higher density are predicted to have higher cortical magnification and vice versa [12] . The RGC mosaics with correlated positional noise therefore imply local fluctuations in the cortical magnification factor on the scale of the noise correlation length . Fig 8C-8E display unfiltered and smoothed model layouts obtained with spatially correlated noisy hexagonal mosaics as well as their amplitude spectra . As expected , orientation domain layouts exhibit a typical column spacing up to higher disorder strengths , when the position disorder was correlated ( compare Fig 8F with Fig 5D ) . Locally , Moiré interference leads to a roughly hexagonal layout of columns that is distorted on larger scales . Both , the orientation of the hexagons as well as column spacings change continuously across the layout . For weak disorder , the amplitude spectrum still exhibits six peaks , indicating a globally hexagonal layout ( Fig 8C , right ) . For intermediate disorder local column spacing and direction of the hexagons varies to the extend that peaks can hardly be identified in the amplitude spectrum of the resulting domain layout . In particular , the spatially varying local column spacing leads to a broader peak in the radially averaged amplitude spectrum with increasing disorder strength ( Fig 8F ) . This is in contrast to the case of uncorrelated disorder ( cf . Fig 5D ) . Note that experimental iso-orientation domain layouts exhibit a similarly broad peak in their marginal amplitude spectra [99] . We quantified the pinwheel density of orientation domain layouts obtained with correlated noisy hexagonal RGCs ( Fig 8G ) as a function of disorder correlation length and disorder strengths . Independent of the disorder correlation length , pinwheel densities plateau around 2 3 for weak disorder and monotonically increase above a critical disorder strength . This critical disorder strength is higher , the larger the correlation length . Thus , model pinwheel densities are inconsistent with the individual values obtained for each species , no matter what the strength of disorder or correlation length is . Fig 8H illustrates that the pinwheel density increases with increasing disorder strength largely because the overall measured column spacing increases , not because additional pinwheels appear in the layouts . In fact , the number of pinwheels per mm2 is almost independent of either correlation length or disorder strength . Finally , we examined whether the statistical connectivity model could reproduce the common design invariants with RGCs distributed in space according to a pairwise interacting point process ( PIPP ) . The PIPP developed by Eglen et al . [49] is currently the experimentally best supported model for RGCs mosaics and was shown by several studies to generate RGC positions which accurately reproduce a variety of spatial statistics of RGC mosaics [31 , 32 , 49 , 82] . The PIPP model generates samples from a statistical ensemble of RGC mosaics by iteratively updating RGC positions to maximize a target joint probability density , specified by pairwise interactions between neighboring RGCs ( for details see Materials & Methods ) . Each PIPP mosaic represents a random realization of a regularly-spaced RGC mosaic with radially isotropic autocorrelograms [31] and lacks long-range positional order . Fig 9A depicts a realization of a PIPP with parameters choosen to reproduce cat RGC mosaics ( for details see Materials & Methods ) . We generated thresholded and smoothed iso-orientation domain layouts from PIPP RGC mosaics as from the ordered mosaics ( Fig 9A and 9B left ) . The lack of long-range positional order in ON and OFF mosaics prevents any Moiré interference between them . Thus , no typical spacing between adjacent columns preferring the same orientation is set in the model layouts ( Fig 9B , right , see also [31 , 32] ) . Spectral power in these layouts is broadly distributed and monotonically decays with increasing spatial frequency . As a consequence , one needs to apply bandpass filtering to obtain orientation domain layouts that are at least qualitatively resembling the experimental data . We used a flexible band-pass filtering function f ( k ) of the following form to the amplitude spectrum: f ( k ) = a | k | β exp ( - | k | 2 b ) , ( 15 ) with a , b , β > 0 . The filter function was normalized such that ∫ d 2 k f ( k ) = 2 π . ( 16 ) With this normalization , one can define the mean column spacing of the resulting layout via Λ = 2 π / k ¯ with k ¯ = ∫ 0 ∞ d k 2 π k f ( k ) = 1 . ( 17 ) By increasing the parameter β , the shape of the filter can be changed from Gaussian lowpass ( β = 0 , Fig 9C , left ) to wide bandpass ( β = 2 , Fig 9C , middle ) and to narrow bandpass ( β = 10 , Fig 9C , right ) . There is an additional degree of freedom in this filter definition , namely how the filter is scaled relative to the absolute physical units mm−1 of the amplitude spectrum . To scan a wide range of filter shapes and column spacings , we varied β between 0 and 10 and choose the scaling such that Λ varied between 0 . 6 mm and 1 . 2 mm , i . e . covering the entire range of experimentally observed mean column spacings in tree shrew , galago , cat , and ferret [23 , 33] . We then measured the pinwheel densities of the resulting statistical connectivity model layouts ( Fig 9E , right ) , where pinwheel density was defined as the number of pinwheels within an area Λ2 . Pinwheel densities were independent of the scale Λ and increased monotonically with increasing spectral width ( decreasing β ) . They are in general substantially larger than the experimentally observed value of 3 . 14 ( see also Fig 9C ) and outside of the single-species/common-design consistency range . Qualitatively , iso-orientation domain layouts generated with the PIPP RGC mosaics resembled those generated from Gaussian random field ( GRF ) [22 , 50 , 51] ( Fig 9C ) . In fact , we find that this resemblance is quantitative . Fig 9E depicts the analytical prediction [51] for the pinwheel density of orientation domains obtained with GRFs with a marginal amplitude spectrum corresponding to the filter function in Eq ( 15 ) ( Fig 9E , left ) . Pinwheel densities for GRF layouts and layouts obtained from PIPP mosaics with the statistical connectivity model are indistinguishable . For the pinwheel density to be consistent with at least the single-species consistency range ( ρ < 3 . 42 ) , amplitude spectra had to be much more peaked ( β ≥ 17 ) than experimentally observed [99] . Finally , we filtered statistical connectivity model layouts with the Fermi-Filter function as used in [23 , 33] with cut-off wavelengths of 0 . 3 mm and 1 . 2 mm ( Fig 9D ) . Again , pinwheel densities were much larger than those observed in the experimental data and outside of the single-species and common-design consistency range . In summary , iso-orientation domain layouts generated by the statistical connectivity model using PIPP RGC mosaics quantitatively resemble layouts derived from Gaussian random fields . Their statistics is distinct from the statistics of experimentally measured layouts .
The statistical wiring model analyzed in the present study is only one representative of possible random wiring schemes . One could argue that alternative , perhaps more realistic , schemes might do a better job at reproducing the experimentally observed pinwheel layouts . There is good evidence that the spatial statistics of RGC mosaics is well approximated by Eglen’s PIPP [31 , 32 , 49] , and , hence , there is little freedom of choice at the retinal level . In contrast , at the next network layer , two main modifications or extensions of the statistical wiring model could be considered: ( i ) adding an additional layer to the feedforward network implementing the transformation of the retinal input structure by the lateral geniculate nucleus ( LGN ) ( ii ) choosing different probabilistic connectivity rules between the retinal/LGN layer and the primary visual cortex . We argue that both modifications of the random wiring approach are unlikely to improve the consistency of the model with experimental data . Regarding ( i ) , Martinez et al . [100] have recently tried to infer the mapping between RGC inputs and LGN relay cells using a statistical connectivity approach . In their model , ON and OFF cell types were homogeneously distributed and their polarity ( ON or OFF ) was inherited from the nearest retinal input . Connection probability between RGCs and LGN neurons was modeled as an isotropic Gaussian function of the relative distance between the RF centers of the presynaptic and postsynaptic partners . With this simple wiring scheme , together with similar connectivity rules for the population of inhibitory interneurons , several spatiotemporal properties of LGN RFs robustly agreed with the experimental data . In the architecture between the retina and the LGN proposed by these authors , the dipoles of ON- and OFF-center cells that characterize the retinal mosaic are transformed into small clusters of same-sign relay cells . The LGN ON-OFF dipoles occur at the boundaries of these clusters with LGN dipole orientations strongly correlating but not necessarily matching dipole orientations in the retina . Notably , dipole density and dipole angle correlation length in the LGN is not increased compared to the retina . The data and modeling by Martinez et al . suggest that the LGN mosaics do not systematically alter the spatial structure of RGC inputs beyond providing an additional source of dipole disorder . Additional disorder imposed by the LGN mosaics would likely add a uniform level of disorder to all spatial frequency components in the unfiltered orientation domain layouts predicted by RGC dipole structure . When hexagonal mosaics are considered , the disorder strength that has to be assumed to match the spatial distribution of RGC cells found in experiment is already rather high [28 , 29 , 31 , 32 , 83] . Additional noise is likely to obstruct any remaining Moiré interference . We therefore speculate that when considering an additional LGN layer , after smoothing , domain layouts for both , the noisy Moiré interference model and the model with PIPP mosaics would be similar to those obtained with PIPP mosaics [31 , 32 , 82] . As we have shown in the present study , the spatial statistics of these layouts resembles those derived from Gaussian random fields and is inconsistent with the data obtained for any of the four species analyzed ( cf . Fig 9 , see also [23 , 50] ) . A similar argument can be made for alternative probabilistic connectivity rules between the retinal/LGN layer and V1 . As our analysis shows , the dipole structure emerging from “realistic” RGC mosaics ( be it very noisy hexagonal mosaics or PIPP mosaics ) is spatially fine-grained because dipole angles vary over short distances in cortical space relative to the typical size of an iso-orientation domain . For this reason , the statistical connectivity model requires an additional smoothing step ( cf . Figs 4 , 5 and 9 ) to yield smooth orientation domain layouts as observed in experiment [65 , 78] . Unless the connectivity rule is assumed to specifically select dipoles with a similar angle from a larger spatial region of the retina , or neurons within an iso-orientation domain are assumed to choose one particular dipole to receive the input from and ignore all other dipoles in the vicinity , such spatial averaging within the cortical layer will always be required no matter what the actual probabilistic connectivity rule is . Domain layouts resulting from such spatial averaging of weakly correlated dipole angles ( see also [32] ) are likely to follow layout statistics that resemble those of Gaussian random fields , independently of the connectivity rule assumed . If neurons are assumed to select specific dipoles out of the repertoire of “available” ones , then the overall spatial layout of RGC dipoles is not informative about the resulting domain layout , which contradicts the main hypothesis of the statistical wiring model . Ultimately , the key experiment to provide support for the random wiring approach consists of determining both , the orientation domain layout and the retinotopic map in a single animal and , in a second step , correlate these with the spatial arrangement of RGCs in the same animal . This challenging experimental task is still awaiting its completion . So far , the only model class able to robustly reproduce all common design parameters , describes the formation of orientation domain layouts as a deterministic optimization process converging to quasi-periodic pinwheel-rich orientation domain patterns associated with and stabilized by a matching system of intrinsic horizontal connections [23 , 101 , 102] . Irregular layouts of orientation domains dynamically emerge as a consequence of large-scale circuit optimization of domain patterns and intrinsic circuits . Is this agreement between model and data good evidence for global circuit optimization or are there simple alternative explanations such as the random feedforward wiring hypothesis that can explain the invariant statistical properties of orientation domains ? Qualitatively , it is in fact tempting to attribute the spatial irregularity and apparent randomness of pinwheel layouts in V1 to some general kind of “biological noise” . In this view , the quantitative laws of pinwheel organization that Kaschube et al . found [23] would then be conceived as outcome of a largely random process underlying the emergence of orientation domains . By now , however , all proposals based on the assumption of disorder as the determinant of spatial irregularity have failed to reproduce the common design parameters and laws that have now been observed in four divergent species . Orientation domain layouts obtained from statistical ensembles of Gaussian random fields [22 , 50 , 51] as well as phase randomized layouts derived from experimental data [23] , exhibit pinwheel densities that are substantially higher than experimentally observed . Importantly , most dynamical models for the development of orientation domains produce such Gaussian random domain layouts during the initial emergence of orientation selectivity [22 , 50] . Therefore approaches based on “frozen” early states of such models are also ruled out by the existing data ( see [103] ) . The present study shows that a mechanistic and biologically plausible feedforward model of the early visual pathway based on ( i ) noisy hexagonal placement of RGCs or ( ii ) a more realistic semi-regular positioning of RGCs generated by the PIPP also generates layouts distinct from experimental observations . These findings illustrate that orientation domains and pinwheels positions , although spatially non-periodic and irregular , follow a rather distinct set of layout laws . These laws cannot easily be accounted for by a spatial irregularity or randomness in the structure of afferent projections to visual cortical neurons . A further conceivable and potentially critical source of stochasticity that is often overlooked is randomness within intracortical circuits . The field approach employed in various models for orientation domain layouts , such as the long-range interaction model , represents an idealization of a complex network , in which every neuron is characterized by its own set of inputs and outputs . These inputs and outputs may , at least to some extent , be stochastic . How and to what extent randomness in intracortical connections can affect and shape orientation domain layouts is not well understood . In that respect , it is interesting to note that model networks for largely stochastic intracortical circuits are able to generate and robustly maintain orientation selective responses to afferent inputs and can lead to highly coherent orientation domains [104 , 105] . Paik & Ringach reported indications of hexagonal order in visual cortical orientation domain patterns of tree shrew , ferret , cat , and macaque monkey [28 , 29] . Two recent studies have casted doubt on the hypothesis that this hexagonal order echoes hexagonal or quasi-hexagonal arrangements of ganglion cell mosaics in the retina [31 , 32] . Hore et al . showed that noisy hexagonal lattices do not capture the spatial statistics of RGC mosaic . Moreover , the positional correlations in measured mosaics extends to only 200–350μm , far less than required for generating Moiré interference [31] . More generally , Schottdorf et al . studied the spatial arrangement of RGC dipole angles in cat beta cell and primate parasol RF mosaics [32] . According to the statistical wiring hypothesis , dipole angle correlations should follow the spatial correlations of preferred orientations in the primary visual cortex , i . e . be positively correlated on short scales ( 0–300 μm ) and negatively correlated on larger scales ( 300–600 μm ) in the retina . By introducing a positive control point process that ( i ) reproduces both , the nearest neighbor spatial statistics and the spatial autocorrelation structure of parasol cell mosaics and ( ii ) exhibits a tunable degree of spatial correlations of dipole angles , they were able to show that , given the size of available data sets , the presence of even weak angular correlations in the data is very unlikely . If not from the structure of RGC mosaics , where does the apparent hexagonal organization in orientation domains come from ? A variety of self-organization models on all levels of biological detail have been shown to generate orientation domains with hexagonal arrangements , e . g . [14 , 92 , 93 , 103 , 104 , 106 , 107] ( notably including the earliest theory for the self-organization of orientation preference by von der Malsburg in 1973 ) . Thus , hexagonal order , even if present , would not provide specific evidence in favor of Moiré interference between RGCs . Future work will have to elucidate whether the long-range interaction model for orientation domains [23 , 101 , 102] can not only explain the common design statistics , but at the same time account for the observed hexagonal order in the visual cortex . Compared to the dense sampling of stimulus space by cortical neurons , the repertoire of detectors on the retina that input into a given cortical area is limited . For the cat visual pathway , Alonso et al . estimated the number of LGN X-relay cells converging onto a single simple cell in V1 to be ∼ 20–40 [27] , based on measuring the probability of finding a connection between individual geniculate and cortical neurons with overlapping receptive fields . This estimate was later confirmed by directly measuring population receptive fields of ON and OFF thalamic inputs to a single orientation column [108] . With an expansion of around 1 . 5–2 . 0 from X-cells in the retina to X-relay cells in the LGN [109 , 110] , each simple cell in V1 receives on average input from only ∼ 10–25 RGCs . This not only implies that random afferent inputs to cortical neurons might seed groups of V1 neurons with similar orientation preferences but also that they might in fact impose substantial biases on the preferred orientation that can be adopted by the cortical neurons . The postnatal development of orientation columns could then be imagined as a dynamical activity-dependent process which refines and remodels an initial set of small biases provided by the RGC mosaic model through Hebbian learning rules and other mechanisms of synaptic plasticity . The up to now most striking experimental evidence that retinal organization can impose local biases in V1 function architecture was revealed by the finding that the pattern of retinal blood vessels can specifically determine the layout of ocular dominance columns in squirrel monkey [111] ( for a modeling study see [112] ) . Dynamical models of orientation column formation generally assume no a priori constraints or biases as to which preferred orientation a given position in the cortical surface can acquire . Usually random initial conditions determine which instance from the large intrinsic repertoire of stable potential domain layouts is adopted . Including seeds and biases derived from RGC mosaics in such models for the dynamical formation of V1 orientation domain layouts may elucidate the potentially complex interplay between a sparse set of subcortical feedforward constraints and self-organization in a dense almost continuum-like intracortical network . The present study provides a detailed description of a candidate set of such subcortical biases and , therefore , can serve as a foundation for such future investigations . The common design invariants comprise four distinct functions in addition to the apparently invariant pinwheel density . As such , they represent a rather specific quantitative characterization of orientation domain layouts . It is , thus , not surprising that entire model classes have been rejected based on whether their predictions match these invariants . Since the discovery of visual cortical functional architecture more than fifty years ago , a large number of models based on a variety of circuit mechanisms has been proposed to account for their postnatal formation ( see [113–115] for reviews ) . Many of these models are explicitly or implicitly based on optimization principles and attribute a functional advantage to the intriguing spatial arrangement of orientation domains . Because many , even mutually exclusive , models could qualitatively account for main features of orientation domain layouts such as the presence of pinwheels or the roughly periodic arrangement of columns , theory could not provide decisive evidence on whether to favor one hypothesis over another . It is only in recent years that the large available data sets have started to allow for a rigorous quantitative analysis of visual cortical architecture and its design principles in distinct evolutionary lineages . To date , abstract optimization models have been analyzed most comprehensively , providing in many case the complete phase diagram . For instance , Reichl et . al . systematically evaluated energy-minimization-based models for the coordinated optimization of orientation preference and ocular dominance layouts [92 , 93] . By quantitatively comparing model solutions to the common design , they were able to rule out a whole variety of otherwise intuitive and plausible principles for their emergence . It is desirable to obtain a similarly quantitative understanding of more detailed models for the formation of orientation domains . In this regard , the analysis of abstract models is informative because there is a many-to-one relationship between detailed models of the visual cortical pathway and those abstract formulations . Abstract models often can be shown to be representative of an entire universality class and , once comprehensively characterized , the questions becomes whether more complex modeling schemes are simply complicated instantiations of such a class . For models of an intermediate degree of realism , semi-analytical perturbation methods can be employed to explicitly provide this mapping . Using this approach , Keil & Wolf studied orientation domain layouts predicted by a widely used representative of a general optimization framework [103] . According to this framework , the primary visual cortex is optimized for achieving an optimal tradeoff between the representation of all combinations of local edge-like stimuli , i . e . all positions in the visual field and all orientations , and the overall continuity of this representation across the cortical surface [116] . While this framework has successfully explained a variety of qualitative aspects of orientation domain design , e . g . [116–118] , the authors found quantitative disagreement with the common design in all physiologically realistic parameter regimes of the representative model [103] . Their analysis enabled an unbiased comprehensive search of the model’s parameter space for a match to the experimental data and indicated alternative more promising optimization hypotheses to explain the experimentally observed V1 functional architecture . Although the statistical wiring model is still rather simplistic , it is hard to make analytical or semi-analytical progress as soon as RGC mosaics with the necessary degree of realism are considered . In this case , the question of whether models account for the cortical architecture can only be answered with the approach we have pursued here , i . e . by systematic comparison between their solutions , experimental data , as well as predictions from minimal approaches . The results presented here show that this approach can indeed be successfully applied to rule out candidate mechanisms as sufficient explanations for the emergence of V1 functional architectures . We expect a re-examination of the quantitative predictions of other modeling approaches to be highly informative about candidate mechanisms for the formation of V1 functional architecture . This present study provides the first systematic assessment as to whether the common design of orientation domains could result from an inherently random process , as realized through the local feedforward structure of the early visual pathway rather than an optimization process coordinated on large scales . Given the disagreement between the layouts predicted by the statistical wiring model and the data , global circuit optimization as proposed by the long-range interaction model currently is the only theory known to be capable of explaining the common design of orientation domains in the primary visual cortex .
Column spacing and pinwheel statistics of both data and simulation were analyzed using the wavelet method introduced in [47 , 119] . This method specifically takes into account that experimentally measured domain layouts often exhibit local variations in column spacing ( as opposed to most model layouts ) and is thus well-suited to unbiasedly compare the pinwheel layouts of model layouts and experimental data . Matlab source code for preprocessing of experimentally obtained layouts , column spacing analysis , and the analysis of pinwheel layouts can be found in the Supplemental Material , along with four example cases from ferret V1 to test the code . The full data set used in the present study is available on the neural data sharing platform http://www . g-node . org/ . For comparison between model orientation domain layouts and experimentally obtained layouts , both were analyzed with the exact same wavelet parameters settings . Raw difference images obtained in the experiments were Fermi bandpass filtered as described in [23] . Filter parameters were adapted to the column spacings of the different species such that structures on the relevant scales were only weakly attenuated ( see [23] ) . To determine the local column spacing of the layouts , we first calculated wavelet coefficients of an image I ( x ) , averaged over all orientations Ψ ( y , Λ ) = ∫ d φ π ∫ d 2 x I ( x ) · ϕ y ( x , Λ , φ ) ( 18 ) where y is the position , φ the orientation and Λ the scale of the wavelet ϕy ( x , Λ , φ ) . We used complex-valued Morlet wavelets composed of a Gaussian envelope and a plane wave ϕ ( x ) = 1 σ exp - x 2 2 σ 2 · exp ( i k ϕ x ) ( 19 ) and ϕ y ( x , Λ , φ ) = ϕ ( Ω - 1 ( φ ) ( y - x ) ) . ( 20 ) The matrix Ω ( φ ) is the two dimensional rotation matrix ( Eq ( 25 ) ) . To compute the wavelet orientation average in Eq ( 18 ) , 16 equally spaced wavelet orientations were used . For a given Λ , the parameters of the Morlet wavelet were chosen as k ϕ = 2 π Λ 1 0 ( 21 ) σ = ξ Λ 2 π . ( 22 ) ξ determines the size of the wavelet and was chosen to be ξ = 7 , as in [23] . This captures column spacing variations on scales larger than 4 hypercolumns while at the same time enabling robust column spacing estimation . To obtain the map of local column spacing Λlocal ( y ) , we calculated the scale Λ with the largest wavelet coefficient Λ local ( y ) = argmax Λ Ψ ( y , Λ ) ( 23 ) for every position y . To estimate the pinwheel density and other pinwheel layout parameters , we used a fully automated procedure proposed in [23] . We refer to the Supplemental Material accompanying [23] for further details . Here , we examine the analytically most tractable variant of the statistical wiring model , in which ON and OFF center cells are localized on perfectly hexagonal lattices L , ( Fig 2B ) , that may exhibit different lattice constants r and r′ , L = 1 0 k + 1 2 1 3 l f ∀ k , l ∈ Z , ( 24 ) where f = r , r′ is the lattice constant . Describing a rotation of the lattice vectors by the rotation matrix Ω ( α ) = cos ( α ) - sin ( α ) sin ( α ) cos ( α ) , ( 25 ) the ON mosaic is rotated by an angle α , the OFF lattice by an angle α′ ( Fig 2B ) . Paik & Ringach found in numerical simulations that in this case , Moiré interference between two hexagonal RGC mosaics results in a hexagonal layout of orientation domains [28 , 29 , 83] . We now first derive an explicit expression for cortical receptive fields RFy spatially varying with y predicted by the model . Calculating the preferred orientation of these receptive fields then provides an explicit expression for the hexagonal domain layouts . The sum in Eq ( 3 ) can be evaluated analytically for rectangular lattices using Jacobi Theta functions [120] . To solve the Moiré interference model , we used the fact that every hexagonal lattice can be written as sum L = L 1 + L 2 of two rectangular lattices with orthogonal base vectors by separating even and odd numbers in l and shifting the l-sum so that the x-component is equal to zero . The two rectangular lattices are L 1 = 1 0 k + 0 3 l f ∀ k , l ∈ Z , L 2 = 1 0 k + 0 3 l + 1 2 1 3 f ∀ k , l ∈ Z . ( 26 ) Evaluating the infinite Gaussian sum ( Eq ( 3 ) ) yields the result for a single sub lattice ( either ON or OFF ) RF α , r , y ON/OFF ( x ) = T Θ 3 b e ϕ , τ Θ 3 b e r , κ + Θ 4 b e ϕ , τ Θ 4 b e r , κ ( 27 ) where b = x σ s 2 + y σ r 2 σ s 2 + σ r 2 ( 28 ) e r = - π r cos ( α ) sin ( α ) ( 29 ) e ϕ = - π 3 r - sin ( α ) cos ( α ) ( 30 ) τ = e - 2 π 2 σ r 2 σ s 2 3 r 2 ( σ r 2 + σ s 2 ) ( 31 ) κ = e - 2 π 2 σ r 2 σ s 2 r 2 ( σ r 2 + σ s 2 ) ( 32 ) T = 2 π σ r 2 σ s 2 3 r 2 ( σ r 2 + σ s 2 ) exp - ( x - y ) 2 2 ( σ s 2 + σ r 2 ) ( 33 ) and Θ3 and Θ4 are the third and fourth Jacobi theta functions [120] . The cortical receptive fields RFy ( x ) are obtained by summing the ON and OFF sublattices RF y ( x ) = RF α , r , y ON ( x ) - RF α ′ , r ′ , y OFF ( x ) , ( 34 ) where α ( α′ ) and r ( r′ ) are the angle and the lattice spacing of the ON ( OFF ) lattice . The inset of a receptive field in Fig 3B shows a plot of Eq ( 34 ) . These receptive fields resemble simple cell receptive fields in V1 with a size of about 1° for our choice of parameters [121 , 122] . An implementation/visualization of the equations for receptive fields can be found in the Supplemental Material . The response Ry of a neuron with receptive field RF ( x ) to a sine wave grating can be calculated using L ( x ) = exp ( −ik x ) as a stimulus in Eq ( 5 ) . Evaluating the integral then corresponds to Fourier transforming the receptive field RF ( x ) . Denoting the Fourier transform of the receptive field as R y ( k ) = 1 2 π ∫ d 2 x RF y ( x ) e - i k x , ( 35 ) we refer to the absolute value | R y ( k ) | as the amplitude spectrum of the receptive field . Given the above definition , the amplitude spectrum represents the response to a sine wave grating with wave vector k = ( k cos ( ϑ ) , k cos ( ϑ ) ) , where ϑ is the grating orientation and k its spatial frequency . A tuning curve for spatial frequencies k and orientations ϑ is given by TC ( ϑ , k ) = | R y ( k cos ( ϑ ) , k sin ( ϑ ) ) | . ( 36 ) We calculated the Fourier transform of Eq ( 34 ) by transforming Eq ( 2 ) and subsequently summing over the two rectangular lattices L 1 and L 2 in Eq ( 26 ) . Interchanging summation and integration is valid because all infinite sums are uniformly convergent . The result is R ( k ) α , r , y ON/OFF = U Θ 3 ( c e ϕ , ν ) Θ 3 ( c e r , ζ ) + Θ 4 ( c e ϕ , ν ) Θ 4 ( c e r , ζ ) , ( 37 ) where U = 2 π σ r 2 σ s 2 3 r 2 exp - i k y - 1 2 k 2 σ s 2 + σ r 2 ( 38 ) c = ( y - i σ s 2 k ) ( 39 ) ν = exp - 2 π 2 σ s 2 3 r 2 ( 40 ) ζ = exp - 2 π 2 σ s 2 r 2 . ( 41 ) The Fourier transform of cortical receptive fields is given by the sum of the ON and OFF sublattice Fourier transforms R y ( k ) = R ( k ) α , r , y ON - R ( k ) α ′ , r ′ , y OFF , ( 42 ) where we suppressed the dependencies on α , α′ , r , r′ on the left hand side . Receptive fields depend on the two scales σr and σs . With increasing σs , more RGCs are pooled to form the cortical receptive field . If the cortical receptive field is dominated by more than two RGCs , it can exhibit multiple ON and OFF subregions . The spatial arrangement of these ON and OFF subregion mirrors the hexagonal lattices of the ON and OFF center RGCs . The parameter σs must be of a minimal size since for very small values of many cortical cells are connected with only a single , dominant RGC input and exhibit no orientation selectivity . Varying σr does not qualitatively change the shape of cortical receptive fields . For simple cell receptive fields with one ON and one OFF subregion , the amplitude spectrum | R ( k ) | will typically look as in Fig 10A . We follow [28 , 82] and define the preferred angle as ϑpref: = arg ( μ ) /2 , where μ = ∫ d 2 k | R ( k ) | · e 2 i arg ( k ) k ∫ d 2 k | R ( k ) | . ( 43 ) While there is consensus about the definition of the preferred orientation , methods for extracting the preferred spatial frequency differ within the literature . Ringach proposed to use kpref = ∣μ∣ [82] , referred to as Center-of-Mass Method . More commonly , the circular variance CV ( k ) [123–125] CV ( k ) = ∫ 0 2 π d ϑ TC ( ϑ , k ) e 2 i ϑ ∫ 0 2 π d ϑ TC ( ϑ , k ) , ( 44 ) is first computed as a measure of orientation selectivity at a given spatial frequency k . Maximizing the circular variance across all spatial frequencies is then performed to obtain an estimate of preferred spatial frequency: k pref = argmax { k } CV ( k ) . ( 45 ) We refer to this method as CV maximization . Finally , one can use the maximum of the amplitude spectrum k pref = argmax { k } | R ( k ) | , ( 46 ) as an estimate of the preferred spatial frequency ( Maximum method ) . We argue that Maximum method and CV maximization in most cases yield similar results . They extract preferred spatial frequencies that one would obtain when searching for the “strongest response” by presenting a set of gratings of varying orientation and spatial frequency to a subject [125–127] . In contrast , estimates made with the center-of-mass method are usually substantially smaller then this intuitive measure ( Fig 10C and 10D ) . As a consequence , the orientation selectivity index defined as OSI = ∫ 0 2 π d ϑ TC ( ϑ , k pref ) e 2 i ϑ ∫ 0 2 π d ϑ TC ( ϑ , k pref ) ( 47 ) for all three methods , will usually be substantially smaller , when estimated with the Center-of-Mass method compared to the other two methods . Among all three methods , the Maximum method has the advantage that its estimates are unaffected by monotonic nonlinearities applied to R commonly used to convert it to a firing rate of a neuron . For this reason , the Maximum method is our method of choice for extracting the preferred spatial frequency from amplitude spectra of receptive fields . Using the equations for the receptive fields of cortical neurons in the Moiré interference model , we extracted the spatial progression preferred orientation and spatial frequency from their squared amplitude spectrum: | ℛ y ( k ) | 2 ∝ exp ( − k 2 ( σ s 2 + σ r 2 ) ) · | Θ 3 α , r Θ 3 α , r + Θ 4 α , r Θ 4 α , r − Θ 3 α ′ , r ′ Θ 3 α ′ , r ′ − Θ 4 α ′ , r ′ Θ 4 α ′ , r ′ ︸ G y ( k ) | 2 , ( 48 ) with abbreviation Θ i α , r Θ i α , r = Θ i ( c e ϕ ( α , r ) , τ ) Θ i ( c e r ( α , r ) , ζ ) . | R ( k ) | 2 is composed of a rotationally symmetric Gaussian envelope and a non-rotationally symmetric part Gy ( k ) varying in space y . To calculate the preferred orientation ϑpref and spatial frequency kpref , we expanded the non-rotationally symmetric part ∣Gy ( ( k ) ∣2 to quadratic order in k: | R y ( k ) | 2 ≈ exp - k 2 ( σ s 2 + σ r 2 ) | G y 0 | 2 + 1 2 k 1 k 2 H y k 1 k 2 , ( 49 ) where H y is the Hessian matrix H y = ∂ 2 | G y | 2 ∂ k 1 2 ∂ 2 | G y | 2 ∂ k 1 ∂ k 2 ∂ 2 | G y | 2 ∂ k 2 ∂ k 1 ∂ 2 | G y | 2 ∂ k 2 2 k = 0 ≡ a y b y b y c y ( 50 ) and G y 0 = G y ( k = 0 ) . Since Gy ( k ) = Gy ( −k ) , this Taylor expansion only contains terms of even power in k . Using the fact that h ( θ ) = cos θ sin θ H y cos θ sin θ = a y cos 2 θ + 2 b y cos θ sin θ + c y sin 2 θ ( 51 ) yields the second directional derivative in the direction of ( cos θ , sin θ ) , ϑpref can be found as the maximum of h ( θ ) , i . e . the direction of largest increase in amplitude spectrum , ϑ pref ( y ) = atan ( a y - c y ) 2 + 4 b y 2 - a y + c y 2 b y . ( 52 ) This formula for ϑpref ( y ) represents an expression for the orientation domain layouts produced by the Moiré interference model for hexagonal RGC lattices . To calculate the smoothed domain layout of the Moiré interference model analytically , we identified the low frequency components of our analytical solution . To this end , we expanded the Jacobi theta functions in Eq ( 42 ) [120] | R y ( k ) | 2 = ∑ j C y j exp - 1 2 σ 2 ( k - a y j ) 2 + C y j exp - 1 2 σ 2 ( k + a y j ) 2 ( 53 ) where C y j and a y j are determined by Eq ( 42 ) . According to this equation , the power spectra of receptive fields in the Moiré interference model is represented by an infinite sum of Gaussians , each mirrored at the origin ( 0 , 0 ) of Fourier space . The preferred orientation of a receptive field represented by such an infinite sum is set by the direction in which the “center-of-mass” of the Gaussians is located . Due to the symmetry of the power under spatial inversion , there are two peaks located at ϑ ( y ) and π + ϑ ( y ) . The direction towards the center-of-mass of the peak is obtained through the complex number μ ( y ) = ∑ j C y j · | a y j | exp ( 2 i arg ( a y j ) ) ( 54 ) with C y j and a y j defined as in Eq ( 53 ) . The preferred orientation then is arg ( μ ( y ) ) /2 . Rewriting this sum and substituting the respective expressions for C y j and a y j , we obtained μ ( y ) = ∑ m , n , o , p f ( m , n , o , p ) exp ( 2 i y ( n e ϕ + m e r - o e r ′ - p e ϕ ′ ) ) , ( 55 ) with coefficients f ( m , n , o , p ) . This is a decomposition of the orientation domain layout of the Moiré interference model into Fourier modes , indexed by four numbers m , n , o , p = 0 , ±1 , ±2 , … . Table 2 lists the first terms of this series in ascending spatial frequency order . By rewriting μ → z and selecting only the lowest contributing spatial frequencies , we obtain Eq ( 7 ) . The phase factor u 0 = e i ( α + α ′ ) e i α ′ r + e i α r ′ e i α r + e i α ′ r ′ ( 56 ) is associated with an overall shift of all preferred orientations . Note that ∣u0∣2 = 1 . Realization of a random distortion field y ( x ) were generated by finding a complex-valued field z ( x ) of which real and imaginary part correspond to dislocations in x and y direction , respectively . We constructed such a field using established methods ( e . g . [51] ) in the Fourier domain . In short , we drew complex-valued amplitudes a ( k ) from a Gaussian distribution satisfying 〈 a ( k ) a ( k ′ ) 〉 = f ˜ ( k ) · δ k , k ′ , where f ˜ ( k ) was a chosen power spectrum , in our case a Gaussian with width 1/σ , σ being the desired correlation length . The corresponding amplitude spectrum was then inversely Fourier transformed to obtain a complex-valued field z ( x ) . Real and imaginary part of this field constitute two independent real-valued Gaussian random fields , both with the desired spatial statistics . We then transformed the coordinates of the hexagonal ON/OFF lattice points ri = ( xi , yi ) according to x i → x i + η ℜ ( z ( r i ) ) and y i → y i + η ℑ ( z ( r i ) ) . For the displacements of ON and OFF lattices , we used two independent complex-valued Gaussian random field realizations . Source code to generate hexagonal RGC mosaics with correlated spatial noise along with Matlab code for visualization is part of the supplementary material to this manuscript . We generated RGC mosaics with a pairwise interacting point process using the code published by Schottdorf et al . [32] derived from the method developed in [49 , 82] . | In the primary visual cortex of primates and carnivores , local visual stimulus features such as edge orientation are processed by neurons arranged in arrays of iso-orientation domains . Large-scale comparative studies have uncovered that the spatial layout of these domains and their topological defects follows species-invariant quantitative laws , predicted by models of large-scale circuit self-organization . Here , we ask whether the experimentally observed layout invariants might alternatively emerge as a consequence of random connectivity rules for feedforward projections from a small number of retinal cells to a much larger number of cortical target neurons . In this random wiring framework , the semi-regular and spatially granular arrangement of retinal ganglion cells determines the spatial layout of visual cortical iso-orientation domains—a hypothesis diametrically opposed to cortical large-scale circuit self-organization . Generalizing a prominent model of the early visual pathway , we find that the random wiring framework does not reproduce the experimentally determined layout invariants . Our results demonstrate how comparison between theory and quantitative phenomenological laws obtained from large-scale experimental data can successfully discriminate between competing hypotheses about the design principles of cortical circuits . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| []
| 2015 | Random Wiring, Ganglion Cell Mosaics, and the Functional Architecture of the Visual Cortex |
Dengue virus ( DENV ) infection of an individual human or mosquito host produces a dynamic population of closely-related sequences . This intra-host genetic diversity is thought to offer an advantage for arboviruses to adapt as they cycle between two very different host species , but it remains poorly characterized . To track changes in viral intra-host genetic diversity during horizontal transmission , we infected Aedes aegypti mosquitoes by allowing them to feed on DENV2-infected patients . We then performed whole-genome deep-sequencing of human- and matched mosquito-derived DENV samples on the Illumina platform and used a sensitive variant-caller to detect single nucleotide variants ( SNVs ) within each sample . >90% of SNVs were lost upon transition from human to mosquito , as well as from mosquito abdomen to salivary glands . Levels of viral diversity were maintained , however , by the regeneration of new SNVs at each stage of transmission . We further show that SNVs maintained across transmission stages were transmitted as a unit of two at maximum , suggesting the presence of numerous variant genomes carrying only one or two SNVs each . We also present evidence for differences in selection pressures between human and mosquito hosts , particularly on the structural and NS1 genes . This analysis provides insights into how population drops during transmission shape RNA virus genetic diversity , has direct implications for virus evolution , and illustrates the value of high-coverage , whole-genome next-generation sequencing for understanding viral intra-host genetic diversity .
With 3 . 6 billion people at risk and nearly 400 million infections annually [1 , 2] , dengue has become the most important mosquito-borne viral disease affecting humans today . Dengue virus ( DENV ) is a positive-sense , single-stranded RNA virus of the family Flaviviridae , genus Flavivirus . The ~10 . 7 kb DENV genome encodes three structural proteins ( capsid [C] , premembrane [prM] , and envelope [E] ) and seven non-structural ( NS ) proteins ( NS1 , NS2A , NS2B , NS3 , NS4A , NS4B , and NS5 ) . DENV is transmitted between humans by the mosquitoes Aedes aegypti and Aedes albopictus . which acquire the virus by taking a bloodmeal from an infected human . Once ingested , DENV first infects and replicates in the mosquito midgut epithelium . It subsequently disseminates through the hemolymph to infect other organs such as the fat body and trachea , finally reaching the salivary glands , where it is secreted into mosquito saliva and injected into a human host during a subsequent blood-feeding event [3] . As is typically the case for RNA viruses , the DENV RNA-dependent RNA polymerase ( RdRp , encoded by DENV NS5 ) operates at low fidelity , resulting in the accumulation of a population of closely-related but genetically-distinct viral genomes , organized around a consensus sequence , within each individual human or mosquito host . Sometimes termed a quasispecies , these intra-host variants are thought to interact cooperatively on a functional level , and to collectively contribute to the overall fitness of the virus population ( reviewed in [4] ) . This has important implications for viral pathogenesis: High fidelity poliovirus mutants , for example , are highly attenuated , demonstrating the importance of genetic diversity and the ability to adapt to changing environments during infection [5 , 6] . Adaptability is especially important for mosquito-borne viruses , which encounter distinct selection pressures when cycling between vertebrate and invertebrate hosts . Even within a single host , intra-host variants generated through multiple rounds of virus replication are subject to evolutionary mechanisms such as bottlenecks , drift , and positive and negative selection pressures . In human-derived DENV populations , highly immunogenic E protein domains also display higher levels of intra-host genetic diversity , suggesting that selection pressures on low frequency population variants operate even during acute infection [7] . In mosquitoes , RNA interference ( RNAi ) , a key antiviral defense mechanism in insects , has been proposed to be a driver of viral intra-host genetic diversity in the Culex-West Nile virus ( WNV , family Flaviviridae ) system [8] , where higher intra-host diversity levels have been reported in the mosquito than in vertebrate hosts [9 , 10] . Host alternation also subjects arboviruses to frequent and significant drops in population size . Within the mosquito and vertebrate hosts , these drops occur during initial establishment of infection and subsequent spread through various tissues and organs , as well as during the process of blood-feeding itself , where only a small percentage of the total virus population circulating in the human actually seeds infection in the mosquito host . It is unclear how such drops shape the diversity and repertoires of intra-host virus populations , and the potential effect they have on transmission . Studies on intra-host genetic diversity in DENV have focused on virus populations in the human , and with few exceptions [7] , have been confined to the sequencing of one or two viral genes [11–14] . It is unclear , however , what happens to intra-host diversity during replication in the mosquito , and particularly after the virus population has gone through a significant drop , such as occurs when a mosquito feeds on a dengue patient that has a viremia around the 50% mosquito infectious dose ( MID50 ) . In this study , we used whole-genome amplification and next-generation Illumina sequencing to characterize DENV intra-host genetic diversity in both patient- and matched mosquito-derived virus populations . Because mosquitoes in our study were infected by allowing them to feed directly on patients , we were able to track changes in viral populations during human-to-mosquito transmission . We used a rigorous variant-calling algorithm to identify low frequency single nucleotide variants ( SNVs ) within a viral population . This unique study design provides important insights into how population drops shape viral genetic diversity , as well as into the differing selection pressures between human and insect hosts .
Study protocols were reviewed and approved by the Scientific and Ethical Committee of the Hospital for Tropical Diseases ( CS/ND/09/24 ) and the Oxford Tropical Research Ethical Committee ( OxTREC 20–09 ) . Written informed consent was obtained from all participants or from the parent or guardian of any child participants on their behalf . The human viremic plasma and DENV2-infected mosquitoes used here originated from the study by Nguyen et al . [15] , in which field-derived Ae . aegypti mosquitoes were allowed to feed on acute dengue cases . Just prior to blood feeding , venous blood was drawn and the plasma fraction stored . Blood-fed mosquitoes were reared until day 14 when the abdomen and salivary glands were dissected and stored . Out of the 83 DENV2-infected patients described by Nguyen et al . [15] , we selected 12 patient plasma samples and 36 paired blood fed mosquitoes ( three per patient ) for investigation of DENV sequence diversity in this study . The rationale for these patient-mosquito pairs was that six of them involved mosquito cohorts that blood fed when the plasma viremia was near to the calculated MID50 ( S1 Fig ) [15] . Hence the seeding of initial infection in these mosquitoes was most likely by a small or very small population of infectious virions , i . e . a major drop in DENV population size moving from human blood to mosquito . Nucleic acid was extracted from plasma samples and from mosquito tissues for quantification of the DENV RNA concentration as described previously [15] . Next-generation whole-genome sequencing of human- and mosquito-derived DENV samples was performed as described in [16] . Viral RNA was extracted from human plasma or mosquito abdomen / salivary gland samples using the QIAamp Viral RNA Mini Kit ( Qiagen ) , and cDNA synthesis was carried out with the Maxima H Minus First Strand cDNA Synthesis Kit ( ThermoScientific ) using a single primer designed to bind to the 3' end of the viral genome . The entire DENV genome was PCR-amplified in 14 overlapping fragments , each about 2kb in length , using the PfuUltra II Fusion HS DNA Polymerase ( Agilent Technologies ) . Primer sequences can be found in S5 Table . PCR products were run on an agarose gel and purified using the Qiagen Gel Extraction Kit ( Qiagen ) . For each sample , equal amounts of all PCR-amplified fragments were combined and sheared on the Covaris S2 sonicator ( Covaris ) to achieve a peak size range of 100–300 bp ( shearing conditions: duty cycle—10%; intensity—5; cycles per burst—200; time—110 seconds ) . Samples were purified with the Qiagen PCR Purification Kit ( Qiagen ) and quality-checked on the Agilent 2100 Bioanalyzer with a DNA 1000 Chip ( Agilent Technologies ) . Library preparation was performed with the KAPA Library Preparation Kit ( KAPA Biosciences ) . After end-repair , A-tailing , and adapter ligation , ligated products in the 200–400 bp range were gel-extracted with the Qiagen Gel Extraction Kit ( Qiagen ) . Samples were subjected to 14 PCR cycles to incorporate multiplexing indices , quantified with the KAPA SYBR FAST qPCR Master Mix ( KAPA Biosciences ) on the LightCycler 480 II real-time thermocycler ( Roche Applied Science ) , and pooled . Paired-end , multiplexed sequencing ( 2 x 101 bp reads ) of libraries was performed on the Illumina HiSeq ( Illumina ) at the Genome Institute of Singapore . Base calling was done with CASAVA 1 . 7; reads that did not pass Illumina’s chastity filter ( CASAVA 1 . 7 user guide ) were removed . Illumina-generated FASTQ files were put through the Viral Pipeline Runner ( ViPR , available at https://github . com/CSB5/vipr ) , which automates the following steps: Iterative mapping of reads against the DENV-2 reference sequence NC_001474 using the Burrows-Wheeler Aligner [17] , generation of a consensus sequence , and calling of low frequency single nucleotide variants ( SNVs ) against the consensus using the LoFreq algorithm [18] . SNVs differ from single nucleotide polymorphisms ( SNPs ) in that the latter occur between the consensus sequences of different individuals . LoFreq has previously been applied to DENV datasets , and its SNV predictions on these datasets have been experimentally validated down to 0 . 5% frequency [18] . To increase specificity and keep only conservative predictions , SNVs that fulfilled any of the following criteria were discarded: located within primer sequences , located adjacent to known homopolymer regions , coverage of <1000X , frequency of <0 . 01 ( 1% ) . Patient DENV clinical samples from the Early DENgue study ( EDEN ) [19] were previously assigned into human-human transmission pairs thought to be separated by a single mosquito [20] . This was based on Sanger sequence similarity , distance between physical addresses , and time between dates of fever onset [20]; the spatiotemporal criteria are consistent with the National Environment Agency's criteria for active transmission [21] . Whole-genome Illumina sequencing and SNV calling for these samples was carried out as described above . To identify mutational hotspots , a scanning window approach was used to scan the DENV genome ( window size of 20 , overlap of five nucleotides ) for an excess of SNVs in a window compared with the genome-wide average ( binomial test; Bonferroni-corrected p-value < 0 . 05 ) . This was done on a per-sample basis . For coldspots , SNVs from all samples were pooled and scanned for windows ( minimum size of 40 ) with a depletion of SNVs ( binomial test; Bonferroni-corrected p-value < 0 . 05 ) [18] . Since haplotype reconstruction is a notoriously difficult problem , we resorted to a simple approach that estimates a lower bound of the number of haplotypes ( viral genomes ) present in a sample . This was done by greedily clustering SNVs based on their allele frequency confidence intervals ( Agresti-Coull at the 0 . 05 level ) . SNVs were sorted by their allele frequency and the SNV with highest allele frequency seeds the first cluster . Variants were added to an existing cluster if their upper confidence interval limit was greater than the cluster minimum , otherwise they form a new cluster . The number of clusters represents a lower bound of the number of distinct viral genomes present in a sample . This clustering approach is implemented as one of the tools that come bundled with LoFreq [18] . The size of the infecting virus population was estimated by simulating 1000 samplings of varying size from the virus population in the human , following a normal distribution with mean SNV frequency matching that in the human . This sampling was carried out for a range of mean SNV frequencies . The sampling error rate was calculated from the simulated SNV frequencies , and p-values were computed using a two-tailed t-distribution ( S3 Table ) .
Nguyen et al . [15] infected field-derived Ae . aegypti mosquitoes with DENV2 by allowing them to blood-feed directly on patients at the acute stage of infection . To track changes in viral intra-population genetic diversity during human-to-mosquito transmission , we performed whole-genome Illumina sequencing of DENV2 populations from 12 patient plasma samples and 36 infected mosquitoes ( three per patient ) derived from that study [15] . We successfully obtained genome-length DENV2 sequences from all 12 patient plasma samples , as well as from 25 mosquitoes ( 21 abdomen-derived and 25 salivary-gland-derived ) . We were unable to PCR-amplify virus from the remaining 11 mosquito samples . Successfully sequenced patient and mosquito samples are described in S1 Table . Single nucleotide variants ( SNVs ) in each DENV2 population were called with the LoFreq variant calling algorithm [18] . At a conservative minimum cutoff of 1% frequency , we observed a total of 1 , 116 SNVs across all human- and mosquito-derived sequence sets . These were evenly distributed across the ~10 . 7 kb DENV2 genome ( S2 Fig ) . We examined four measures of intra-sample diversity , calculated on a per sample basis [7]: a ) the number of SNVs , b ) the sum of SNV frequencies , c ) the average SNV frequency , and d ) the standard error of the mean ( SEM ) SNV frequency ( Fig 1 ) . These measures capture different aspects of diversity: the number of SNVs indicates how many distinct variant positions there are along the viral genome; the sum and mean of SNV frequencies indicate how commonly these variants occur and whether they are distributed across many or a few positions; the SEM indicates whether sample diversity comes from dominant variant genomes , minor variant genomes , or a mix of both . All four diversity measures varied widely among individual samples ( Fig 1A–1D ) . No statistically significant differences in any of these four measures were observed between DENV populations from human plasma , mosquito abdomens and mosquito salivary glands ( p > 0 . 05 , Kruskal-Wallis test ) ( Fig 1A–1D ) , indicating that intra-sample diversity levels remain similar during replication in human and mosquito hosts . Although we cannot experimentally follow transmission of DENV2 from mosquitoes into humans , we had access to patient plasma samples from the Early DENgue ( EDEN ) study carried out in Singapore from 2005–2007 [19] . DENV1 and DENV3 samples from this study were previously assigned to human-human transmission pairs thought to be separated by a single mosquito , based on Sanger sequence similarity and spatiotemporal relationships ( distance between patients' homes and time between dates of fever onset ) between members of a pair [20] . Illumina deep sequencing of these samples revealed that measures of intra-sample diversity also did not change significantly between members of a human-human transmission pair ( S3A–S3D Fig ) , suggesting that passage through a mosquito does not affect intra-host viral diversity levels upon subsequent replication in the human . Although levels of intra-host diversity remain unchanged , horizontal transmission may alter the SNV repertoire in a viral population if existing SNVs are lost and new ones generated . To examine this , we tracked individual SNVs during transmission from human to mosquito abdomen to mosquito salivary glands ( Fig 2 ) . Of 267 SNVs present in human plasma-derived DENV2 populations , only 26 ( 9 . 7% ) were observed again in any of the associated mosquitoes ( abdomen , salivary glands , or both ) , and only 37 of 478 SNVs present in the mosquito abdomen ( 7 . 7% ) were observed again in the salivary gland ( Fig 2A ) . This indicates that viral populations are able to quickly restore their diversity upon replication in a new compartment , but predominantly with a very different SNV repertoire that most likely arises through random mutation . It is possible that SNVs not observed in the mosquito abdomen or salivary gland had dropped below our conservatively-chosen level of detection ( frequency < 1% ) . We observed relatively few instances of SNVs being maintained across transmission stages ( Fig 2B ) . These SNVs were present at a significantly higher frequency than those that were lost ( Fig 3A and 3B ) , suggesting that higher frequency SNVs are more likely to be detectably transmitted . Maintained SNVs were never seen in more than one human-mosquito group , but were observed in more than one mosquito per human approximately half the time ( 12 out of 26 cases; in 8 of these cases , SNVs were maintained in all three mosquitoes that had fed on the same human ) ( Table 1 ) . No significant difference in frequency was observed between SNVs maintained in one mosquito and SNVs maintained in more than one mosquito ( Fig 3C ) . SNVs may also be maintained because they play roles in increasing virus fitness , and selection for these SNVs may result in an increase in their frequencies over the course of transmission . We did not , however , observe this—the frequencies of most maintained SNVs ( both synonymous and non-synonymous ) remained similar in human- , mosquito abdomen- , and mosquito salivary gland-derived DENV populations ( Fig 4 ) , suggesting that most SNVs are neutral in the mosquito . There were several exceptions—for example , the frequency of 3989 T>C ( NS2A ) nearly doubles from 0 . 11 in the human to 0 . 20 in the mosquito , the frequency of 1745 T>C ( E ) nearly halves from 0 . 26 in the human to 0 . 16 in two separate mosquitoes , and the frequency of 507 T>C ( M ) increases 8-fold from 0 . 02 in the human to 0 . 16 in the mosquito abdomen , but drops back to 0 . 07 in the salivary gland ( Fig 4 ) . It is possible that these more dramatic changes in SNV frequency may be due to selection for fitter variants , but it is difficult to differentiate this from stochasticity resulting from drops in population size . In the EDEN human-human transmission pairs , 15 of 130 SNVs ( 11 . 5% ) present in the first member of the pair were observed again in the second member ( Fig 2C and Table 2 ) . The majority of these SNVs were from a single pair , where 12 out of 23 SNVs ( 52 . 2% ) in the first member were maintained ( S2 Table ) . Thus , similar to what we observed for human-mosquito pairs , linked human-human pairs of DENV clinical samples also yielded predominantly different SNV repertoires , although there may be exceptions , as observed here . While the majority of SNVs may drop below our detection limit during passage through the mosquito , it is possible that , once transmitted back into the human , selection pressures there may again drive up the frequencies of SNVs that impact virus fitness . Given that ~90% of SNVs detected in the human were lost , presumably due to chance , during transmission to the mosquito , we wanted to estimate the size of the population drop at which this would occur . For an assumed range of plasma viremia levels from 102 to 107 infectious virions / ml , we estimated the maximum size of the virus population infecting the mosquito for which loss of SNVs could be attributed to random chance . This was done by simulating 1000 samplings of varying sizes from the viral population in the human . For SNVs with frequencies from 0 . 01 to 0 . 03 ( the vast majority of SNVs in our dataset ) , the maximum infecting virus population size for SNV loss to be attributable to chance was 100 virions ( S3 Table ) . This represents the number of virions that establish a productive infection in the mosquito midgut . Since a 2 μl bloodmeal [22 , 23] taken from a patient with plasma viremia equivalent to the MID50 of 2x106 RNA copies / ml [15] would be expected to contain 4000 RNA copies , this estimate suggests that the viral population size actually infecting the mosquito midgut is reduced at least 40X , probably due to the bulk of viral genomes being non-infectious . Low frequency DENV variants that arise during virus replication are subject to immune selection pressure . To examine differences in selection pressures between hosts , we compared the ratio of the number of non-synonymous to synonymous SNVs ( #NS/#S ) for each gene in human- versus mosquito-derived virus populations . Because of the small numbers of human plasma-derived DENV2 SNVs from this study , we pooled these with SNVs detected in DENV1 and DENV3 populations from the EDEN study plasma samples [19] . We are aware that selection pressures may differ across DENV serotypes , but this approach allowed us to broadly compare the vertebrate and invertebrate hosts . Human #NS/#S ratios were significantly higher than mosquito ratios for the prM , E , and NS1 genes ( Fig 5A ) . While non-significant , the ratio for C in the human was nearly double that in the mosquito ( 2 . 17 versus 1 . 18 ) . It is striking that the products of these genes are actively targeted by the human antibody response ( reviewed in [24] ) , while no antibody response exists in the mosquito immune repertoire . Although these ratios did not reach the levels formally required for positive selection , we note that positive selection is usually measured at consensus level over time scales much longer than a single transmission [25] . We are instead interested in selection pressures acting on low frequency variants generated over multiple virus replication cycles in each host . We found no significant differences in #NS/#S ratios between mosquito abdomen and salivary gland for any of the viral genes ( Fig 5B ) , suggesting that host might be more important than body compartment . The much smaller number of SNVs in these two datasets , however , makes it difficult to rule out distinct selection pressures acting in separate mosquito tissues . We also used a scanning window approach to identify mutational hot- and coldspots in the DENV genome; i . e . regions containing a statistically significant excess or lack of SNVs compared with the genome-wide average ( Fig 6 ) . We identified a hotspot in E , the gene encoding the DENV envelope protein , in a single human-derived sample from Vietnam , but not in any of the mosquito-derived samples ( Fig 6 ) . An E hotspot was also found in a DENV3 human-derived sample from the EDEN study ( S4B Fig ) . These hotspots were located in E domains ( ED ) II and I respectively , which are known targets of the antibody response [24] . Taken together with our finding that the E #NS/#S ratio is significantly higher in humans compared to mosquitoes ( Fig 5A ) , we speculate that immune pressure on the E protein may play a bigger role in generating diversity in the DENV population in humans than in mosquitoes . Hotspots in NS3 , which encodes the viral serine protease and helicase , were detected in two mosquito-derived samples . While not significantly different between humans and mosquitoes ( Fig 5A ) , the NS3 #NS/#S ratio in mosquitoes was significantly higher than the average across all viral genes ( p = 0 . 027 , exact binomial test ) , suggesting that NS3 may experience a stronger selection pressure than other viral genes in the mosquito . We also identified a 3'UTR hotspot in one mosquito-derived sample . Adaptation to mosquitoes has been proposed as the major driver of evolution in the 3'UTR of chikungunya virus ( CHIKV ) [27] , an alphavirus of the family Togaviridae; it will be intriguing to test this hypothesis in DENV , a flavivirus . The coldspot in prM detected in mosquito-derived samples corresponds to a conserved region of the gene , which has been reported to be important for the association of prM with E during viral assembly [18 , 28] . We also detected a coldspot in NS5 in mosquito-derived samples , which spans the junction between the methyltransferase and polymerase domains of the protein . The lack of SNVs in these regions suggests the presence of functionally important residues . 27 SNVs ( 2 . 4% of all SNVs ) were found to encode low frequency premature stop codon mutations . These were spread across the coding region of the DENV genome in both human- and mosquito-derived DENV populations ( S4 Table ) . Previous studies have proposed that defective RNA viruses can be transmitted through complementation by co-infection of host cells with functional virus [29 , 30]; however in our dataset we did not observe transmission of any of these SNVs . To determine how SNVs were distributed into distinct viral genomes , we clustered SNVs in each sample based on frequency , reasoning that SNVs present on the same viral genome would be present at similar frequencies . This gave us an estimate of the minimum number of distinct variant viral genomes in each sample ( Fig 7A ) ; the actual number is likely to be higher in many samples , since different genomes could be present at the same frequency , and lower frequency SNVs tend to cluster together without good separation . Similar to the diversity measures in Fig 1 , we found no significant difference in the minimum number of distinct variant genomes between virus populations derived from humans , mosquito abdomens , and mosquito salivary glands ( p > 0 . 05 , Kruskal-Wallis test ) ( Fig 7A ) . We next used network diagrams to visualize transmission patterns of viral genomes ( Fig 7B ) . In this example ( patient-mosquito group 641 ) , no consensus changes were observed as the virus was transmitted from human to mosquito abdomen to mosquito salivary gland . A maximum of only one SNV from each variant genome ( shown as circles radiating out of the consensus sequence ) was maintained across transmission stages ( Fig 7B ) . This situation was similar across all our other patient-mosquito groups , where it was most common for only one SNV per variant genome to be maintained across transmission stages , with a maximum of two SNVs being maintained as a unit . This suggests that SNVs in each sample were spread out across many variant genomes instead of being clustered together on one genome , and that the actual number of distinct genomes per sample is higher than our predicted minimum number .
Here we report the use of next-generation , whole-genome DENV sequencing to follow within- and between-host differences in viral populations in naturally-infected humans and their infected Ae . aegypti mosquito counterparts . We observed that the vast majority ( >90% ) of SNVs in a population were lost upon virus transmission from human to mosquito . Based on this , we estimated that a maximum of ~100 virions infect the mosquito midgut , with the actual number varying with the level of viremia in the human . This number is consistent with previous studies that used fluorescent reporter viruses to show that , even with high titer bloodmeals , oral infection with WNV or Venezuelan equine encephalitis virus ( VEEV , family Togaviridae ) resulted in a relatively small number of midgut cells being infected , on the order of 15 for WNV [31] and 100 for VEEV [32] . The VEEV study detected a higher than expected frequency of dually-infected cells , suggesting that only a small percentage of midgut cells are susceptible to infection [32] . A 2 μl bloodmeal from a human host where the DENV plasma viremia is 2x106 RNA copies per ml [15] would be expected to contain 4000 RNA copies . Our estimate suggests that the number of virions that actually infect the mosquito midgut is at least 40X lower . This drop , compatible with a bottleneck scenario , could be due to a number of factors , such as the presence of non-infectious particles , preference for certain cell types , and interference with receptor binding , perhaps by the midgut microbiota or proteins in the bloodmeal . For example , the particle-to-PFU ( plaque-forming unit ) ratio for DENV has been reported to be on the order of 103 to 104: 1 [33 , 34] . High particle-to-PFU ratios are typical for animal viruses [35] , and suggest the presence of many non-infectious particles in a population . Studies on the effect of population drops on arboviral intra-host genetic diversity have yielded varying results . WNV diversity has been reported to decrease during dissemination from the midgut of Culex pipiens to the salivary gland and saliva [36] . Another study reported that WNV genetic diversity was maintained during infection in the mosquito , and , unlike us , observed considerable maintenance of particular variants ( cloned sequences ) from one mosquito body compartment to another . The authors concluded that in the mosquito , population drops did not significantly impact WNV genetic diversity , and that the maintenance of diversity was more likely due to variation in the input virus population rather than the generation of new mutants [37] . In contrast , intra-host diversity in our DENV dataset was maintained through the generation of a new SNV repertoire at each transmission stage , such that population drops in the mosquito abdomen and salivary glands impact viral repertoire but not overall diversity levels . The high error rate ( 10−4 [4] , approximately one mutation per 11 kb DENV genome ) and burst size ( ~103−104 genomes per cell [38 , 39] ) of the virus means that a new array of SNVs is quickly regenerated . It will be interesting to determine if genetic diversity is also maintained in virus populations in mosquito saliva , which this study did not sequence . Infectious DENV loads ( PFUs ) in saliva vary greatly from mosquito to mosquito , but are in general much lower ( ~10-100X ) than in the salivary gland or carcass [40 , 41] , suggesting that the population injected into a human will have gone through yet another drop . However , our observation that genetic diversity also remains similar across members of EDEN human-human transmission pairs suggests that diversity is quickly restored upon replication in the human . We observed that SNVs that were maintained between transmission stages were present at significantly higher frequencies than SNVs that were not maintained . A previous next-generation sequencing analysis of human DENV clinical samples from Nicaragua found that intra-host variants that were also observed at the inter-host level ( i . e . in consensus sequences of circulating Nicaraguan DENV strains ) were present at higher frequencies than variants that were only observed at the intra-host level [7] . This suggests that the initial abundance of a SNV may affect its chances of being maintained and eventually becoming the consensus base . This initial abundance may be subject to stochastic processes; indeed , non-synchronous infection , where early infecting virions have an advantage , has been reported to be a major contributor towards genetic drift in human immunodeficiency virus ( HIV ) populations [42] . We were unable to follow transmission from the mosquito back into the human , but the EDEN human-human transmission pairs ( thought to be separated by a single mosquito ) allowed us to make hypotheses about this process . Similar to the human-mosquito pairs , paired human-human DENV clinical samples also showed predominantly different SNV repertoires . We estimate that the majority of SNVs present in a human sample will fall below our detection limit during passage through the mosquito , where most SNVs appear to be neutral . However , once transmitted to a second human host , selection pressures there may once again drive up the frequencies of SNVs that impact virus fitness in the human . Our observation that several viral genes display higher #NS/#S ratios in human- than in mosquito-derived samples is consistent with this idea , but SNVs would have to be followed over a chain of multiple linked transmissions in order to further investigate this . Over multiple rounds of replication in the human or mosquito , viral populations grow exponentially and generate a range of low frequency variants that are subject to selection pressures . Our data suggest that variants in the prM , E , and NS1 genes experience stronger immune pressures in the human than in the mosquito , consistent with the fact that these gene products are known targets of the human antibody response ( reviewed in [24] ) , which has no equivalent in mosquitoes . The envelope protein E is the main antigen on the virion surface and the target of neutralizing antibody [43 , 44] . Intra-host genetic diversity in the various E protein domains has been reported to positively correlate with immunogenicity in humans , suggesting that immune-driven selection occurs even during short-term , acute dengue infection [7] . We further observed E mutational hotspots in human- but not mosquito-derived DENV populations , leading us to hypothesize that the antibody response in humans is a major driver of viral diversity . The mosquito immune response lacks an adaptive arm , but is able to mount potent responses against virus , bacteria , and fungi through the Toll , IMD , and JAK/STAT immune signaling pathways ( reviewed in [45] ) . Compared to the situation in vertebrates , relatively little is known about the molecular mechanisms by which these pathways act against DENV , and a better understanding is required before selection pressures on DENV in the vector can be characterized . In addition to the classical immune signaling pathways , RNAi is also a major antiviral defense mechanism in the mosquito [40 , 46] . WNV intra-host genetic diversity is thought to be driven by RNAi , with the parts of the WNV genome most likely to be targeted by RNAi also being the most diverse [8] . A separate study made use of artificially diverse WNV strains to show that high intra-host genetic diversity was associated with increased viral fitness in Culex mosquitoes [47] . Although it has been proposed that purifying selection is relaxed in the Culex mosquito host for WNV [10 , 48] , these studies suggest that diversity may be actively selected for in insect vectors . This idea is supported by our observation that almost entirely new DENV SNV repertoires were generated at each transmission stage , but requires further testing in the DENV-Ae . aegypti system . We found no evidence for differences in selection pressures between the mosquito abdomen and salivary gland . Still , because the SNV sample size was small , this does not rule out differences in immune pressures between body compartments , and #NS/#S ratios did diverge for several genes . A potential driver of differing immune pressures is the midgut microbiota , which has been shown to impact mosquito physiology and vector competence for human pathogens ( reviewed in [49] ) . Depletion of the gut microbiota increases mosquito susceptibility to DENV [50]; this is thought to occur at least partly because the microbiota trigger a basal level of immune activity [50–52] . It will be interesting to examine the impact of the gut microbiota on DENV diversity . After phasing SNVs into distinct viral genomes , we observed that in most cases , only one SNV from a predicted variant genome was maintained across transmission stages , with a maximum of two being maintained as a unit . This suggests that SNVs in a sample are spread out across many variant genomes instead of being clustered together on one genome , and is consistent with the RdRp error rate of 1x10-4 [4] , which corresponds to approximately one mutation per 11 kb DENV genome . Despite the growing number of studies characterizing intra-host genetic diversity , the impact of this diversity on virulence or disease severity for arboviruses in vertebrate hosts is not well understood . One study unexpectedly found that mouse morbidity and mortality was negatively correlated with WNV intra-host genetic diversity [10] . The authors suggest that this could be due to the parental virus clone being highly pathogenic , such that most mutations would result in a decrease in pathogenicity . Alternatively , adaptation to mosquitoes during passage could also have resulted in a loss of pathogenicity [10] . Another study found lower intra-host diversity in patients with severe dengue than with mild dengue [12]; however the authors state that it is difficult to know if these differences are the cause or the consequence of disease severity . In other studies , no association was found between DENV intra-host diversity and disease severity [7 , 14] . It will be important to address the question of whether disease severity is associated with the genetic diversity of the infecting viral population , rather than with diversity after the onset of symptoms . | Dengue virus ( DENV ) is transmitted between humans through the bite of infected female Aedes aegypti mosquitoes . Virus populations experience significant drops in size and are subject to differing selection pressures as they cycle between human and mosquito hosts . Subsequent changes in viral intra-host genetic diversity may have consequences for the adaptability and fitness of the virus population as a whole but are poorly understood . To study the impact of human-to-mosquito transmission on DENV populations , we allowed mosquitoes to feed directly on patients with acute dengue infections , then deep-sequenced DENV populations from patient plasma samples and from the abdomens and salivary glands of corresponding mosquitoes . These matched samples allowed us to estimate the size of the population drop that occurs during establishment of infection in the mosquito , track changes in viral intra-host variant repertoires at different stages in transmission , and investigate the possibility of host-specific immune selection pressures acting on the virus population . These novel insights improve our understanding of DENV population dynamics during horizontal transmission . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| []
| 2015 | Tracking Dengue Virus Intra-host Genetic Diversity during Human-to-Mosquito Transmission |
Circulation is an important delivery method for both natural and synthetic molecules , but microenvironment interactions , regulated by endothelial cells and critical to the molecule's fate , are difficult to interpret using traditional approaches . In this work , we analyzed and predicted growth factor capture under flow using computer modeling and a three-dimensional experimental approach that includes pertinent circulation characteristics such as pulsatile flow , competing binding interactions , and limited bioavailability . An understanding of the controlling features of this process was desired . The experimental module consisted of a bioreactor with synthetic endothelial-lined hollow fibers under flow . The physical design of the system was incorporated into the model parameters . The heparin-binding growth factor fibroblast growth factor-2 ( FGF-2 ) was used for both the experiments and simulations . Our computational model was composed of three parts: ( 1 ) media flow equations , ( 2 ) mass transport equations and ( 3 ) cell surface reaction equations . The model is based on the flow and reactions within a single hollow fiber and was scaled linearly by the total number of fibers for comparison with experimental results . Our model predicted , and experiments confirmed , that removal of heparan sulfate ( HS ) from the system would result in a dramatic loss of binding by heparin-binding proteins , but not by proteins that do not bind heparin . The model further predicted a significant loss of bound protein at flow rates only slightly higher than average capillary flow rates , corroborated experimentally , suggesting that the probability of capture in a single pass at high flow rates is extremely low . Several other key parameters were investigated with the coupling between receptors and proteoglycans shown to have a critical impact on successful capture . The combined system offers opportunities to examine circulation capture in a straightforward quantitative manner that should prove advantageous for biologicals or drug delivery investigations .
The bioavailability of molecules as they circulate through the bloodstream is a crucial factor in their signaling capability . Half-life in circulation can determine the effectiveness of a drug simply by regulating the opportunities a molecule has to interact with the vessel wall . Although in vivo measurements are routinely made by researchers to monitor serum levels of molecules and to determine half-lives , interactions in the microenvironment are not easily measured or observed . While some molecules may have a long circulation life , many may have only a single opportunity to interact with the blood vessel walls before being filtered through the liver or kidneys . In addition , even molecules with a long circulation life may still face impediments to direct interaction with the endothelium . This , for example , is the case with vascular endothelial growth factor ( VEGF ) when bound to bevacizumab , a monoclonal antibody to VEGF [1] , [2] . Bevacizumab has been shown to increase the circulating concentration of VEGF in cancer patients when compared to patients not undergoing therapy because of the increased half-life of the growth factor-antibody complex; however the complex is unable to bind to VEGF receptors [3] making delivery of the VEGF questionable . In order to better understand the vessel microenvironment and to accurately monitor drug interactions in the context of that microenvironment , better tools are needed to provide meaningful measurements that can predict the fate of molecules in circulation . Many important measurements have and continue to be made using in vitro mammalian tissue culture methods but there are obvious limitations to the traditional two-dimensional culture approach . In circulation , the influence of flow on whether a molecule remains in the fluid phase or binds to the vessel wall can be a dominant factor . This influence cannot be ascertained in static tissue culture studies . For example , the velocity of blood in the aorta is ∼400 mm/sec while at the capillary level it is less than 1 mm/sec [4] . This reduction in velocity allows the exchange processes at the capillary level to take place more efficiently [4] and it likely also affects the activity of molecules in circulation that rely on cell surface binding in order to fulfill their roles . While direct measurement of this binding process is difficult , our model makes use of a commercial bioreactor with endothelial-lined hollow tubes operating under pulsatile flow to mimic the vascular environment architecture and to directly measure the loss of molecules as they pass through these hollow fibers . We have used a single pass method to allow better assessment of the effect of flow in either retaining molecules in the circulation or permitting their interaction with vessels . Our approach also makes use of a bolus administration , since this is a typical way in which drugs would be delivered in a clinical setting . The binding of fibroblast growth factor-2 ( FGF-2 ) to its cell surface receptor ( FGFR ) and the role of heparan sulfate proteoglycans ( HSPG ) in regulating the process have been of research interest for many years because of their role in angiogenesis , the growth of new blood vessels from existing vessels . Knowledge of how these processes work could aid in the development of new therapeutics to control tumor growth and assist clinically in the treatment of chronic wounds . In order to understand the mechanism of FGF-2-mediated cell proliferation , a multitude of experimental studies have been undertaken [5] and , in the past two decades , several computational models of FGF-2 binding to its receptor FGFR and HSPG have been proposed [6]–[11] . Insight can be gained through experiment-coupled modeling that could not otherwise be readily obtained . Nugent and Edelman [11] were among the earliest researchers to develop a simple model that includes three species , FGF-2 , FGFR and HSPG . They measured kinetic binding rate constants experimentally and used their model to analyze the data thereby providing a foundation for investigating the complexity of FGF-2 binding . A similar approach was used by Ibrahimi et al [9] to investigate stepwise assembly of a ternary FGF-2-FGFR-HSPG complex in conjunction with their surface plasmon resonance measurements . We introduced more complexity into the FGF-2 binding model with the inclusion of heparin binding [12] , receptor dimerization [8] , and formation of alternative HSPG-FGFR species [13] . Recent models have moved towards including intracellular signaling [14] . With the exception of work by Filion and Popel [7] , [15] , which included diffusive transport , previous simulation work has been based on a static tissue culture environment that may be quite different from the dynamic in vivo environment of blood vessels . We introduced a computational model based on a flow environment in which the competitive binding of FGF-2 , FGFR , and HSPG in a pulsatile flow environment was addressed to mimic blood vessel-like hollow fibers [16] , [17] . In this paper we use an updated version of that model to explore how specific parameters such as flow rate impact FGF-2 capture and receptor binding , and compare our results with experimental studies . Insights with regard to the importance of surface coupling and ligand depletion zones within the fluid phase were found . The described simulation package provides a new and valuable way to investigate growth factor capture and can be easily extended to other biologically relevant molecules and drugs .
BAECs ( passage 10 ) , cryopreserved in liquid nitrogen , were cultured in Dulbecco's modified Eagle's medium ( DMEM-low glucose , phenol red-free , Invitrogen Corporation , Grand Island , NY ) , supplemented with penicillin ( 100U/mL , Invitrogen Corporation , Grand Island , NY ) , streptomycin ( 100µg/mL , Invitrogen Corporation , Grand Island , NY ) , glutamine ( 2mM , Invitrogen Corporation , Grand Island , NY ) , and 5% newborn calf serum ( Invitrogen Corporation , Grand Island , NY ) . When a sufficient number of cells were grown ( passage 11∼13 ) , they were transferred to the hollow fiber cartridge . The FiberCell polysulfone plus endothelial cartridges ( C2025 , FiberCell Systems Inc . , Frederick , MD ) , also called hollow fiber bioreactors , contain 20 capillaries which are 12 cm long , 700 µm I . D . , 300 µm wall , 0 . 1µm pore size , 53 cm2 lumen surface area ( Figure 1A ) . They were activated with 70% ethanol ( Fisher Scientific , Houston , TX ) , followed by multiple washes with sterile distilled water . The cartridges were then coated using 5 µg/mL fibronectin ( Sigma Aldrich , St . Louis , MO ) in phosphate buffered saline ( PBS , Invitrogen Corporation , Grand Island , NY ) . BAECs ( passage 11∼13 ) were inoculated into the cartridges ( 0 . 7–1×107 cells/cartridge ) 24 hours after the coating and placed in an incubator for 4 hours ( rotated 180° after 2 hours ) without flow in order to promote cell attachment . The BAEC culture cartridges were then linked to the FiberCell pump system ( FiberCell Systems Inc . , Frederick , MD ) and media circulated through the system at ∼2 . 6 mL/minute ( 5 . 2 mm/sec ) . The flow system was maintained in the incubator ( 37°C , 5% CO2 ) at all times except during the experiment periods . Cell growth and viability was monitored by measurement of the cell glucose consumption from the medium once a day with OneTouch UltraSmart blood glucose monitoring system ( Lifescan , Inc . , Milpitas , CA ) . The flow system and cell-lined cartridges were removed from the incubator , gently washed once with warmed ( 37°C ) PBS ( 60 mL ) , and then maintained in circulating 125 mL serum-free medium ( DMEM-low glucose , phenol red-free , supplemented with 0 . 05% gelatin in PBS ) in a sterile room-temperature tissue culture hood ( Thermo Scientific , Waltham , MA ) . After establishing flow at the desired rate ( low rate: 0 . 60∼0 . 68 mL/min ( 1 . 2–1 . 36 mm/sec ) ; high rate: 1 . 6–1 . 8 mL/min ( 3 . 2–3 . 6 mm/sec ) or 2 . 9–3 . 0 mL/min ( 5 . 8–6 . 0 mm/sec ) ) with a CellMax Quad pump ( Spectrum Laboratories , Inc . ) for about 2 minutes , flow was stopped to allow the growth factor of interest ( FGF-2 ( Sigma Aldrich , St . Louis , MO ) , EGF ( R&D Systems Inc . , Minneapolis , MN ) and VEGF ( R&D Systems Inc . , Minneapolis , MN ) ) ( 0 . 11 mL ) to be injected into the inlet . After the injection , the flow was resumed and the flow media collected ( two drops/fraction ) for the desired time period . The flow pattern was assumed to be sigmoidal based on previous studies [18] , [19] . The cartridges were then gently washed with warmed PBS supplemented with 0 . 3M NaCl ( 10 mL ) followed by one wash with 10 mL PBS and a wash of the whole flow system with PBS ( 60 mL ) . The system was returned to the same culture media and flow rates as described under Preparation of BAECs , allowing at least 24 hours before the next experiment . The media fractions collected during the binding experiments were stored at 4°C and analyzed with ELISA kits ( R&D Systems Inc . , Minneapolis , MN ) within the next 24∼48 hours . Dynamic viscosity of the test cell culture medium was measured using a DV-II++ Pro Programmable cone-plate viscometer ( cone #CPE-40; Brookfield Engineering Laboratories; Boston , MA ) according to the manufacturer's instructions . Viscosity measurements were made for a range ( 375 to 750 sec−1 ) of shear rates ( to confirm Newtonian fluid behavior ) at room ( i . e . , 25°C ) and physiologic ( i . e . , 37°C ) temperatures . Heparan sulfate expression was measured in static tissue culture dishes and in the flow cartridge by heparinase treatment of cells , collection of the cleaved glycosaminoglycans , and quantitation using a dimethylene blue colorimetric assay [20] , [21] . Cells in static culture contained 4 . 3+/−0 . 31×10−6 µg of heparan sulfate/cell and cells in cartridge hollow fibers contained 1 . 1+/−0 . 09×10−6 µg of heparan sulfate/cell , reflecting an ∼75% reduction in cell surface heparan sulfate under flow ( 0 . 63 mL/min ( 1 . 26 mm/sec ) ) . Heparinase III ( 0 . 01 unit/0 . 11mL , Seikagaku Corp . , Japan; 0 . 2unit/0 . 11mL , Sigma Aldrich , St . Louis , MO ) , chondroitinase ABC ( 0 . 2 unit/0 . 11mL , Seikagaku Corp . , Japan ) and keratanase ( 0 . 33unit/0 . 11mL , Sigma Aldrich , St . Louis , MO ) were utilized to observe their effect on growth factor flow and binding . In some experiments , the enzymes ( heparinase III , chondroitinase ABC and keratanase ) were mixed together as an enzymatic cocktail solution at the above concentrations . Cartridges were treated for 20 minutes at 37°C , washed with warmed PBS ( 10 mL ) , and growth factor studies performed as described above . Non-specific binding of FGF-2 in the system was determined to be primarily due to the inlet reservoir . The reservoir chamber was removed from the cartridge , growth factors were injected into the inlet of the cartridges with a syringe , and flow was initiated . Fractions were collected as they exited the reservoir . Growth factors were measured before injection and compared to the sum of the collected fractions . The difference between the input amount and the amount collected constituted the nonspecific binding in our experiments . For FGF-2 ( 1 . 0+/−0 . 1 ng ) , the amount retained in the reservoir was 29+/−2 . 8% of the FGF-2 added ( SD , n = 3 ) . Additional nonspecific binding within the hollow fibers was assumed to be minimal . The concentrations of FGF-2 , EGF , and VEGF in the collected fractions were measured by ELISA . The flow rate of each experimental run was determined from the total volume collected divided by the total flow time . To visualize the BAECs cultured in the flow system , cartridges were washed with PBS supplemented with 0 . 5M NaCl to extrude the endothelial cell lining from the hollow fibers and then the cell linings were fixed with 4% paraformaldehyde ( Electron Microscopy Sciences , Hatfield , PA ) in PBS for 10 minutes . Three washes with PBS ( one minute per wash ) followed and the cell linings permeabilized with PBS supplemented with 0 . 03% Triton and 1% BSA for 3 minutes on a shaker platform at room temperature . The cells were then treated with 10 µg/mL 4′ , 6-diamidino-2-phenylindole ( DAPI ) ( Sigma Aldrich , St . Louis , MO ) in PBS supplemented with 0 . 03% Triton and 1% BSA for 20 minutes , followed by three PBS washes for 2 minutes each at room temperature . The cells were then visualized and photographed using a Nikon Eclipse TE 2000E fluorescent microscope ( Nikon , Melville , NY ) at an excitation wavelength of 350 nm ( Figure 1B ) . The computational model is based on the physical dimensions of the bioreactor although the system is scalable to other desired dimensions . The domain of the simulation is the hollow-fiber portion of the cartridge ( Figure 1 ) . The computational model has three coupled parts: ( 1 ) the medium flow equations; ( 2 ) the convective mass transport equations of growth factor in the flow; ( 3 ) the binding kinetics equations on the wall of the fibers [8] , [16] . In order to solve the coupled equations numerically and efficiently , the following assumptions are made: ( 1 ) the walls of the hollow fibers are rigid and nonporous; ( 2 ) the flow is axisymmetric and laminar; ( 3 ) the fluid is incompressible , Newtonian and isothermal; ( 4 ) all of the hollow-fiber capillaries within the cartridge have the same dimensions , flow rate , cell densities and entrance conditions; and ( 5 ) the cells are packed tightly and distributed evenly on the wall of the hollow-fiber capillaries . Entrance effects of the flow are ignored [22] , [23] and , consequently , the flow within the fibers is treated as fully developed flow in which the radial velocity is neglected . A uniform mesh is used . The kinetic pathways are shown in Figure 2 and the equations and parameter values are included in Tables 1 and 2 , respectively . In our experimental system , FGF-2 is injected into the inlet reservoir where it is assumed to quickly reach a uniform concentration . The concentration of FGF-2 in the reservoir is assumed to decrease gradually as fluid is pumped into the reservoir prior to distribution into the capillaries with each pulse cycle as:where is the volume of the reservoir , is the volume of fluid flowing into the fibers at each pulse , is the current and is the previous concentration of FGF-2 in the reservoir . , where is the amount of FGF-2 injected . The pump pulse cycle was measured experimentally and determined to be ∼36 strokes/min at a flow rate of 1 . 4 mm/sec . Pulsatile flow is treated in the following manner . A pulse of fluid volume enters the pre-pump inlet reservoir ( 0 . 4 mL volume ) , from which a continuous flow of fluid having an axial velocity greater than or equal to zero enters the cell-lined fibers in the cartridge . The axial velocity is oscillatory but with only positive terms . Entrance effects are considered negligible [23] . The velocity of the fluid in the axial direction is determined with the following formula [17]:where qs is the average volumetric flow rate , Nf is the number of fibers inside the cartridge , R is the radius of a fiber , ω = 2π/T is the angular frequency of the pulsatile flow , and T is the pump pulse cycle . Good agreement between the simulation and experimental results was determined based on two criteria: an amount criterion and a curve-matching criterion . The amount criterion is defined as:where Mexp is the outflow amount of protein determined experimentally , Msim is the outflow amount determined within the simulations and M is the amount of FGF-2 entering the capillary . The curve-matching criterion is calculated in the following way . The FGF-2 exit profile curve is not a continuous curve but is a series of discrete values at different time intervals . This makes use of traditional curve matching algorithms difficult . Our method aligns the initial exit times for the simulations and experiments and then calculates the distance between points on the two outflow curves using the following formula:where N is the total number of time intervals . and is the amount of FGF-2 exited at the ith time interval in experiment and simulation , respectively . The curve-matching criterion is defined as:A special program written in C/C++ that operates under Windows XP or Vista operating system has been built for solving this model and has been described previously [16] , [17] . The interface allows users to easily set parameters related to the simulation such as FGF-2 injected concentration , flow rate , mesh size , time step , and total simulation time via either configuration text files or from the computer interface . The mass transport of FGF-2 within the fiber is visualized in real time during the simulation process . A Linux version of the software is also available however it lacks a user interface tool and there is no real time visualization . The binary code can be downloaded from www . cs . uky . edu/~czhanb/research . html . In the simulations there are 800 , 000 cells/fiber or 16 , 000 , 000 cells/cartridge , a value which was obtained from the experimental system . The tolerance for solving the mass transport PDEs was set at 10−12 . The relative tolerance for solving the kinetic ODEs was set at 10−8 and the absolute tolerance was 10−12 . All experiments were performed a minimum of three times in independent cartridges . The mean of all replicates ± standard deviation of those replicates is presented except where discrete measurements were used to more closely represent small changes in initial concentration . Significance ( p<0 . 05 ) was determined using a Student t-test with a two-tail distribution and unequal variance ( Excel , Microsoft ) .
Endothelial cells line blood vessels and are the initial entry point for access of blood-borne proteins to the underlying tissue . Our investigations focused on flow and the impact it has on endothelial cell capture of growth factors , which are important regulators of cell and tissue activity . To better approximate the microenvironment of a blood vessel , we seeded bovine aortic endothelial cells into the FiberCell cartridge system and cultured the cells under flow ( Figure 1A ) . Cell viability was confirmed for up to 8 weeks and cell density was ∼0 . 3×106/cm2 . The geometry is clearly more similar to in vivo than typical cell culture dishes but it was important to obtain a uniform and confluent monolayer of cells within the cartridge system to correctly perform and analyze experiments . To confirm this , cartridges were treated with a high salt wash to extrude the cell-based vessel and the cells were fixed and imaged ( Figure 1B ) . An incision was made at one end to expose the lumen and demonstrate the continuity of the cell layer . The average fluid velocity in human capillaries is <1 mm/sec [4] . We hypothesized that capture of regulatory growth factors from solution would be significant at these flow rates thereby facilitating growth factor activity . Using the lowest velocity setting with the standard pulsatile pump included with the Cellmax system ( ∼1 . 3 mm/sec , ∼0 . 65 mL/min ) , FGF-2 ( 5 . 0±0 . 4 ng ) was injected into the cartridge inlet reservoir and flow was commenced . As shown in Figure 3 , there is a delay in FGF-2 appearance in the outflow corresponding to the time for FGF-2 to travel through the cartridge and exit the system . The majority of FGF-2 added exited the cartridge as a large peak approximately 1 mL ( or 1 . 5 min at this flow rate ) after flow was initiated . Non-specific binding within the injection cartridge reservoir was measured directly ( 31+/−2 . 5% ) . Specific binding within the cell-lined hollow fibers accounted for 9+/−2 . 5% of total FGF-2 added to the cartridge at this concentration and ∼13% of the FGF-2 entering the cell-lined fibers , after taking into account non-specific binding ( Figure 3 ) . The results shown in Figure 3A are from three independent experiments conducted using three different cartridges illustrating the reproducibility of the system . Repeat runs conducted using the same cartridge as well as runs using radiolabeled FGF-2 instead of unlabeled FGF-2 both produced similar results ( data not shown ) . The peak appearance time or volume in the outflow from the cartridge was insensitive to FGF-2 injection concentration in the range studied ( data not shown ) . However , the size of the FGF-2 peak correlated with the injection concentration with the highest peak corresponding to the highest concentration of FGF-2 added ( Figure 3B ) . The accuracy of our measurements took into consideration specific losses that occurred with injection ( i . e . tube , syringe , needle , and reservoir ) . Rather than averaging datasets with variable FGF-2 reservoir values , we therefore present them as discrete results . A plot of total FGF-2 retained at these discrete concentration points shows a dose responsive binding curve , reflecting the linear portion of the binding curve expected at sub-saturation ligand concentrations ( Figure 3C ) . Heparan sulfate proteoglycans ( HSPG ) are ubiquitous molecules found on virtually all cells including endothelial cells and have been shown to regulate heparin-binding growth factor binding and activity in tissue culture [6] , [24]–[28] . FGF-2 is a heparin-binding molecule associated with a number of physiologic and pathologic processes [29] and , therefore , the role of HSPG in regulating FGF-2 retention under flow was examined . Although the binding affinity of FGF-2 for HSPG has been shown to be lower than the affinity for the FGF receptor , these HSPG sites can provide up to a thousand fold more binding sites for FGF-2 [6] , [24] significantly impacting the cell binding “potential” for heparin-binding growth factors . Cartridges were treated with heparinase , an enzyme specific for heparin and heparan sulfate , and FGF-2 outflow quantified . After heparinase treatment , FGF ( ∼1 ng ) was injected and pumped through the cartridge . Almost 74% of the total FGF-2 added to the system was recovered in the outflow , compared to ∼46% of the total FGF-2 recovered from the non-heparinase treated cartridge prior to subtraction of non-specific binding . The amount of FGF-2 retained in the cartridge after heparinase treatment corresponded to the measured level of non-specific binding and thus indicated no specific binding to cell-lined fibers in the absence of HSPGs ( Table 3 ) . In contrast , 25% of the FGF-2 pumped through untreated cartridges was retained after subtraction of non-specific binding . Although FGF-2 can bind to its receptor in the absence of HSPG stabilization , that binding , based on the apparent KD of the receptor for FGF-2 in the absence of heparan sulfate , the lower level of FGFR generally found , and the ligand-receptor exposure time under flow , would be expected to be at least ten-fold lower than in the presence of HSPG [24] and our data certainly support this . To ensure that the effect with heparinase under flow was due to the specific removal of heparan sulfate and not a general effect due to enzymatic treatment of the cartridge or the enzyme incubation process , the cartridges were treated with keratanase , an enzyme having no specific known target on these cells . Keratanase , as opposed to heparinase , had no significant effect on FGF-2 retention ( Table 3 ) . Interestingly , there was a small but reproducible reduction ( ∼9% ) after chondroitinase treatment on FGF-2 retention compared to control . Chondroitin sulfate proteoglycans are typically found on vascular surfaces but FGF-2 has not been shown to bind directly to chondroitin sulfate [30] , [31] . It is not known at this time what the cause for the reduced binding is , although it has been reported that both chondroitin sulfate and dermatan sulfate under certain circumstances are able to influence FGF binding [32]–[34] . VEGF , a heparin binding protein , and EGF , which does not bind heparin , were next tested in this system . Both the initial appearance time and outflow volume for the protein as well as the general shape of the outflow peak for both VEGF and EGF were similar to FGF-2 ( Figure 4 ) . To ensure that the measured effects seen with heparinase-treatment on FGF-2 retention were due to specific responses of the growth factor to the removal of heparan sulfate and not a general response by all proteins , flow studies were done with VEGF and EGF following enzymatic treatment . EGF retention and outflow were unaffected by treatment with a cocktail of heparinase , chondroitinase , and keratanase ( Table 4 ) . Treatment with heparinase without chondroitinase or keratanase also had no effect on EGF retention or outflow ( data not shown ) . In contrast , VEGF showed a significant decrease in specific retention between control and heparinase treated cartridges ( 16+/−5 . 8% versus −2 . 5+/−6 . 1% VEGF retained ) indicating the critical role HSPG can have in heparin-binding growth factor capture under flow . The lack of a change in EGF binding or outflow profile under heparinase treatment is supportive that there are no gross changes in the cell glycocalyx that might impact the shear stress in the system . Capture of FGF-2 by endothelial cells within the vasculature is a critical step in growth factor activity and our bioreactor is an excellent tool for investigating the capture process . However , it has limitations with regard to quantification of cellular binding behavior . The cartridges are expensive for short-term experiments and culture time and preparation can be relatively lengthy . Visualization of individual cell behavior within the culture is not feasible . In addition , the ability to predict the capture of molecules by cells under flow has value across a wide range of areas and the development of a flow-based tool for the design and testing of mechanisms related to retention is desireable . Our computer model was designed based on media flow equations and mass transport equations [35] with cell surface reaction equations to reflect the cell-growth factor interactions ( see Materials and Methods-Model development ) . To validate the model , simulations were performed using the variables ( ie FGF input concentration and flow rate ) specific for an experimental series and a comparison was made . Experimental trials were run in which FGF-2 ( 0 . 92 ng ) was added to the reservoir , pumped through the cartridge , and outflow collected and analyzed for FGF-2 . FGF-2 in the outflow showed a characteristic peak outflow approximately 100s after flow was initiated at 0 . 63 mL/min ( 1 . 26 mm/sec ) and 17±6 . 3% of the input FGF-2 was retained within the cartridge after non-specific binding was subtracted ( Figure 5 ) . Simulations performed using the same input FGF-2 value and flow rate were run and comparison was made between the simulations and experimental outflow from control ( Figure 5A ) or heparinase-treated ( Figure 5B ) cartridges . We defined good agreement based on two criteria; the amount of FGF-2 recovered and the curve similarity . Criteria one requires the relative difference in FGF-2 outlow from the experimental and simulation studies to be less than 1% while the second criteria compares the actual amounts of FGF-2 exiting from the experimental and the simulation system ( see Materials and Methods ) . We did note that FGF-2 retention with the simulations was very dependent on the level of HSPGs with higher densities resulting in too much retention via HSPG-FGF-2 binding and subsequent FGFR coupling while lower HSPG densities resulted in too little retention ( data not shown ) . Comparison of simulation results with our heparinase-treated data showed fine agreement with regard to our criteria when non-specific loss in the reservoir was subtracted . Capillary flow is generally steady , and gradually becomes pulsatile at higher flow rates . We conducted simulations and in vitro experiments to compare steady and pulsatile flow at a low flow rate ( 0 . 6 mL/min , 1 . 2 mm/sec ) to determine whether our model would predict differences between FGF-2 interactions using steady and pulsatile flow . Simulations predicted no difference in FGF-2 binding at low flow using pulsatile flow conditions versus steady flow in either the FGF binding down the cell-lined hollow fiber ( Figure 6A ) or in the profile of the outflow ( Figure 6B ) . In vitro experiments were performed using a syringe pump for steady flow and the bioreactor's pulsatile flow pump ( Figure 6C ) . FGF-2 outflow measurements indicated no overall change at 0 . 6 mL/min ( 1 . 2 mm/sec ) suggesting that , at low rates typical of capillary flow , no significant change in FGF-2 interactions takes place . Our experimental system does not allow easy separation between internalized FGF-2 and that bound to the cell surface or visualization of FGF-2 distribution within the cell-lined hollow fiber . Using our computer model we examined how FGF-2 would be distributed with respect to time after flow was initiated ( Figure 7 ) . At a relatively low flow rate ( 0 . 63 mL/min , 1 . 26 mm/sec ) , the FGF-2 in the reservoir had essentially all entered the hollow fibers by 150s and the peak outflow of FGF-2 was evident ∼200s after flow was initiated corresponding to the time when the bulk FGF-2 had exited the hollow fibers . Later times showed cell-bound FGF-2 either internalized or dissociated from the cell surface with little chance to reassociate . The vast majority of binding is predicted to occur near the entrance to the cell-lined hollow fibers as opposed to the middle or end of the fibers ( Figure 7B ) . The impact of time was more pronounced in the front section also as fluid entering the hollow fiber after ∼150s was devoid of FGF-2 ( <0 . 1% of initial FGF-2 ) . Increasing the diffusion rate for FGF-2 in solution by increasing the diffusion coefficient by an order of magnitude is predicted to have a negligible impact on FGF-2 capture in the front of the capillary but increased significantly the FGF-2 bound down the length of the cell-lined hollow fiber . This was due to changes in the depletion zone near the cell-lined walls ( Figure 8 ) . After 44s , an FGF-2 depletion zone near the surface was evident which was reduced when the diffusive transport of FGF-2 was increased . The replenishment of FGF-2 near the wall promoted greater FGF-2 binding as complex formation is a second-order process and illustrates the importance of surface depletion in growth factor capture . Our simulations indicate that depletion near the cell surface impacts binding and suggests that residence time in the vicinity of the cell surface is important . We therefore looked at how flow impacted cell binding of FGF-2 . Simulations predict that cell binding is significantly diminished with increased flow rate ( Figure 9A ) although the basic result of high binding at the entrance and reduced binding down the cell-lined hollow fiber was consistent across flow rates examined ( data not shown ) . This difference was evident regardless of the concentration of FGF-2 introduced to the system with the difference being more pronounced at higher flow rates ( Figure 9B ) . Reduction in binding due to the loss of HSPG is less evident at higher flow rates where the specific binding was already greatly reduced . This inverse relationship between flow and cell binding is potentially important especially at these relatively low flow rates . The highest rate used in our simulations ( ∼3 mL/min , ∼6 mm/sec ) is considerably lower than average arterial flow rates ( 100–400 mm/sec ) in larger vessels of the circulatory system [4] suggesting that , with a short half-life , retention may be relevent only in small vessels with lower velocities . Note that simulations were run to a constant time rather than volume to reduce small fluctutations in retained FGF-2 due to dissociation effects . Experimentally , we found results that were consistent but not quantitatively exact with this model prediction ( Table 5 ) . FGF-2 retention in the hollow fibers was virtually eliminated under medium ( ∼1 . 7 mL/min , 3 . 4 mm/sec ) and higher flow rates ( 3 . 0 mL/min , 6 mm/sec ) , a significant reduction compared to binding at 0 . 62 mL/min ( 1 . 24 mm/sec ) ( Table 3- control group ) . The simulations , in contrast , did show some level of binding even at the highest level but this likely reflects the idealized conditions used for the model system ( i . e . uniform receptor and HPSG densities , free access to coupling between FGF-2 bound molecules ) . Heparinase treatment showed no significant further reduction in retention at the higher flow rates in agreement with the simulation results . Simulations indicated no difference in FGF-2 binding under our pulsatile flow conditions versus steady flow ( data not shown ) . Additional experiments were performed using a syringe pump with steady flow rather than pulsatile flow . FGF-2 outflow measurements indicated no overall change at 0 . 62 mL/min ( 1 . 2 mm/sec ) ( data not shown ) . Qualitatively the experimental results agreed with the simulation predictions for the overall effect of flow rate on retention although the model suggested higher retention levels for the control case and closer agreement between control and heparinase at both higher flow rates . FGF-2 binding affinity and concentration , along with binding partner density , regulates the capture process for FGF-2 from the fluid phase . We therefore examined using our simulations how varying the affinity of FGF-2 for either HSPG ( Figure 10A ) or FGFR ( Figure 10B ) while holding all other parameters at their baseline value would impact retention . Decreasing the affinity ( i . e . increasing KD ) for HSPG had a dramatic effect on retention reducing it to 40% of baseline capture at the lowest value examined . The association rate constant had a greater impact than the dissociation rate constant although both followed similar trends . Somewhat surprisingly , increasing the affinity of the interaction by reducing the value of the dissociation rate constant of FGF-2 for HSPG did not alter FGF-2 binding likely due to the strong coupling present between FGFR and HSPG in the presence of FGF-2 , making strict HSPG-dissociation somewhat irrelevant . For the same reason , FGF affinity for FGFR did not have a strong impact on FGF-2 capture since the vast majority of FGF-2 interacting with FGFR was via FGF-2-HSPG coupling . Cells typically express significantly more HSPG than FGFR and we next asked how varying the cell surface densities of these binding sites would impact FGF-2 capture . In the absence of FGFR , a typical density of HSPG in our cartridge ( 2 . 5×105 #/cell ) resulted in significant binding of FGF-2 in the absence of FGFR that is essentially doubled when FGFR density is 1×106 #/cell , a two-fold increase in binding sites ( Figure 11A ) . FGFR typically are expressed at densities of approximately 1×104 #/cell thereby keeping the primary signaling receptor at a controlled level . This is predicted to result in an order of magnitude less overall FGF-2 binding than that found at typical HSPG levels but which is increased in a similar way when HSPG are present . The combination of the two surface binding sites ( FGFR and HSPG ) is critical . For example , when 1 . 0×104 FGFR are present , the retained FGF-2 is increased to ∼0 . 25ng from a value of ∼0 . 14ng without the FGFR . Looking at cell binding at the entrance of the cell-lined hollow fiber as a function of time after FGF-2 has been introduced with constant FGFR ( 1×104 #/cell ) and variable HSPG , we found that there was a significant increase in bound FGF-2 at the higher HSPG ( 1×105 #/cell ) when compared to the lower values and that the FGFR binding was essentially all coupled to HSPG ( Figure 11B ) . When there are fewer HSPG , there is a lower percentage of coupled binding at least at earlier times as well as lower overall FGFR complexes . The results with the FGF-2-HSPG affinity simulations and the density studies indicated the importance of coupling in facilitating effective FGF-2-FGFR interactions . We next looked at how varying the coupling rate constant impacted binding and internalization using simulations ( Figure 12 ) . In the absence of HSPG-FGFR coupling ( kc = 0 ) , there is a reduction in peak binding of FGF-2 and the majority of FGF-2 bound is not internalized but dissociates and exits from the system in the outflow . Even with a low level of coupling , the FGF-2 binding and internalization is dramatically increased until a peak effect is seen with kc = 0 . 01 ( #/cell ) −1 min−1 . If we looked at later times in the simulation ( Figure 12B ) , we would find that a large fraction of the FGF-2 injected is bound during the initial pass and that this bound FGF-2 is largely internalized with little exiting the system . If coupling between HSPG and FGFR is eliminated ( Figure 12C ) , this is not the case . In this scenario , the cells bind a smaller but still significant level of FGF-2 during the initial pass but this FGF-2 is not retained and nearly all of the FGF-2 captured ultimately exits the system in the outflow . To further illustrate the importance of the coupling process , simulations were performed with cell-lined hollow fibers having only HSPG ( 2 . 5×105 #/cell ) in the front 25% of the tube and both FGFR ( 1×104 #/cell ) and HSPG ( 2 . 5×105 #/cell ) in the back 75% of the fiber ( Figure 13 ) . The entrance area ( front 25% ) did not include internalization of FGF-2 by HSPG modeling an ECM-like section , however , the overall outcomes are not significantly changed when internalization is included ( data not shown ) . HSPGs in this front section were able to capture FGF-2 but there is a significant rise in retention in the back section where both HSPG and FGFR are present . This is not simply due to the increase in binding sites due to the addition of FGFR as increasing HSPG by an equivalent level to that of the HSPG plus FGFR did not lead to the same increase in retention ( data not shown ) . Moreover , this increase in retention is lost when the dissociation rate for FGF-2-FGFR-HSPG is reduced to that of FGF-2-HSPG and only nominally increased when the coupling rate is eliminated , reflecting the increased affinity of FGFR compared to HSPG for FGF-2 ( data not shown ) . The effect is evident at both low and high flow rates . Finally , we used simulations to ask whether dissociation from HSPG in an ECM-like section could lead to increased binding downstream due to slow dissociation of the growth factor and prolonged availability of the growth factor for downstream binding . When the HSPG density in the front 25% zone was increased to 5×106 HSPG/cell , a large increase in overall retention of FGF-2 in the front section was evident resulting in a decrease in FGF binding in the HSPG-FGFR section ( back 75% ) due to a depletion of FGF-2 in the fluid zone near the cells . This was evident at both 5 ( Table 6 ) and 10 min ( data not shown ) . In contrast , a low level of HSPG ( 5×104 or less ) in the entrance section did not lead to significant binding in this zone and results in increased binding of FGF-2 in the final 75% section . FGF-2 in the fluid phase was at a higher concentration at later times after FGF-2 injection when there were more HSPG in the front section due to dissociation from the HSPGs; however , under flow conditions , this dissociated FGF-2 is not predicted to grow to a high enough concentration to meaningfully impact downstream receptor binding . This is an important difference between flow and static culture studies .
Circulation is an obligatory process for the maintenance of human life . The proper balance of solid and fluid components , flow and pressure , and chemical content are all tightly regulated to maintain homeostasis . Within these limits , however , wide fluctuations can occur . The effects of the regulatory processes that are in place to deal with these fluctuations are not well characterized . Often the overall effects can be easily measured but not the changes in the microenvironment that come together to drive these effects . Although traditional tissue culture studies have added a wealth of knowledge in such areas , they often lack the capability to emulate the in vivo environment . In the study of the effect of flow in regulating vessel wall interactions , for example , three-dimensional studies can provide valuable information . Three-dimensional studies have been used previously to measure the effects of flow on cell populations [18] , [36]–[39] . We have chosen such an approach to measure the effect of flow on heparin binding protein delivery . By employing a single pass method to focus on the initial growth factor-vessel wall interaction we were able to more directly measure the effect of flow on the bioavailability of these growth factors . We measured substantial binding of all growth factors ( FGF-2 , VEGF , and EGF ) at the lowest flow rate tested ( 0 . 61–0 . 66 mL/min , 1 . 22– 1 . 32 mm/sec ) . Had a traditional two-dimensional approach been used instead , these factors would have had few limitations on their rebinding potential since in a closed system they would not be subject to the flow that would remove them from the vessel as is typical of normal circulation . In the case of the heparin binding proteins ( FGF-2 and VEGF ) , removal of heparan sulfate sites via enzyme digestion resulted in a significant increase in growth factor outflow ( i . e . non-retention within the vessel ) , suggesting an important regulatory role for these proteoglycans in ligand capture . This is not necessarily surprising given the large number of binding sites these proteoglycans provide on normal cell surfaces . Certainly , it has been shown by us and others that HSPGs are important regulators of FGF-2 binding to FGF receptors in tissue culture [28] , although not essential for the interaction [6] , [24] , [27] . Their importance with regard to capture under flow has , however , not been shown previously and suggests a critical role in the circulation . An equally significant influence on FGF-2 , VEGF , or EGF binding , regardless of heparin binding characteristics however , was the flow rate . By increasing the flow rate by less than a factor of three ( ∼1 . 8 mL/min , 3 . 6 mm/sec ) a significant increase was seen in growth factor outflow , reflecting the absence of specific binding taking place on vessel surfaces . A higher flow rate ( ∼3 . 0 mL/min , 6 mm/sec ) showed no further increase in FGF-2 outflow above that observed at the medium flow rate with both showing retention levels equivalent to that evident in the absence of heparan sulfate . This correlation of flow rate and outflow of growth factors suggests a strong regulatory effect and an environment in the bloodstream that reduces the probability of capture significantly at flow rates typically measured in arteries [4] . Although pulsatile flow is undoubtedly important in increasingly larger vessels and higher flow rates , both simulations and experiments showed that at the low flow rate typical of capillaries it had no significant effect on FGF-2 interactions when compared to steady flow . The removal of chondroitin sulfate created a small but significant increase in FGF-2 outflow . This is interesting since a number of published findings found no significant affinity between FGF-2 and chondroitin sulfate [30] , [31] . It is possible that under flow conditions subtle changes in chondroitin sulfate modifications allow for some weak interaction . Others have reported the ability of FGF-2 to bind chondroitin sulfate under certain circumstances [32]–[34] . EGF binding was , however , unaffected by treatment with a heparinase , chondroitinase and keratanase cocktail suggesting the chondroitinase effect was not universal . How this effect is manifest is currently under further study . The minimum size of capillaries has been shown to be relatively fixed across species regardless of size [40] and is a basic assumption in the general model of allometric scaling laws proposed by West et al [41] . This suggests an optimum environment for the exchange of gases , nutrients , and the removal of waste products that is likely rooted in fundamental physical laws . In order to best make use of these environmental conditions blood flow must also be optimal . Our data demonstrate an inverse correlation between flow rate and probability of capture . Although the presence of heparan sulfate is crucial to FGF-2 capture at low flow rates , at higher flow rates the overriding regulator seems to be the flow rate itself which , based on our results , would all but preclude efficient FGF-2 binding to vessel walls in a single pass under all but the slowest flow conditions . The expectation of lower binding at increasingly higher flow rates might be somewhat expected but the relatively small increase in flow rate required to ablate binding was surprising . Other influences , such as viscosity , and the presence of competing molecules were not addressed in this work . These are ongoing studies as we begin to add complexity to the system so as to form even more accurate models of circulation . The advantage of this method is that the conditions can be monitored and controlled much as two dimensional culture systems can be but include the three dimensional architecture and flow characteristics that are part of normal blood flow . This approach has obvious potential in the testing of both endogenous molecules and pharmaceuticals in order to provide a better perspective of molecular interactions in the microenvironment of blood vessels . The importance of HSPGs in FGF-2 binding and signaling has been shown in many systems [6]–[11] and is a generally accepted feature for heparin-binding growth factors . Our work builds upon those studies and shows the critical importance of HSPGs in FGF-2 capture under flow ( Figure 3 ) . In this paper , we explore the impact of this critical component in detail using our computational model and show the parameters that regulate this process . In particular we show that the two-step coupling process and the accompanying decrease in dissociation are essential for effective retention of FGF-2 in a flow situation . HSPG can mediate both the heparin-binding growth factor-receptor interaction at the cell surface and the accumulation and storage of these growth factors in the extracellular matrix [42] , [43] . Removal of HSPG from the cell surface by enzymatic digestion greatly impairs FGF-2 activity in vitro and inhibits neo-vascularization in vivo [27] , [28] , [44] . HSPG interacts with FGFR directly [45] , [46] and FGF-2 binding to cell surface HSPG can facilitate FGF-2 binding to FGFR , which in turn can result in activation of intracellular signaling cascades . Using our simple model under flow , we show in several ways that the coupling step is critical for FGF-2 retention . Elimination of coupling or decreasing the rate constant describing that interaction has a dramatic effect on both FGF-2 bound and internalized with essentially no internalization or effective binding when coupling is eliminated ( Figure 12 ) . Reducing the density of HSPG ( Figure 11 ) or the affinity of FGF-2 for HSPG ( Figure 9 ) significantly reduces the amount of FGF-2 bound to both the cell surface and to FGFR . In addition , simulations with only low levels of HSPG ( Figures 11 , 12 – entrance zone ) or FGFR ( data not shown ) do not exhibit high retention but , when both HSPG and FGFR are present ( Figure 13 ) , the combination of both increases retention . This is evident independent of flow rate . The ability of flow to regulate the level of binding suggests how crucial the presence of HSPG is on the vessel wall , in order to increase the probability of capture of heparin-binding molecules especially given the short half-lives of some growth factors in circulation . Under the flow condition , simulations predict that the majority of FGF-2 binding occurs at the entrance to the cell-lined hollow fiber ( Figure 7 ) . In our simulations set up to match the experimental conditions , FGF-2 enters at its highest concentration and thus is most likely to bind under those conditions . Once binding occurs , there is a depletion of FGF-2 in the fluid phase near the cell surface ( Figure 8 ) . Under flow , this zone can be replenished via diffusion as increasing the diffusion coefficient increases the concentration in this zone ( Figure 8 ) and ultimately leads to higher binding down the cell-lined hollow fiber . We had postulated that FGF-2 bound in the entrance zone of the cell-lined hollow fiber would eventually dissociate and rebind further down the tube but this does not appear to be the case . Even when binding is extremely high at the entrance , FGF-2 that dissociated from the entrance was not in high enough concentration to impact downstream binding and was eventually washed out of the system ( data not shown ) . In a non-flow system this would likely not be the case and exemplifies the importance of including flow in studies . In conclusion , a simulation program previously developed by us but enhanced for our specific cell investigations of FGF-2 binding under flow [16] , [17] performed well when compared to our experimental endothelial cell-lined bioreactor . Our simulations suggest that: ( 1 ) The amount of FGF-2 bound to FGFR is dominated by HSPG and the coupling rate constant , and this triad ( FGFR-HSPG-FGF-2 ) is the key to FGF-2 capture; ( 2 ) The amount of FGF-2 bound is proportional to the diffusivity of the growth factor in solution and inversely proportional to the flow rate; ( 3 ) Flow rate and diffusivity will affect the FGF-2 outflow profile and the distribution of FGF-2 bound along the cell-lined hollow fiber wall; ( 4 ) The majority of FGF-2 binding occurs in the entrance zone of the cell-lined hollow fiber; and ( 5 ) most FGF-2 effectively bound by FGFR and HSPG will be internalized rather than dissociated . The simulation environment can provide additional information and insight into capture of FGF-2 that is not easily accessible from experimental work . We have applied the model to our in vitro bioreactor system but it has potential to be used for other growth factors as well as other cell systems where flow and capture are pivotal such as in drug and biologicals delivery testing . | In this work we have investigated the role of a family of cell surface molecules , proteoglycans , in blood vessel capture of proteins important to normal and diseased states under flow conditions . We developed a computer model to analyze and predict these events and , using an experimental system incorporating endothelial-lined hollow fibers as model blood vessels , tested our predictions . We found that both proteoglycans and flow exert significant influence over growth factor binding to the vessel wall . Removal of proteoglycans significantly reduced binding of these proteins; and flow rates slightly higher than that seen in capillaries had a similar effect , albeit in a different way . This knowledge will increase our understanding of interactions inside blood vessels and help to design more efficient pharmaceuticals . Also , our computer model has the potential to test the ability of existing and future drugs and biologics to successfully target blood vessels . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"cell",
"biology",
"computational",
"biology"
]
| 2010 | Endothelial Cell Capture of Heparin-Binding Growth Factors under Flow |
Many critical events in the Plasmodium life cycle rely on the controlled release of Ca2+ from intracellular stores to activate stage-specific Ca2+-dependent protein kinases . Using the motility of Plasmodium berghei ookinetes as a signalling paradigm , we show that the cyclic guanosine monophosphate ( cGMP ) -dependent protein kinase , PKG , maintains the elevated level of cytosolic Ca2+ required for gliding motility . We find that the same PKG-dependent pathway operates upstream of the Ca2+ signals that mediate activation of P . berghei gametocytes in the mosquito and egress of Plasmodium falciparum merozoites from infected human erythrocytes . Perturbations of PKG signalling in gliding ookinetes have a marked impact on the phosphoproteome , with a significant enrichment of in vivo regulated sites in multiple pathways including vesicular trafficking and phosphoinositide metabolism . A global analysis of cellular phospholipids demonstrates that in gliding ookinetes PKG controls phosphoinositide biosynthesis , possibly through the subcellular localisation or activity of lipid kinases . Similarly , phosphoinositide metabolism links PKG to egress of P . falciparum merozoites , where inhibition of PKG blocks hydrolysis of phosphatidylinostitol ( 4 , 5 ) -bisphosphate . In the face of an increasing complexity of signalling through multiple Ca2+ effectors , PKG emerges as a unifying factor to control multiple cellular Ca2+ signals essential for malaria parasite development and transmission .
Malaria is caused by vector-born protozoan parasites of the genus Plasmodium , which cycle between mosquitoes and humans . Waves of fever arise from the synchronised egress of merozoites from erythrocytes , an event that must be followed by the invasion of fresh red blood cells ( RBCs ) for the asexual replicative cycle to continue . Precise timing of egress is crucial for parasite survival as premature or late egress leads to noninvasive merozoites [1] . Parasite transmission to mosquitoes relies on gametocytes , sexual precursor stages that are developmentally arrested in the blood but that resume their development within seconds of being taken up into a mosquito blood meal . Gametocytes respond rapidly to environmental signals including a small mosquito molecule , xanthurenic acid ( XA ) , and a concomitant drop in temperature [2] . Egress of gametes from the host erythrocyte occurs within 10 min of gametocyte ingestion by the mosquito and is followed by fertilisation . Within 24 h zygotes transform into ookinetes , which move actively through the blood meal to colonise the epithelial monolayer of the mosquito midgut . Each successful ookinete transforms into an extracellular cyst that undergoes sporogony . Eventually thousands of sporozoites are released from each cyst and invade the salivary glands of the mosquito . Once transmitted back into another human , they first replicate in the liver before invading the blood stream . This complex life cycle requires a high degree of coordination to allow the parasites to recognise and respond appropriately to stimuli from their environment . However , the underlying signal transduction pathways remain poorly understood . Reverse genetics and pharmacological studies have identified 3′-5′-cyclic guanosine monophosphate ( cGMP ) as an important second messenger for regulating the development of malaria parasites . In Plasmodium , cGMP levels are tightly controlled at the level of synthesis by two membrane-associated guanylyl cyclases ( GCs ) and degradation by four cyclic nucleotide phosphodiesterases ( PDEs ) , all of which show stage specificity in their expression [3] . GCα has resisted knockout attempts in the human malaria parasite Plasmodium falciparum [4] and in Plasmodium berghei [5] , a parasite infecting rodents , suggesting GCα is essential in asexual blood stages . In contrast gcβ and pdeδ could be deleted in asexual blood stages and the mutants revealed critical functions for both enzymes in gametocytes of P . falciparum [6] and in ookinetes of P . berghei , where the deletion of gcβ results in a marked reduction of gliding motility that could be reversed by the additional deletion of pdeδ [5] , demonstrating a key role for cGMP in regulating ookinete gliding . The only known downstream target of cGMP in malaria parasites is a cGMP-dependent protein kinase , PKG [7] , which according to current evidence is essential in asexual blood stages of P . falciparum [6] and P . berghei [5] . Work in Toxoplasma gondii and Eimeria tenella , coccidian parasites that are related to Plasmodium , identified PKG as the primary target for two structurally distinct anticoccidal compounds , the trisubstituted pyrrole compound 1 ( C1 ) and the imidazopyridine-based inhibitor compound 2 ( C2 ) . Both compounds achieve high selectivity over PKG of humans by exploiting an unusually small gatekeeper residue within the active site of all apicomplexan PKG enzymes [8] . Mutating the threonine gatekeeper residue of apicomplexan PKG to a larger residue renders parasites resistant to both inhibitors . This provided a powerful genetic tool to study PKG function , first in tachyzoites of Toxoplasma gondii , where PKG was found to be important for egress from the host cell , secretion of micronemes , and gliding motility [9] , and later in P . falciparum , where PKG was shown to be important for the initial activation of gametocytes in response to environmental triggers and for replication of asexual blood stages [4] , [10] . Inhibition of PKG in P . falciparum resulted in the accumulation of mature segmented schizonts , which did not rupture and failed to release merozoites . A C1-insensitive PKG allele also reversed inhibition of schizont rupture , ruling out off-target effects of C1 as responsible . Recently PKG was shown to operate upstream of a Ca2+-dependent protein kinase , CDPK5 [11] , and to control exocytosis of two secretory organelles , called exonemes and micronemes , which contain proteins essential for merozoite egress [1] . In mammals cGMP regulates diverse and important cellular functions , ranging from smooth muscle contractility [12] to retinal phototransduction [13] . It mediates cellular response to a range of agonists including peptide hormones and nitric oxide [14] . In Plasmodium neither the upstream regulators of cGMP signalling have been identified , nor the cellular targets and downstream effector pathways through which PKG regulates the two distinct biological processes that are schizont egress and gametocyte activation . In the present study , we generate P . berghei transgenic lines that express a resistant PKG allele . Using a chemical genetic approach we first show that PKG controls the gliding motility that ookinetes rely on to reach and penetrate the midgut epithelium of the mosquito during transmission . We then use a global analysis of protein phosphorylation by quantitative mass spectrometry to identify pathways that operate downstream of PKG in gliding ookinetes . We chose phosphoinositide metabolism as a putative effector pathway for further validation and demonstrate that PKG controls phosphoinositide synthesis including the production of phosphatidylinositol ( 4 , 5 ) -biphosphate ( PI ( 4 , 5 ) P2 ) , the precursor of inositol ( 1 , 4 , 5 ) -trisphosphate ( IP3 ) , whose synthesis triggers mobilisation of intracellular Ca2+ [15] . This leads us to hypothesise that a major function for PKG is to control intracellular Ca2+ levels in malaria parasites , through the regulation of phosphoinositide metabolism by lipid kinases . This study presents strong evidence in support of this idea by showing in three life cycle stages and two Plasmodium species that activation of PKG is critically required to regulate cytosolic Ca2+ levels . PKG emerges as a universal regulator that controls ookinete gliding , gametocyte activation , and schizont rupture .
The pkg gene appears to be essential in blood stages of P . berghei since it could not be disrupted . So far only pharmacological evidence implicates PKG as the effector kinase of cGMP in gliding ookinetes [5] . To facilitate genetic studies in P . berghei we replaced pkg with a modified allele , pkgT619Q-HA , in which the threonine gatekeeper residue was mutated to a larger glutamine residue together with a C-terminal triple HA epitope tag ( Figure S1 and Figure S2A ) . The equivalent gatekeeper mutation in P . falciparum PKG confers resistance to the selective inhibitors C1 and C2 [4] , [10] . A transgenic control line without the T619Q mutation , pkg-HA , was also generated and the resistance marker was removed from both cloned lines by negative selection to enable subsequent genetic modifications ( Figure S2A ) . We observed no effect of the T619Q mutation on asexual growth rate , gametocyte and ookinete formation , midgut oocyst numbers , salivary gland sporozoite numbers , and sporozoite infectivity to mice ( Figure S3 ) . To assess the role of PKG in gliding we recorded time-lapse movies of in vitro cultured ookinetes in thin layers of matrigel . Ookinetes expressing PKG-HA were strongly inhibited by C2 ( Figure 1A and Figure 1B ) with a half-maximal effect of ∼100 nM ( Figure 1C ) . Expression of PKGT619Q-HA , in contrast , conferred complete resistance to C2 up to at least 5 µM , demonstrating that PKG is the critical target for C2 and essential for ookinete gliding . Inhibition of PKG by C2 was as potent as disrupting cGMP production genetically by deleting gcβ ( Figure 1D ) . In contrast , interfering with degradation of cyclic nucleotides through complete deletion of pdeδ ( Figure S2B ) had the opposite effect , resulting in a marked increase in average gliding speed ( Figure 1D ) . These results show that PKG is a key effector kinase for cGMP in regulating ookinete gliding . In some animal cells , activated PKG can translocate to the nucleus and control transcription [16] , [17] . We therefore sequenced mRNA from wild-type and gcβ mutant ookinetes but failed to reveal a notable pattern of differential expression ( Figure S4A and Figure S4B ) , suggesting PKG does not regulate gene expression in ookinetes . We next designed two experiments to measure the effect of altered cGMP signalling on the global phosphorylation state of ookinete proteins using mass spectrometry . In the first experiment , we looked for long-term molecular changes in nonmotile gcβ mutant parasites as compared to gliding wild-type ookinetes ( Figure 2A ) . We used triplex stable isotope labelling in culture ( SILAC ) to measure differences between wild-type ( medium label ) and mutant parasites ( heavy label ) from five biological replicates . By performing SILAC-based quantitative proteome profiling on 1% of the material , we first identified labelled tryptic peptides from 1 , 312 proteins . Of these proteins , 763 could be quantified with high stringency , which revealed no notable differences between the mutant and the wild-type proteomes ( Figure S4C and Table S1 ) , suggesting the overall protein composition of the gcβ mutant was normal . To compare phosphorylation patterns we next performed SILAC-based quantitative phosphoproteomics using the remaining material from each replicate ( Figure 2A ) . Analysing phosphopeptides enriched by immobilised metal ion chromatography ( IMAC ) , we identified 6 , 375 phosphorylation sites , 5 , 002 of which were detected with high confidence ( class I sites according to [18] ) . Only 96 class I sites exhibited significantly altered phosphorylation in the gcβ mutant as compared with wild-type ookinetes ( Figure 2B ) . Table S1 lists ookinete proteins and their phosphorylation sites , and Table S2 shows the significantly regulated sites in the gcβ mutant . In a second experiment , we asked which ookinete proteins show rapid changes in phosphorylation when PKG is inhibited by C2 for 2 min ( Figure 2C ) . For this experiment , we exposed ookinetes expressing either PKG-HA or PKGT619Q-HA to 0 . 5 µM C2 , which blocks gliding only in PKG-HA parasites . Recognising that in the first experiment SILAC had potentially favoured the identification of regulated sites in the part of the proteome that turns over most rapidly and thus incorporates more isotope label [19] , we now opted for a label-free strategy . We detected an even larger number of 1 , 634 phosphorylated proteins , on which we mapped 7 , 277 unique phosphorylation sites with high confidence ( Table S1 ) . Data from six biological replicates lead us to conclude that 266 sites belonging to 193 different proteins were reproducibly regulated in response to inhibition of PKG ( Figure 2D and Table S2 ) . A significantly less phosphorylated site ( 6-fold , p = 0 . 03 ) in gcβ ookinetes was serine S694 in the activation loop of the kinase catalytic domain of PKG itself . Activation loop phosphorylation is a common mechanism for regulating protein kinase activity , including in mammalian PKG [20] . Down-regulation of S694 probably reflects a state of reduced PKG activity , as was expected in the gcβ mutant . In contrast , phosphorylation of PKG S694 was not affected within 2 min of adding C2 , suggesting the kinetics of PKG dephosphorylation is slow . However , C2 reduced phosphorylation of S2072 in GCβ and increased phosphorylation of S310 in PDEδ , suggesting possible mechanisms for rapid feedback regulation of cGMP levels , as happens in mammalian cells [21] , [22] , reinforcing the notion that these enzymes act in the same pathway as PKG to regulate ookinete gliding . Inhibition of PKG also resulted in a rapid 6-fold reduction in the phosphorylation of S11 in the N-terminal leader peptide upstream of the kinase domain of CDPK3 , a Ca2+-dependent protein kinase important for ookinete gliding [23] , indicating its function is linked closely with PKG . To identify mechanisms of regulation by PKG we asked which cellular pathways were enriched among the proteins with regulated phosphorylation sites ( Figure 2E , see Table S2 for gene IDs and site information ) . Treatment with C2 had the greatest impact on phosphorylation sites of inner membrane complex ( IMC ) proteins and components of the gliding motor , such as the two glideosome-associated proteins GAP45 and GAPM2 , and IMC1b . Microtubule-associated proteins were also enriched , including several dynein and kinesin-related putative motor proteins of unknown function . In marked contrast , regulated phosphosites in the gcβ mutant were most abundant in mRNA-interacting proteins involved in splicing and 3′ polyadenylation , and in components of Plasmodium P-bodies , such as the RNA helicase , DOZI ( development of zygote inhibited ) , and a putative trailer hitch homolog , CITH , which are both essential for ookinete formation by stabilising translationally repressed mRNAs in the female gametocyte [24] . Also deregulated were multiple sites in a family of Alba domain-containing proteins that form part of the DOZI snRNP complex [24] . Furthermore , regulated phosphoproteins in the gcβ mutant were enriched for components of clathrin and COPI-coated vesicles , including a putative clathrin coat assembly protein , AP180 , the beta subunit of the coatomer complex , as well as a number of putative regulators of vesicular trafficking , which together point to an important role for protein trafficking in ookinete gliding . Finally , the gcβ mutant had a notable abundance of regulated phosphorylation sites in enzymes involved in the metabolism of inositol phospholipids and their regulators . Importantly , many representatives from this group of proteins were also regulated in response to C2 ( highlighted in Figure 2B and Figure 2D ) , suggesting they may be more direct targets of PKG than some of the other regulated phosphoproteins . Given that enzymes in the inositol phospholipid biosynthetic pathway were identified by both phosphoproteomic approaches , we chose to investigate this pathway in more detail . Phosphorylated phosphatidylinositol lipids have important roles in vesicle trafficking and as a source of secondary messengers in signal transduction . Their biosynthesis from phosphatidyl-1D-myo-inositol ( PI ) is mediated by lipid kinases . The P . berghei genome encodes four putative lipid kinases to convert PI first to phosphatidylinositol 4-phosphate ( PI4P ) and then to phosphatidylinositol ( 4 , 5 ) -bisphosphate ( PI ( 4 , 5 ) P2 ) ( Figure 3 ) . Hydrolysis of the latter by a PI-specific phospholipase C ( PI-PLC ) gives rise to the secondary messenger inositol ( 1 , 4 , 5 ) -trisphosphate ( IP3 ) , which plays an important role in P . berghei gametocytes , where it is responsible for the mobilisation of Ca2+ from internal stores , leading to activation and gametogenesis [25] . All four PI kinases were detected in the ookinete phosphoproteome and three contained sites that were less phosphorylated upon inhibition of PKG or disruption of gcβ ( Figure 3 ) . An important regulator of phosphoinositide metabolism and membrane trafficking is the phosphatidylinositol transfer protein Sec14 [26] , the phosphorylation of which was reduced in closely adjacent sites in both experiments . PIP5K activity of the P . falciparum orthologue of PBANKA_020310 is controlled by a small G protein of the ADP-ribosylation factor ( ARF ) family [27] , which cycles between an inactive GDP-bound and an active GTP-bound form . In other eukaryotes , the active state of ARF results from its interaction with a guanine nucleotide exchange factor ( ARF-GEF ) that forces ARF to adopt a new GTP molecule in place of a bound GDP , whereas the inactive state results from hydrolysis of GTP facilitated by a GTPase activating protein , ARF-GAP . Putative ARF-GEF and ARF-GAP proteins are encoded in the P . berghei genome , and these also have GCβ/PKG-dependent phosphosites ( Figure 3 ) . Taken together these data led us to hypothesise that phosphoinositide metabolism is important for ookinete gliding and regulated by PKG . To test whether the PKG-dependent phosphorylation of enzymes associated with phosphoinositide metabolism has a direct role in ookinete motility , we used experimental genetics to infer the role of putative PI kinases . Both the putative PI4K ( PBANKA_110940 ) and the putative PIP5K ( PBANKA_020310 ) were unable to be genetically disrupted ( unpublished data ) , suggesting these genes may be essential for asexual growth , although both loci could be modified ( Figure S2C and Figure S2D ) . One of the most strongly down-regulated phosphorylation sites in the gcβ mutant was S534 of PI4K . To assess the importance of this residue , we generated allelic replacement constructs to mutate S534 to alanine , either on its own or in combination with a nearby phosphorylation site , S538 ( Figure S2D ) . At the ookinete stage , pi4kS534A and pi4kS534A/S538A clonal mutants showed a significant decrease in gliding speed compared with a control line , pi4kS534 ( Figure 4A and Figure S2D ) , which would be consistent with phosphorylation of S534 in PI4K contributing to the regulation of phosphoinositide metabolism in vivo . A direct link between PKG and phosphoinositide metabolism was also supported by the location of PIP5K-HA , which rapidly redistributed from the cell periphery to the ookinete cytosol in ookinetes treated with 0 . 5 µM C2 ( Figure 4B and Figure 4C ) . To define more precisely the role of PKG in regulating phosphoinositide metabolism , we examined the effect of C2 on the phospholipid composition of gliding ookinetes . Analysing total lipid extracts from purified ookinetes by mass spectrometry , we detected PI , phosphatidylcholine ( PC ) , phosphatidylethanolamine ( PE ) , phosphatidylglycerol ( PG ) , cardiolipin ( CL ) , and several other minor phospholipids . For all of these we then determined changes in their relative abundance upon inhibition of PKG ( Figure S5A and Figure S5B ) . Exposing gliding ookinetes to 0 . 5 µM C2 for 10 min resulted in a marked relative increase in PI , while peaks corresponding to PIP and PIP2 molecular species were reduced ( Figure 4D ) . No significant differences in PC , PE , CL , or PG were detected ( unpublished data ) . In control experiments , C2 had no effect on phospholipid composition of ookinetes expressing PKGT619Q-HA ( Figure 4E ) . These data suggest a link between PKG and PI4P synthesis and therefore probably with Ca2+ signalling in ookinetes . We therefore examined next whether PKG is a positive regulator of Ca2+ release . To measure Ca2+ levels in life ookinetes , we inserted an expression cassette for the free Ca2+ reporter pericam into the redundant p230p locus of the PKG-HA and PKGT619Q-HA lines ( Figure S2E ) . Pericam is a fusion protein comprising calmodulin , GFP , and the M13 peptide corresponding to the calmodulin-binding domain of skeletal muscle myosin light chain kinase [28] . Binding of Ca2+ to the EF hands of calmodulin causes the latter to interact with the M13 peptide , which in turn modulates the fluorescence properties of GFP . We expressed a ratiometric form of pericam , in which Ca2+ shifts the excitation peak from 415 to 494 nm . Dual excitation imaging detects changes in Ca2+ levels through shifts in the ratio of the Ca2+ bound to the unbound form of pericam [28] . Ca2+-bound and -unbound pericam were uniformly distributed throughout the ookinete cytosol ( Figure S6A ) and were sensitive to changes in intracellular Ca2+ induced by the Ca2+ ionophore ionomycin ( Figure S6B ) . Addition of 0 . 5 µM C2 to PKG-HA-pericam ookinetes resulted in a rapid shift within 15 s from Ca2+-bound to -unbound reporter if compared to C2-resistant PKGT619Q-HA-pericam parasites ( Figure 5A ) . There was no such response in the solvent control ( Figure 5B ) . These results demonstrate that PKG activity is critical for high cytosolic Ca2+ levels to be maintained in gliding ookinetes . It remains unknown whether ookinetes glide constitutively in vivo or whether their behaviour responds to internal or external stimuli . In contrast , gametocytes , the developmentally arrested sexual precursor stages that circulate in the blood stream in a developmentally arrested form , become activated by well-defined environmental triggers within seconds of being taken up by a feeding mosquito . Activation is mediated by a drop in temperature and the concomitant exposure to a chemical stimulus from the mosquito , XA [2] . These well-defined triggers offer an opportunity to ask how PKG and Ca2+ interact in response to physiological agonists . At a permissive temperature XA triggers PI ( 4 , 5 ) P2 hydrolysis in P . berghei gametocytes , presumably by activation of PI-PLC [25] , which results in the rapid mobilisation of Ca2+ from internal stores . In P . falciparum , on the other hand , XA enhances GC activity in membrane preparations [29] , and PKG regulates gametocyte activation [10] . If and how signalling through cGMP and Ca2+ are linked during gametocyte activation has not been addressed . To ask if in P . berghei gametocytes PKG regulates Ca2+ release in response to XA , we introduced into the dssu or cssu locus of the marker-free PKG-HA and PKGT619Q-HA lines an expression cassette for a reporter protein that is based on the Ca2+-dependent photoprotein , aequorin ( Figure S2F ) [30] . XA triggered a transient luminescence response that peaked rapidly after a characteristic lag phase 10 s after stimulation ( Figure 5C ) , as described previously [31] . In parasites expressing PKG-HA this response was dose-dependently blocked by C2 with an IC50 of around 3 µM ( Figure 5C ) . C2 acted through inhibition of PKG , because the PKGT619Q-HA-GFPaeq line was completely resistant to even higher concentrations of the inhibitor . We conclude that the rapid activation of PKG within seconds of exposing gametocytes to their natural agonist mediates gametocyte activation through the mobilisation of Ca2+ and is therefore most likely a key event for the transmission of Plasmodium to the mosquito . To explore whether Ca2+ release via PKG signalling regulates blood stages in Plasmodium parasites , we turned to the major human pathogen P . falciparum . In asexual blood stages of P . falciparum PKG is critically required for schizonts to rupture and for merozoites to egress [4] . C1 and C2 were shown to block schizont rupture by inhibiting PKG , while conversely , an inhibitor of cGMP-PDE , zaprinast , raises cellular cGMP levels and triggers premature egress through the rapid discharge of micronemes and exonemes from the intracellular parasite that is strictly dependent on parasite PKG [1] . Since discharge of secretory organelles by Plasmodium schizonts also depends on Ca2+ we asked whether activation of PKG by zaprinast triggers egress by controlling intracellular Ca2+ levels . In synchronised P . falciparum schizonts loaded with the fluorescent Ca2+ sensor Fluo-4 , addition of 100 µM zaprinast led to a marked increase in free cytosolic Ca2+ within 20 s ( Figure S6C ) . This rapid Ca2+ response was mediated by PKG , because it could be inhibited by the simultaneous administration of C2 with a half-maximal effect of ∼2 µM ( Figure 6A ) . Importantly , C2 was completely ineffective in blocking zaprinast-induced Ca2+ release in parasites expressing a resistant T618Q allele of PfPKG . To ask if PKG controlled schizont Ca2+ mobilisation by regulating phosphoinositide metabolism , we studied the schizont lipidome ( Figure S5C ) . Zaprinast triggered a rapid depletion of PIP , PIP2 , and PIP3 , and a concomitant rise in PI from total lipid extracts within 10 s , consistent with the kinetics of Ca2+ mobilisation by the same treatment . Pre-exposing infected erythrocytes to 1 µM C2 blocked the zaprinast-induced depletion of phosphorylated PIs , but only in parasites expressing wild-type PKG and not in parasites expressing the C2-resistant T618Q allele of PKG ( Figure 6B ) . Uninfected erythrocytes , in contrast , contained very little phosphorylated PIs ( Figure S5D ) . Taken together , these data show that activation of PKG by the PDE inhibitor zaprinast triggers a cellular Ca2+ response in P . falciparum schizonts that is accompanied by the rapid initiation of PIP2 hydrolysis .
All biological processes in apicomplexan parasites known to require PKG are also thought to rely on the release of Ca2+ from intracellular stores [11] , [31] , [32] , [35] , [36] , which in turn activates distinct stage-specific effector pathways including members of a family of plant-like Ca2+-dependent protein kinases ( CDPKs ) [37] . Merozoite egress requires CDPK5 [11] , cell cycle progression to S-phase in activated male gametocytes is mediated by CDPK4 [31] , and ookinete gliding relies on CDPK3 [23] , [35] . By combining a range of genetically encoded and chemical Ca2+ reporter systems with PKG gatekeeper mutants in both P . berghei and P . falciparum , this study demonstrates clearly that PKG controls cytosolic Ca2+ levels in all these life cycle stages . The critical importance of PKG upstream of parasite cytosolic Ca2+ is thus of universal relevance to parasite development in blood and transmission stages , as it holds true for agonist-induced signalling in gametocytes , in constitutively gliding ookinetes , and after artificial PKG activation triggered by a PDE inhibitor in erythrocytic schizonts . Cross-talk between second messengers is common in eukaryotic cells . In mammalian vascular smooth muscle cells , for instance , cGMP regulates cytosolic Ca2+ negatively , chiefly through PKG1β , which reduces Ca2+ release from internal stores by phosphorylating the IP3 receptor in the ER membrane [38] . Phototransduction , in contrast , relies on a direct inhibitory interaction of cGMP with a cyclic nucleotide-gated cation channel in the plasma membrane , which appears to have been lost from the apicomplexan genomes during evolution [39] . We here present evidence for a positive interaction between cGMP and Ca2+ signalling in malaria parasites that invokes a different mechanism by involving regulation of phosphoinositide metabolism . Studying gametocyte activation in P . berghei , we previously identified hydrolysis of PIP2 and generation of IP3 by PI-PLC as a critical event upstream of Ca2+ mobilisation by XA [25] . IP3-dependent Ca2+ release also operates in P . falciparum blood stages [34] , although genetic evidence for an IP3 receptor in Plasmodium is still missing . Because canonical G-protein coupled receptors and heterotrimeric G-proteins , which typically regulate PI-PLC , are absent from the genomes of Plasmodium species , it has remained unclear how IP3 production in apicomplexan parasites is regulated . Our lipidome analysis of P . falciparum schizonts provides strong biochemical evidence that activating PKG with the help of a PDE inhibitor that raises cellular levels of cGMP [1] leads to the rapid and PKG-dependent hydrolysis of PIP2 at the same time as cellular Ca2+ increases sharply . This would be consistent with a role for PKG for the activation of PI-PLC . Whether this involves phosphorylation remains unknown , as our proteometric studies failed to detect PI-PLC with confidence . Intriguingly , inhibiting PKG in gliding ookinetes , where Ca2+ is already elevated , revealed a different link between PKG and phosphoinositide metabolism . In this situation , inhibiting PKG did not cause PIP2 to accumulate , as would be expected if its primary role was to promote PI-PLC activity . Instead , inhibition of PKG revealed phosphorylation of PI as a rate-limiting step . More tentative evidence that in ookinetes PKG regulates signalling at the point of PI phosphorylation comes from two additional observations . First , mutations in phosphorylated serine residues in PI4K reduce ookinete gliding speed . Second , a type I PIP5K of ookinetes localises to the cell periphery in a PKG-dependent manner . Dissociation of PIP5K from the plasma membrane or possibly the IMC of the ookinete could merely be a consequence of substrate depletion [40] . On the other hand , PIP5K , the P . falciparum ortholog of which encodes a functional type I PIP5K that is activated by ARF [27] , may also provide a more active link to cGMP signalling , as it contains EF-hand-like motifs of a kind typically found in the neuronal Ca2+ sensor family of proteins , which intriguingly can function as activators of membrane GCs in other eukaryotes [41] . More work is clearly required to evaluate lipid kinases and PI-PLC as direct substrates of PKG; to establish their precise roles , and those of their products , in linking PKG to Ca2+ in Plasmodium; and to determine their relative importance in schizonts , ookinetes , and gametocytes . A recent study has proposed that in the mammalian central nervous system PKG regulates presynaptic vesicle endocytosis by controlling PI ( 4 , 5 ) P2 synthesis indirectly via the small GTPase RhoA and Rho kinase [42] . However , a different mechanism must operate in apicomplexan parasites , which appear to lack a Rho signalling pathway . Our identification of PKG as a positive upstream regulator of cytosolic Ca2+ levels extends current models of how schizonts rupture in the bloodstream and how XA activates gametocytes in the mosquito . Importantly , it puts a spotlight on the GCs and PDEs that control cGMP levels in the parasite and which must provide the next layer of regulation for some of the key events in the life cycle of malaria parasites . Apicomplexan zoites glide with the help of transmembrane adhesins , which are secreted apically from micronemes , and then translocate posteriorly , where they are shed and left behind in a trail of vesicles that also contain other membrane proteins [43] . Sustained gliding must be fuelled by the continuous synthesis of substantial quantities of proteins and lipids that need to be trafficked to new micronemes , which are then transported to the apical pole of the ookinete for regulated secretion . In the gcβ mutant the striking abundance of regulated phosphoproteins with likely functions in vesicular transport ( Figure 7A ) , but also of chaperones ( Figure 2E ) , reflects a profoundly dysregulated secretory system ( Figure 7B ) , which may result from the role of PKG as a regulator of phosphoinositide metabolism in at least two ways . First , the changed phosphorylation pattern in vesicular transport proteins may result from a “traffic jam” following a block in IP3/Ca2+-dependent regulated secretion of micronemes at the apical end of the ookinete . Second , because in other eukaryotes PIP and PIP2 control many essential cellular processes in regulating membrane dynamic and vesicular trafficking [44] , PKG-mediated changes in PI phosphorylation could alter vesicular transport more directly and at different stages in the secretory pathway . Microneme biogenesis is poorly understood in Plasmodium , but work in T . gondii tachyzoites has identified endosomal sorting signals that traffic micronemal proteins from the Golgi to an endosomal-like compartment before they are packaged into micronemes . In the absence of either dynamin-related protein B ( DrpB ) or the VPS10/sortilin homolog TgSORTLR , proteins destined for micronemes fail to be targeted from the Golgi to secretory organelles and instead enter the constitutive secretion pathway . TgSORTLR is thought to be an essential cargo receptor to transport microneme and rhoptry proteins to endosomal-like compartments of the T . gondii tachyzoite . In other eukaryotes retrograde transport of sortilin to the Golgi relies on its interaction with the conserved retromer complex . We here report regulated phosphorylation for P . berghei homologues for another retromer component , VPS35 , as well as for DrpB ( T690 and S736 , respectively; Figure 7A ) . Similarly , a VPS9 homolog we found regulated in S767 is a likely conserved activator of Rab5 GTPases , which is relevant because in T . gondii Rab5A and Rab5C are essential for targeting proteins to unique subsets of micronemes [45] . Although many of the regulated phosphosites we report here are probably not phosphorylated directly by PKG , these examples illustrate that our study provides a rich source of leads for future research into gliding motility and microneme biogenesis in Plasmodium . Our data demonstrate that by regulating phosphoinositide metabolism PKG controls multiple essential cellular processes including critical Ca2+ signals and probably vesicular trafficking . Members of a promising class of new antimalarials , the imidazopyrazines , target PI4K [46] . Our analysis provides a rationale for the activity of imidazopyrazines against multiple life cycle stages of Plasmodium and highlights how a stage transcending signalling pathway can regulate different critical steps during parasite development . Future work will have to refine the knowledge about the nature of PKG-mediated signalling by identifying direct PKG substrates and the genetic or environmental factors controlling PKG activity .
All animal experiments were conducted under a license from the UK Home Office in accordance with national and European animal welfare guidelines . P . berghei ANKA wild-type strain 2 . 34 and transgenic lines made in the same background were maintained in female Theiler's Original outbred mice and infections monitored on Giemsa-stained blood films . Exflagellation was quantified 3 d postinfection by adding 4 µl of blood from a superficial tail vein to 150 µl exflagellation medium ( RPMI 1640 containing 25 mM HEPES , 4 mM sodium bicarbonate , 5% FCS , 100 µM XA , pH 7 . 4 ) . Between 13 and 16 min after activation , the number of exflagellating microgametocytes was counted in a haemocytometer and the RBC count determined . The percentage of RBCs containing microgametocytes was assessed on Giemsa-stained smears , and the number of exflagellations per 100 microgametocytes was then calculated . For Ca2+ assays , gametocytes were separated from uninfected erythrocytes on a nycodenz cushion made up from 48% of a nycodenz stock ( 27 . 6% w/v Nycodenz in 5 . 0 mM Tris-HCl [pH 7 . 20] , 3 . 0 mM KCl , 0 . 3 mM EDTA ) and RPMI1640 medium containing 25 mM HEPES , 5% FCS , 4 mM sodium bicarbonate , pH 7 . 30 . Gametocytes were harvested from the interphase . For ookinete cultures , parasites were maintained in phenyl hydrazine-treated mice . Ookinetes were produced in vitro by adding one volume of high gametocyteamia blood in 20 volumes of ookinete medium ( RPMI1640 containing 25 mM HEPES , 10% FCS , 100 µM XA , pH 7 . 5 ) and incubated at 19°C for 16–18 h . For the gcβ mutant experiment , customized RPMI1640 medium ( Invitrogen ) containing 13C6 , 15N2 L-lysine and 13C6 , 15N4 L-arginine ( gcβ mutant ) or D4 L-lysine and 13C6 L-arginine ( wild-type ) was used . Conversion efficiency was determined by live staining of ookinetes and activated macrogametes with Cy3-conjugated 13 . 1 monoclonal antibody against p28 . The conversion rate was determined as the number of banana-shaped ookinetes as a percentage of the total number of Cy3-fluorescent cells . For biochemical analysis and Ca2+ assays , ookinetes were purified using paramagnetic anti-mouse IgG beads ( Life Technologies ) coated with anti-p28 mouse monoclonal antibody ( 13 . 1 ) . For motility assays , ookinete cultures were added to an equal volume of Matrigel ( BDbioscience ) containing DMSO or C2 on ice , mixed thoroughly , dropped onto a slide , covered with a cover slip , and sealed with nail polish . After identifying a field containing ookinetes , time-lapse videos were taken ( 120×; 1 frame every 20 s , for 20 min ) on a Leica M205A at 19°C . Movies were analysed with Fiji and the Manual Tracking plugin ( http://pacific . mpi-cbg . de/wiki/index . php/Manual_Tracking ) . For transmission experiments batches of ∼50 female Anopheles stephensi , strain SD500 , mosquitoes were allowed to feed on infected mice 3 d after intraperitoneal injection of infected blood . Unfed mosquitoes were removed the day after . Infected mosquitoes were maintained on fructose at 19°C and oocyst numbers were counted on dissected midguts 7 d after feeding . Sporozoite numbers were determined on day 21 by homogenising dissected salivary glands and counting the released sporozoites . To determine sporozoite infectivity to mice 21 d after infection , infected mosquitoes were allowed to feed on naïve mice or 20 , 000 freshly isolated sporozoites in RPMI 1640 containing 1% penicillin/streptomycin were injected in a volume of 100 µl into the tail vein . Mice were then monitored daily for blood stage parasites . Tagging , knockout , and allelic replacement constructs were generated using phage recombinase mediated engineering in Escherichia coli TSA ( Figure S1 ) ; PlasmoGEM vectors PbG01-2397b11 , PbG02_C-11d08 , PbG02_C-25e06 , and PbG01-2399h12 containing PBANKA_100820 ( PKG ) , PBANKA_110940 ( PI4K ) , PBANKA_020310 ( PIP5K ) , and PBANKA_100820 ( PDEä ) encoding genes , respectively , ( Figure S2 ) were generated from genomic DNA library clones in linear pJAZZ-OK vectors ( Lucigen ) as described ( Ref . 47 and http://plasmogem . sanger . ac . uk/ ) . Oligonucleotides used are shown in Table S3 . Point mutations were introduced with a two-step strategy using λ Red-ET recombineering in E . coli . The first step involved the insertion by homologous recombination of a Zeocin-resistance/Phe-sensitivity cassette surrounded by sequences 5′ and 3′ of the codon of interest amplified using primer pairs specific for the gene of interest ( GOI ) : goi-delF/goi-delR . Recombinant bacteria were then selected on Zeocin . After verification of the recombination event by PCR , a second round of recombination exchanged the Zeocin-resistance/Phe-sensitivity cassette with a PCR product containing the desired mutation surrounding the codon of interest amplified using goi-mutF/goi-mutR primer pairs . Bacteria were selected on YEG-Cl kanamycin plates . Mutations were confirmed by sequencing vectors isolated from colonies sensitive to Zeocin . Generation of knockout and tagging constructs was performed using sequential recombineering and gateway steps as previously described [47] . Lambda red-ET recombineering was first used to introduce a bacterial selection marker amplified into the gDNA insert , such that the target gene is either deleted or prepared for 3′-end tagging . The bacterial marker was then replaced with a selection cassette for P . berghei in a Gateway LR Clonase reaction in vitro . The modified library inserts were then released from the plasmid backbone using NotI and used to transfect P . berghei . For the pericam construct , the P . berghei hsp70 promoter ( PBANKA_071190 ) was first cloned using SacII/XhoI and hsp70-F/hsp70-R primers into a p230p targeting vector , which also contains a resistance cassette encoding human DHFR [48] . This was followed by the pericam coding sequence [28] downstream of the promoter , using XhoI only and primers pc-F and pc-R . The vector was linearised using HindIII/EcoRI . The expression vector for the GFP-aequorin chimeric gene was transfected in PKG-HA and PKGT619Q-HA parasites as previously described [31] . Schizonts for transfection were purified from overnight cultures and transfected with 1–5 µg of linearised DNA as previously described [49] . Electroporated merozoites were injected intraveinously into a naïve mouse . Resistant parasites were selected by pyremithamine supplied in the drinking water . All transgenic parasites were cloned by limiting dilution cloning , and correct integrations of the targeting constructs were verified by diagnostic PCRs and sequencing . Negative selection of PKG-HA and PKGT619Q-HA parasites expressing yFCU was performed through the administration of 5 fluorocytosine via the drinking water [50] , and dilution cloning was subsequently carried out allowing for further genetic modifications . Fluorescence measurements of purified ookinetes expressing Pericam were performed on single cells . Purified ookinetes were immobilised on Lab-Tek II chambered coverglass slides using Cell-Tak as described previously [51] . Immobilised ookinetes were imaged in PBS ( Ca2+ and Mg2+ free ) supplemented with 10 mM glucose . Confocal images were acquired with a LSM510 laser scanning confocal microscope ( Zeiss ) , a Plan-Apochromat 63×/1 . 4 oil Ph3 objective and a LP505 filter using 405 nm or 488 nm excitation . For time lapse image acquisition , cells were imaged with a Fluar 40×/1 . 30 oil objective on a Zeiss Axiovert 200 M inverted microscope equipped with an AxioCam MRm camera using the multidimensional acquisition mode of the Axiovision software . Reporter fluorescence was monitored using a 21HE FURA filter set ( exposure time , 50 ms ) and a 38HE filter set ( exposure time , 300 ms ) at a rate of 1 frame per 5 s for 35 cycles . The fluorescence of individual ookinetes was subsequently analysed using the Axiovision Physiology Module . Aequorin reconstitution and luminometric Ca2+ detection from purified P . berghei gametocytes were performed as previously described [31] . First , purified gametocytes were washed 3 times in coelenterazine loading buffer ( CLB - PBS , 20 mM HEPES , 20 mM Glucose , 4 mM sodium bicarbonate , 1 mM EGTA , 0 . 1% w/v bovine serum albumin , pH 7 . 2 ) . Reconstitution was then achieved by shaking ∼108 gametocytes , in 0 . 5 ml CLB , supplemented with 5 µM coelenterazine for 30 min at 19°C . Loaded gametocytes were washed twice in CLB and were then suspended in 10 ml RPMI 1640 , 5% FBS , 4 mM sodium bicarbonate , pH 7 . 2 . For luminescence measurements , 150 µl of the gametocyte suspension were injected into the same volume of ookinete medium containing DMSO or C2 in a 96-well assay plate of an Orion II microplate system luminometer . For each sample 50 luminescence readings were acquired over 35 s . Changes in the levels of intracellular free Ca2+ were measured using Fluo-4 ( Sigma ) loaded P . falciparum late stage schizonts that had been cultured in albumax II ( Fisher Scientific ) and human A+ erythrocytes ( washed whole blood from the National Blood service ) as previously described [4] . Excitation was measured using a SPECTRAmax microplate fluorometer . Levels of free Ca2+ were compared to a baseline read prior to the addition of a test reagent . Schizonts were purified magnetically ( Macs; Milteny Biotec ) and pelleted by centrifugation for 2 min at 500 g . Parasites were resuspended in 10× warm Ringer Buffer ( 122 . 5 mM NaCl , 5 . 4 mM KCl , 0 . 8 mM MgCl2 , 11 mM HEPES , 10 mM D-Glucose , 1 mM NaH2PO4 ) to 1–2×108 parasites/ml , and 2 µl of 5 mM Fluo-4 was added to 1 ml of parasite preparation . Cells were incubated with Fluo-4 at 37°C for 45 min and washed twice in warm Ringer buffer and incubated for 20 min for de-esterification followed by a further two washes . The pellet was resuspended in Ringer buffer and plated out on the bottom half of a 96-well plate . Compound dilutions were plated out on the top of the plate and the excitation of the cells measured at 20 s intervals for a period of 3 min to achieve a baseline read . The cells were then transferred onto the test compound and read for a further 5 min . Ookinete immunofluorescence assays were performed as previously described [52] . For HA staining after fixation with 3% paraformaldehyde in PBS , ookinetes were permeabilised with 0 . 1% Triton X-100/PBS and blocked with 2% BSA/PBS . Primary antibodies were diluted in blocking solution ( rat anti-HA , 1∶200 ) . Anti-rat Alexa488 was used as a secondary antibody together with DAPI ( all from Life Technologies ) , all diluted 1∶200 in blocking solution . Confocal images of ookinetes were acquired with a LSM510 laser scanning confocal microscope ( Zeiss ) . Approximately 10 µg of wild-type and gcβ mutant RNA were extracted in duplicate from purified ookinetes using the RNeasy kit ( Qiagen ) . Depletion of ribosomal RNA and highly abundant transcripts and sequencing library construction were performed as previously described [53] . The library was end-sequenced on an Illumina GAII instrument . TopHat [54] was used to map the Illumina reads against the P . berghei ANKA reference genome . Read counts and reads per kilo base per million mapped reads ( RPKM ) values were calculated for each gene . Differential expression was analysed using DESeq [55] . Data plots were created using R , and Artemis [56] was used to visualise transcriptome data . For the gcβ mutant proteome profiling , SILAC-based quantitative proteome profiling was performed essentially as previously described [19] . Briefly , 40 µg of total protein was used for each of the five replicates ( 20 µg from wild-type ( K4/R6 ) and 20 µg gcβ mutant ( K8/R10 ) ) pooled . Protein gels were stained with colloidal Coomassie blue , and each lane was excised and cut into 12 bands that were destained and in-gel digested overnight using trypsin . Extracted peptides were suspended using 0 . 5% formic acid were analysed online using an Ultimate 3000 Nano/Capillary LC System ( Dionex ) coupled to an LTQ Orbitrap Velos hybrid mass spectrometer ( Thermo Electron ) equipped with a nanospray ion source . Peptides were desalted online using a micro-Precolumn cartridge ( C18 Pepmap 100 , LC Packings ) and then separated using a 70 min RP gradient ( 4%–32% acetonitrile/0 . 1% formic acid ) on a BEH C18 analytical column ( 1 . 7 µm , 75 , 235 µm id×10 cm , Waters ) and analysed using a Top10 CID method . For the phosphoproteomic analyses , ∼5 mg of total proteins were extracted as previously described [19] from ∼5 . 108 purified ookinetes for each of the five and six biological replicates of the gcβ mutant and C2 experiments , respectively . Solubilised proteins were processed according to the FASP procedure [57] , [58] and digested with Trypsin Gold ( Promega ) . Peptides were collected by centrifugation and addition of ammonium bicarbonate and further desalted using Sep-Pak Light C18 cartridge ( Waters ) . IMAC purifications were performed as described previously [59] , with the following modifications: Peptides were resuspended in IMAC loading buffer ( 50% acetonitrile , 0 . 1% TFA ) and incubated with pre-equilibrated Phos-Select beads ( Sigma ) for 1 h at room temperature . The beads were then transferred to a TopTip ( Glygen ) and washed once with IMAC loading buffer , 1% acetic acid , and then water . Phosphopeptides were eluted with 100 µl ammonia water pH 11 and acidified using formic acid . Phosphopeptide samples were analysed online using an Ultimate 3000 Nano/Capillary LC System ( Dionex ) coupled to an LTQ Orbitrap Velos hybrid mass spectrometer ( Thermo Scientific ) equipped with a nanospray ion source . Data were analysed using MaxQuant version 1 . 0 . 13 . 13 and Mascot server 2 . 2 ( Matrix Science ) and MaxQuant version 1 . 3 . 0 . 5 for the gcβ mutant experiment and C2 experiment , respectively [60] . MaxQuant processed data were searched against a combined mouse and P . berghei protein identified by the GeneDB database . A protein false discovery rate ( FDR ) of 0 . 01 and a peptide FDR of 0 . 01 were used for identification level cutoffs . Class I phosphorylation site was defined with a localisation probability of >0 . 75 and a score difference of >5 [18] . For the gcβ mutant experiment , phosphorylation sites with a p value for detection of significant outlier ratio ≥0 . 01 , a ratio count ≥6 , and at least a 3-fold change were defined as being regulated in the mutant . For the C2 experiment , data were filtered , and two-sample t testing was performed with a permutation-based FDR calculation in Perseus ( 1 . 3 . 0 . 4 ) as described previously [61] . Phosphorylation sites within an FDR of 0 . 05 with at least a 1 . 5-fold change were defined as being regulated by C2 treatment . See Text S1 for details of protein extraction , data acquisition , and analysis . Total lipids were extracted using a modified Bligh and Dyer method from ∼5 . 108 purified P . berghei ookinetes or 50 µl of packed cell volume of purified P . falciparum schizont per replicate . Ookinetes were first washed with PBS , suspended in 100 µl PBS , and transferred to a glass tube . Cells were lysed with 375 µl of 1∶2 ( v/v ) chloroform∶methanol and vortexed for 15 min . Samples were made biphasic by the addition of 125 µl of CHCl3 and 125 µl of H2O . After centrifugation at 1 , 000 g at room temperature for 5 min , the lower phase was transferred to a new glass vial and dried under nitrogen . Lipids were subsequently dissolved in a mixture of chloroform∶methanol ( 1∶2 ) and acetonitrile∶iso-propanol∶water ( 6∶7∶2 ) and analysed with a Absceix 4000 QTrap , a triple quadrupole mass spectrometer equipped with a nanoelectrospray source . Samples were delivered using either thin-wall nanoflow capillary tips or a Nanomate interface in direct infusion mode ( 125 nl/min ) . The lipid extracts were analysed in both positive and negative ion modes using a capillary voltage of 1 . 25 kV . Tandem mass spectra ( MS/MS ) scanning ( daughter , precursor , and neutral loss scans ) was performed using nitrogen as the collision gas with collision energies between 35 and 90 V . Each spectrum encompasses at least 50 repetitive scans . Tandem mass spectra ( MS/MS ) were obtained with collision energies as follows: 35–45 V , PC/SM in positive ion mode , parent-ion scanning of m/z 184; 35–55 V , PI/IPC in negative ion mode , parent-ion scanning of m/z 241; 35–65 V , PE in negative ion mode , parent-ion scanning of m/z 196; 20–35 V , PS in negative ion mode , neutral loss scanning of m/z 87; and 40–90 V , all glycerophospholipids ( including PA , PG , and CL ) were detected by precursor scanning for m/z 153 in negative ion mode . MS/MS daughter ion scanning was performed with collision energies between 35 and 90 V . Assignment of phospholipid species is based upon a combination of survey , daughter , precursor , and neutral loss scans , as well as previous assignments [62] . The identity of phospholipid peaks was verified using the LIPID MAPS: Nature Lipidomics Gateway ( www . lipidmaps . org ) . The enrichment analysis was based on a list of 3 , 167 P . falciparum genes assigned to 212 metabolic pathways and functional groups present in Malaria Parasite Metabolic Pathways database ( http://priweb . cc . huji . ac . il/malaria , June 2013 ) . Annotations were transferred to P . berghei orthologs identified by the GeneDB database ( http://www . genedb . org ) . Enrichment for pathways in sets of genes was calculated using a one-tailed Fisher's exact test . | Malaria , caused by Plasmodium spp . parasites , is a profound human health problem . Plasmodium parasites progress through a complex life cycle as they move between infected humans and blood-feeding mosquitoes . We know that tight regulation of calcium ion levels within the cytosol of the parasite is critical to control multiple signalling events in their life cycle . However , how these calcium levels are controlled remains a mystery . Here , we show that a single protein kinase , the cGMP-dependent protein kinase G ( PKG ) , controls the calcium signals that are critical at three different points of the life cycle: ( 1 ) for the exit of the merozoite form of the parasite from human erythrocytes ( red blood cells ) , ( 2 ) for the cellular activation that happens when Plasmodium sexual transmission stages are ingested by a blood-feeding mosquito , and ( 3 ) for the productive gliding of the ookinete , which is the parasite stage that invades the mosquito midgut . We provide initial evidence that the universal role of PKG relies on the production of lipid precursors which then give rise to inositol ( 1 , 4 , 5 ) -trisphosphate ( IP3 ) , a messenger molecule that serves as a signal for the release of calcium from stores within the parasite . This signalling pathway provides a potential target to block both malaria development in the human host and transmission to the mosquito vector . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"second",
"messenger",
"system",
"mechanisms",
"of",
"signal",
"transduction",
"cgmp",
"signaling",
"host-pathogen",
"interaction",
"microbiology",
"parasitology",
"parasite",
"physiology",
"phosphoinositide",
"signal",
"transduction",
"crosstalk",
"cell",
"movement",
"signaling",
"signaling",
"pathways",
"calcium",
"signaling",
"protein",
"kinase",
"signaling",
"cascade",
"signaling",
"in",
"cellular",
"processes",
"biology",
"phospholipid",
"signaling",
"cascade",
"pathogenesis",
"calcium",
"signaling",
"cascade",
"signal",
"transduction",
"calcium-mediated",
"signal",
"transduction",
"molecular",
"cell",
"biology",
"signaling",
"cascades"
]
| 2014 | Phosphoinositide Metabolism Links cGMP-Dependent Protein Kinase G to Essential Ca2+ Signals at Key Decision Points in the Life Cycle of Malaria Parasites |
Antiretroviral therapy ( ART ) has reduced morbidity and mortality in HIV-1 infection; however HIV-1-associated neurocognitive disorders ( HAND ) persist despite treatment . The reasons for the limited efficacy of ART in the brain are unknown . Here we used functional genomics to determine ART effectiveness in the brain and to identify molecular signatures of HAND under ART . We performed genome-wide microarray analysis using Affymetrix U133 Plus 2 . 0 Arrays , real-time PCR , and immunohistochemistry in brain tissues from seven treated and eight untreated HAND patients and six uninfected controls . We also determined brain virus burdens by real-time PCR . Treated and untreated HAND brains had distinct gene expression profiles with ART transcriptomes clustering with HIV-1-negative controls . The molecular disease profile of untreated HAND showed dysregulated expression of 1470 genes at p<0 . 05 , with activation of antiviral and immune responses and suppression of synaptic transmission and neurogenesis . The overall brain transcriptome changes in these patients were independent of histological manifestation of HIV-1 encephalitis and brain virus burdens . Depending on treatment compliance , brain transcriptomes from patients on ART had 83% to 93% fewer dysregulated genes and significantly lower dysregulation of biological pathways compared to untreated patients , with particular improvement indicated for nervous system functions . However a core of about 100 genes remained similarly dysregulated in both treated and untreated patient brain tissues . These genes participate in adaptive immune responses , and in interferon , cell cycle , and myelin pathways . Fluctuations of cellular gene expression in the brain correlated in Pearson's formula analysis with plasma but not brain virus burden . Our results define for the first time an aberrant genome-wide brain transcriptome of untreated HAND and they suggest that antiretroviral treatment can be broadly effective in reducing pathophysiological changes in the brain associated with HAND . Aberrantly expressed transcripts common to untreated and treated HAND may contribute to neurocognitive changes defying ART .
HAND is a common complication of HIV-1 infection in the nervous system presenting a varied spectrum of clinical manifestations with cognitive , motor and behavioral symptoms . Currently three conditions of increasing severity are recognized as components of HAND: HIV-1-associated asymptomatic neurocognitive impairment , HIV-1-associated mild neurocognitive disorders ( MND ) , and HIV-1-associated dementia ( HIV-D or HAD ) [1] . HAND remains prevalent in populations with access to highly active ART , despite the efficacy of these therapies in controlling viral load and ameliorating viral load-associated clinical and neuroradiologic abnormalities [2]–[6] . By some accounts , up to 50% of HIV-1-infected individuals will develop some form of HAND regardless of access to currently available ART [7]–[9] . Under ART , HAND has became milder , its course more protracted and variable in symptoms , and it now overlaps with aging processes and potentially with other neurodegenerative diseases in AIDS patients [7] , [10] , [11] . The etiology of persistent cognitive deficits in patients on ART remains unclear . Studies of cerebrospinal fluid from individuals with HAND have sometimes supported contradictory conclusions; in the absence of viral replication , some authors observe associations among dementia , abnormal neurometabolites , and inflammatory phenotypes; others , with non-inflammatory states or markers of neurodegeneration [12]–[16] . There have been limited studies of brain tissues from treated patients with HAND to evaluate what biologic pathways remain abnormally regulated under ART , with studies largely focused on single cell types or molecules [17] . Comprehensive analysis of the spectrum of molecular abnormalities in the brain that may underlie HAND in the presence of ART has not yet been undertaken . Here we used functional genomics to conduct comparative analysis of genome-wide gene expression profiles in brain tissues from treated and untreated patients who died with HAND . Functional genomics has been applied with success to identification of complex molecular pathways in carcinogenesis and to better detection , classification , and prognosis of some cancers [18]–[23] . This approach has also been used extensively to investigate the transcriptome correlates of HIV-1 infection in peripheral tissues from HIV-1-infected patients including lymph nodes [24]–[26] , CD4+ T cells [27] , [28] , monocytes [29] , B cells [30] , and gastrointestinal mucosa [31] . Some of these studies determined the effects of ART on HIV-1 transcriptomes in patients , revealing categories of treatment-responsive genes as well as aberrantly expressed transcripts that may serve as targets for future therapies [24] , [27]–[31] . In a related approach , a recent study evaluated gene expression profiles of blood monocytes as a function of ART and neuropsychological impairment of HIV-1-infected patients [32] . Interestingly in this case , there was no correlation between changes in blood monocyte transcriptomes under treatment and clinical HAND [32] . Generally , this research benefited from the ability to serially sample peripheral tissues and individual cell types in living individuals , allowing ongoing evaluation of treatment . In contrast , analyses of human brain tissues by functional genomics can only be conducted retrospectively in autopsy tissues and are complicated by the multicellular interactions underlying central nervous system diseases [33] . Nonetheless , large-scale , cross-sectional gene expression profiling of brain tissues has revealed potential pathogenic pathways in Alzheimer's disease ( AD ) [34] , [35] , Parkinson's disease [36] , [37] , chronic schizophrenia [38] , multiple sclerosis [39] , and viral encephalitis [40] . With respect to HIV-1 infection in the brain , investigators reported transcriptional changes in selected gene categories such as anion channels in the frontal cortex of patients who died with HAND [41] , [42] . Another group reported aberrant expression of genes specific to HIV-1 encephalitis ( HIVE ) [43] , established a correlation between use of methamphetamine and up-regulation of interferon genes in these patients [44] , and investigated the role of microRNA in gene regulation in HIV-1-infected brain [45] . To our knowledge , these studies did not consider the effects of ART on brain gene dysregulation in HAND . The Manhattan HIV Brain Bank ( MHBB; member of the National NeuroAIDS Tissue Consortium ) follows a cohort of advanced-stage , HIV-1-infected individuals with a high prevalence of well-characterized cognitive dysfunction; with entry to the study , participants agree to be organ donors upon death . The antiviral treatment status of study participants is monitored while in the program . Thus , brain tissues obtained by this program provide an opportunity to examine the potentially diverse processes underlying HAND in the ART era . We report herein the gene expression profiles of individuals with HAND focusing on the impact of ART on these profiles .
We assayed virus burden and cellular gene expression on archived brain tissues from 15 HIV-1-infected patients with HAND and six HIV-1-negative subjects with no neurological or neuropathological abnormalities . Information for this study group is summarized in Table 1 and Methods . All assays were performed on parallel samples from deep white matter within the anterior frontal lobe , an area implicated in HAND and HIV-1-associated neuropathologies [46] , [47] . Custom consensus primers based on sequences of HIV-1 amplified from the brain were used for reliable measurement of HIV-1 in the brain by quantitative real-time PCR ( QPCR ) ( see Methods ) ; the results were confirmed by standard PCR and hybridization with a specific probe ( Table 1 and Supplementary Figure S1 ) . Pre-mortem plasma HIV-1 burdens are plotted for comparison ( Supplementary Figure S1 ) . With the exception of patient 30015 , the HIV-1 brain burdens in untreated patients correlated with presentation of HIVE; on average , these patients had about 180-fold more viral RNA per µg total RNA than patients without HIVE . Patient 30015 had limited HIVE pathology and no detectable virus in the brain ( Table 1 ) . Patients on ART , both with and without HIVE , had lower virus burdens in the brain than untreated patients , results consistent with a previous study in a different cohort [48] . In patients with HIVE the reduction was 50% and 95% at the DNA and RNA levels respectively , while virus was undetectable in treated patients without HIVE . Notably , there was no correlation between brain and plasma viral loads and brain virus burdens were independent of patient's age , gender , ethnic background , or postmortem interval . Global gene expression profiles of patient brain tissues were determined on Affymetrix GeneChip Array Human Genome U133 Plus 2 . 0 Arrays . Three independent array experiments were performed , each comprising a subset of HIV-1-positive samples and controls , with most samples tested either in duplicates or repeated in independent runs and some samples tested three times . Replicate gene sets of the same samples were averaged , yielding 21 final brain tissue datasets for 21 subjects . Preliminary hierarchical cluster analysis of complete datasets from HAND patients indicated that the primary biological variable in clustering of these datasets was whether or not patients were on ART at the time of death ( not shown ) . Interestingly , this analysis also indicated that brain transcriptomes from untreated patients with HIVE did not cluster independently of non-HIVE transcriptomes ( Figure 1A ) , suggesting that they are statistically similar across their entire datasets . This result was surprising because untreated patients with HIVE had on average higher brain virus burdens than patients without HIVE ( Table 1 and Supplementary Figure S1 ) and HIV-1 infection is known to induce cellular gene expression [49] . To confirm findings of cluster analysis we performed global gene set analysis using GAzer software [50] to identify biological pathways that were most significantly altered in HIVE positive and negative groups compared to uninfected controls ( Figure 1B ) . GAzer employs parametric analysis of sets of co-regulated genes across complete microarray datasets , independent of arbitrary fold change ( FC ) value limits , thus increasing statistical power of detection of biological differences between datasets [50] , [51] . Figure 1B depicts the eight most altered pathways in HAND and HAND/HIVE datasets versus controls and Supplementary Table S1 lists all aberrant pathways and detailed statistics of GAzer analysis including Z-score , p-and q-values , and Bonferroni correction value . Consistent with previous array studies in HIVE patients [43] , [44] , datasets of patients with HIVE showed greater dysregulation of immune responses and endogenous antigen presentation pathways than those without HIVE ( Figure 1B ) , possibly reflecting effects of high virus burdens in the brain in untreated HIVE ( Table 1 ) . However , the two HAND patient groups were generally similar with respect to the ranking and extent of change ( indicated by all four statistical measures ) of the majority of altered gene sets ( Figure 1B and Supplementary Table S1 ) . Because of their overall similarities in both cluster and gene set analyses , we chose to pool microarray datasets from HAND patients with and without encephalitis for analysis of antiviral drug effects . Seven of the 15 patients in our cohort were treated using different antiretroviral drug combinations ( Table 1 ) enabling investigation of the extent of HIV-1-induced changes in the brain and potential differences between treated and untreated patients at the level of brain transcriptomes . Patient microarray datasets were pooled separately into untreated ( HAND , n = 8 ) and treated ( ART , n = 7 ) groups and each group was compared to the uninfected control pool C ( n = 6 ) . We used a cut-off of 1 . 5 FC and p-value<0 . 05 to identify significantly dysregulated genes . For some analyses , we also established a subgroup ARTa excluding low adherence subjects ART3 and ART5 . The complete lists of FC values and accompanying statistics for HAND , ART , and ARTa are provided in Supplementary Tables S2 and S3 . Overall , we identified 2073 dysregulated transcripts in the untreated HAND group and 333 and 145 transcripts in the ART and ARTa groups , respectively ( Table S2 ) . The Venn diagram in Figure 2A depicts the number of genes dysregulated in brain tissues of untreated and treated patients with HAND and the overlap in the dysregulated genes among the three groups tested . Excluding multiple probes for the same gene and transcripts of undefined function at the time of this writing , the HAND group had 1470 genes with significantly altered expression compared to 260 in ART and 107 in ARTa ( Figure 2A , Supplementary Tables S2 and S3 ) . About two-thirds of dysregulated genes ( 947 ) in untreated patients were up-regulated and 95 of these had FC values of ≥3 . 0 and p = 10−2–10−7 , among down-regulated genes , 58 had FC of ≤−3 . 0 and p of 10−2–10−4 , suggesting that HIV-1 infection profoundly alters the brain transcriptome in untreated patients with HAND and indicating significant conformity of molecular profiles of disease in this cohort . In treated patients , down-modulated genes predominated , the FC values ranged from 3 . 95 to 1 . 5 for up-regulated genes and from −1 . 5 to −2 . 87 for down-regulated genes , and p-values were 5×10−2–7 . 3×10−5 ( Supplementary Tables S2 and S3 ) , indicating less uniformity of brain gene expression in this group . These results indicate marked differences between aberrant brain transcriptome profiles of untreated and treated patients , the latter showing 6–14-fold ( depending on treatment compliance ) fewer dysregulated genes with generally lesser dysregulation of expression than in untreated patients . To confirm this observation we conducted gene ontology analysis using GAzer to examine cellular processes affected by HIV-1 infection in the brain in our patient groups . Altered gene sets ( biological pathways ) were identified by comparing each patient group to HIV-1-negative controls; Supplementary Table S4 lists these pathways for HAND , ART , and ARTa in the order of their significance as determined by the Z-score , p and q values , and Bonferroni statistics [50] , [51] . Figure 2B shows the eight most dysregulated biological pathways in the HAND datasets , displaying the extent of dysregulation in the same pathways in ART and ARTa datasets . Up-regulated pathways in untreated HAND patients included immune responses , inflammation , response to virus , and complement activation while synaptic transmission , neurogenesis , ion transport , cell adhesion , and signal transduction were down-regulated . The statistical significance for the eight major pathway changes reached Z-scores of 6–14 and p , q and Bonferroni values of ≥10−8 ( Figure 2B and Supplementary Table S4 ) . In contrast , samples from treated patients showed either fewer significantly altered pathways ( ART and ARTa panels in Figure 2B ) or lesser extent of dysregulation of the remaining pathways displayed ( e . g . , the ART panel in Figure 2B ) . The extent and kind of changes in gene ontology processes in the treatment compliant ARTa group were limited compared to changes seen in untreated patients: of the 8 most up-regulated HAND pathways only endogenous antigen presentation and processing were also up-regulated in ARTa with Z-scores less than 2 . 5 , low significance compared to other pathways ( p = 0 . 018 and 0 . 026 , respectively ) , and no down-regulated gene sets in these categories were identified ( Figure 2B and Supplementary Table S4 ) . Notably , some pathways that were significantly down-regulated in HAND were significantly up-regulated in ARTa including neurotransmitter secretion ( Z = 4 . 26; p = 2×10−5 ) and synaptic transmission ( Z = 2 . 9; p = 0 . 0037 ) ( Supplementary Table S4 ) . Dysregulation of selected genes detected by microarrays was confirmed in adjacent tissues by QPCR . Genes were chosen by previous demonstration of their link to HAND [52]–[56] . Representative gene expression values are shown in Figure 2C and the complete list is provided in Supplementary Table S5 . Consistent with microarray data , brain tissues from untreated patients showed significant up-regulation of complement component 3 ( C3 ) , macrophage antigen CD68 ( CD68 ) , and protein-tyrosine phosphatase receptor-type C PTPRC ( also known as CD45R ) ; and down-regulation of neuronal cyclin-dependent kinase 5 , regulatory subunit 2 ( CDK5R2 ) ( only significant for p39 ) , neuronal marker microtubule-associate protein 2 ( MAP2 ) and the SNARE protein complexin 1 ( CPLX1 ) , compared to brains of patients on ART . As also noted in other studies [57] , confirmatory QPCR analyses generally yielded higher FC results than parallel microarray analyses ( Supplementary Table S5 ) . Four of the changes in gene expression in the brain were also tested at the protein level by immunohistochemistry ( Figure 2D ) . We observed increased expression of CD68 , C3c and CD45R proteins in untreated HAND compared to uninfected control tissue , and an amelioration of the protein dysregulation in patients with HAND on ART . MAP2 protein was down-regulated in HAND brains compared with brains from uninfected subjects , and it was partially restored in patients under treatment for HIV-1 infection . Taken together , our findings parallel recent reports of the effects of ART on gene expression in peripheral tissues [24] , [27] , [29] in that treatment of HIV-1 infection by ART is also accompanied by marked reduction of the virally-induced dysregulation of gene expression in brain tissues . Treatment adherence further reduces the number of genes and biological pathways affected . In the next level of analysis , we conducted unsupervised hierarchical clustering for 2073 dysregulated transcripts implicated in HAND ( Figure 2A and Supplementary Table S2 ) to visualize and group individual gene expression profiles from all 21 subjects in this study ( Figure 3A ) . We used normalized Robust Microarray Average ( RMA ) values from duplicate samples for each subject for greater statistical power ( Methods ) . Three main clusters of subjects were identified . With the exception of HAND5 , untreated HAND patients clustered as one group ( cluster 2 in Figure 3A ) distinct from the other two clusters . ART-treated HAND patients clustered with uninfected controls in two co-mingled clusters , cluster 1 containing subjects C1 , C2 , C3 and ART1 and cluster 3 including subjects C4 , C5 , C6 , ART2 , ART3 , ART4 , ART5 , ART6 , and the remaining untreated HAND5 . Compared to cluster 3 , cluster 1 was more phylogenetically distant from untreated HAND , but overall clusters 1 and 3 , separately or combined , were significantly different from cluster 2 . The t-test values computed using RMA values for all transcripts in cluster 1 vs . cluster 2 , cluster 3 vs . cluster 2 , and cluster 1+cluster 3 vs . cluster 2 were 1×10−221 , 7 . 8×10−25 , and 1 . 96×10−24 , respectively . These results demonstrate that for the 2073 transcripts tested , gene expression in brain tissues of patients on ART tends to resemble that of uninfected subjects , indicating that ART is associated with a profound reduction in the extent of dysregulation of HAND-related genes in the brain . These results also suggest that the differences related to HIV-1 disease ( and treatment ) override potential differences resulting from genetic heterogeneity of individual donors . The statistical similarity of ART transcriptomes with HIV-1-negative controls with respect to presumptive HIV-1-impacted genes ( Figure 3A ) prompted us to directly compare gene expression changes of ART brain tissue to expression in HAND brain tissue , without filtering microarray results through HIV-1-negative controls . This analysis inquires whether control of HIV-1 replication by ART removed viral perturbations to cellular gene expression in the brain , analogous to longitudinal studies of peripheral tissues from HIV-1-infected subjects pre and post-ART [24] which are not feasible for the brain . Using normalized RMA datasets for over 54 , 000 transcripts detected by the U133 chipset , we compared 7 treated patients to 8 untreated patients delimiting the results by FC of 1 . 5 and t-test value of 0 . 05; the complete list these differentially expressed transcripts is shown in Supplementary Table S6 . In calculating the ratio of gene expression in ART to HAND , positive FC values in ART indicate an association of increased gene expression with treatment and negative FC indicate reduced cellular gene expression associated with treatment . Overall , we identified 640 significantly up-regulated and 276 down-regulated genes in this analysis ( Supplementary Table S6 ) . Samples from patients with treated HAND show increased expression ( relative to untreated HAND ) of many genes involved in neuronal functions including synaptoporin ( SYNPR , FC 6 . 55 and p = 6 . 5×10−4 ) , neurofilament , light polypeptide ( NFEL , FC 6 . 15 , p = 4 . 6×10−4 ) , and synaptotagmin IV ( SYT4 , FC 4 . 16 , p = 5 . 8×10−4 ) compared to samples from untreated patients . Conversely , genes involved in immune activation including CD74 antigen ( CD74 , FC −2 . 49 , p-0 . 013 ) , complement component 1 , q subcomponent , C chain ( C1QC , FC −2 . 47 , p = 0 . 022 ) , and interferon-induced protein with tetratricopeptide repeats 2 ( IFIT2 , FC −2 . 35 , p = 0 . 007 ) were reduced in expression in treated HAND patient samples relative to samples from untreated patients . For reference , Supplementary Table S6 also lists respective ART microarray data normalized to HIV-1-negative controls . Notably , all but 7 of the genes in treated patients that increased in expression relative to untreated HAND were unchanged in the ART/HIV-negative control comparison ( Supplementary Table S6 ) , suggesting that treated patients express these genes at normal ( control ) levels . Similar normalization of expression ( 272 out of 276 ) was found for genes that were down-modulated in ART versus HAND . To put these findings in the context of the biological pathways potentially affected by ART , we employed GAzer to compare the complete ART microarray datasets to those from HAND . Figure 3B depicts 16 biological processes that were most significantly changed in ART relative to HAND and Supplementary Table S7 provides complete statistics for this analysis . Twelve up-regulated processes in treated versus untreated HAND , with Z-scores ranging from 9 . 54 to 5 . 02 , were related to neuronal function and repair . The down-modulated pathways included , in decreasing order of significance , immune responses , inflammatory responses , apoptosis , and responses to stress . These results suggest that ART reverses many dysfunctional processes of untreated HAND represented in gene ontology analysis . To provide an alternative view of the extensive up-regulation of cellular processes in ART relative to HAND , up-regulated genes with p-values of ≤0 . 01 included in Supplementary Table S6 were analyzed by STRING to identify predicted gene interaction networks ( Figure 4 ) . In this analysis ART was associated with improved synaptic transmission including synaptic vesicle system , nervous system development including cytoskeleton associated proteins , and GABA neurotransmission networks . These findings strongly suggest that by reducing HIV-1 replication , ART also reduces triggers to aberrant gene expression in the brain . The HAND patients in this study share cognitive dysfunction whether they were untreated or treated with ART ( Table 1 ) . To begin to identify transcripts that may contribute to HAND development or persistence despite ART , we used t-test to compare significantly changed genes in untreated ( Supplementary Table S2 ) and treated ( Supplementary Table S3 ) patients with HAND; a gene whose expression was not significantly different in the HAND versus ART comparison ( p>0 . 05 ) was considered similarly dysregulated in both groups of patients relative to uninfected subjects . We have identified 43 such up-regulated and 42 down-regulated genes; they are listed grouped into biological categories in Supplemental Table S8 , selected genes with their expression statistics are listed in Figure 5A , and their heatmap expression profiles in all subjects in this study are shown in Figure 5B . Considering functional characterization , genes related to immune responses were up-regulated in both HAND and ART samples including complement receptor 1 ( CR1 ) , chemokine , CXC motif , ligand 2 ( CXCL2 ) , major histocompatibility complex , class II HLA-DQB1 and interferon-mediated antiviral responses including interferon-induced protein with tetratricopeptide repeats 1 ( IFIT1 ) ; interferon-induced protein 44 ( IFI44 ) ; myxovirus resistance 1 ( MX1 ) ; 2′ , 5′-oligoadenylate synthetase 1 ( OAS1 ) , and signal transducer and activator of transcription 1 ( STAT1 ) . Cell cycle pathway was dysregulated in both treated and untreated HAND patients , with some sets of genes up-regulated and others down-regulated ( Figure 5A and 5B ) . Over-expression of selected transcripts in interferon or chemokine pathways in both HAND and ART brain samples was confirmed by real-time PCR ( Figure 5C ) , over-expression of HLA Class II alleles and proliferating cell nuclear antigen ( PCNA ) was also demonstrated by immunohistochemistry ( Figure 5D ) . Of particular interest , common down-regulated genes in treated and untreated patients with HAND included myelin-related genes myelin-associated oligodendrocyte basic protein ( MOBP ) , myelin transcription factor 1 ( MYT1 ) and myelin basic protein ( MBP ) . Down-regulation of MOBP and MYT1 was confirmed by real-time PCR , with the MYT1 gene being particularly suppressed in the ART group ( FC = −38 . 38 , p = 1×10−5 ) ( Supplementary Table S5 ) . This result is consistent with histopathological detection of myelin pallor in autopsy brain tissues from some patients with HAND [58] , [59] , although other explanations also exist [60] . Pearson's formula was applied to determine the correlation of the level of expression of each transcript with HIV-1 load in plasma , cerebrospinal fluid ( CSF ) and brain . Examples of genes correlating positively with plasma viral load are shown in Figure 6A , including histocompatibility loci HLA-B-G and F , interferon-gamma-inducible protein 30 ( IFI30 ) , OAS1 , and Cathepsin S ( CTSS ) . Transcripts with a positive ( >0 . 5 ) or negative ( <−0 . 5 ) correlation with viral load were analyzed using the gene ontology software Expression Analysis Significance Explorer ( EASE ) to identify the pathways that correlated most with viral load ( Figure 6B ) . Pathways positively correlated with viral load were similar in the three compartments ( brain , CSF , plasma ) , including several immune activation responses . The main difference among the three compartments was the level of significance of the changes , as represented by the EASE score . All the pathways implicated in immune response correlated better with plasma viral load than with brain viral load . Even though CSF data were available for only 9 of the 15 infected patients , the positive correlations for CSF were higher than those for brain , although less strong than those observed for plasma . The categories of antigen presentation and processing were correlated only with viral load in plasma . Pathways down-regulated in correlation with viral load differed depending on the compartment tested . No biological pathway correlated negatively with brain viral load . Pathways negatively correlated with plasma viral load included synaptic transmission , cell communication , transmission of nerve impulse , organogenesis and neurogenesis . A wide variety of metabolic pathways negatively correlated with CSF viral load . Overall , grouping genes engaged in similar biological functions indicates that gene groups induced in the brain correlated best with virus burden in the periphery and that virus burden in the brain , unlike viral load in plasma or CSF , was uncorrelated to suppression of expression of any gene group .
The foundation of this work is a new comprehensive database of global gene expression profiles in brain tissues of patients with HAND . The profiles described here complement published datasets from previous array studies in HAND [41]–[44] with important differences . For the first time in this disease , we analyzed brain transcriptomes on the basis of the antiretroviral treatment of patients , revealing two distinct , largely non-overlapping groups of aggregate gene expression profiles termed ART and untreated HAND . The ART and untreated HAND profiles also formed separate clusters in unsupervised hierarchical cluster analysis [61] of individual patient datasets , confirming that they are phylogenetically distant from each other based on a large sets of aberrantly expressed genes used in this analysis ( Figure 3A ) . Importantly , the hierarchical clustering distinction between treated and untreated patients was statistically more prominent than potential distinctions in gene expression related to other patient characteristics in our cohort including HIVE and intravenous drug use ( Figure 3A and Supplementary Figure S2 ) . Therefore , in most analyses in this work we considered datasets from patients with and without HIVE as one disease category , stratified only on the basis of ART . These results suggest that ART , through effects on HIV-1-associated changes in cellular gene expression [27] , [62] , [63] , is one of the key biological variables governing the extent and pattern of gene dysregulation in molecular profiles of HIV-1 brain disease . Another important difference with previous HAND array studies concerns the histological regions of the brain tested . We evaluated frontal deep white matter , whereas most of the previous studies focused on the neighboring cortical gray matter [41]–[44] or small gene sets in both brain regions [41] . White matter is the primary site of HIV-1 infection and HIV-1-associated neuropathologies [46] , [47] and frontal cortex is one of the sites of synaptic and dendritic damage consequent to this infection [64] , [65] . Transcriptomes from these two areas reflect different regional physiologies and therefore different aspects of HIV-1 neuropathogenesis . With these caveats , a meta-analysis summarized in Supplementary Table S9 identified a small number of genes involved in interferon-related responses ( IFIT1 , IFITM1 , IFI44 , MX1 ) , synaptic functions ( SYN1 , SYN2 , GABRG2 , MAP2 ) , and cell cycle ( CDC42 , CDK5R1 , R2 ) that were dysregulated in common in the present and previously published HAND datasets . These genes and biological pathways may represent features of HIV-1-associated neuropathogenesis common to white and gray matter . During the last twenty years , individual inflammatory and neurodegenerative mediators in HAND brains were demonstrated by immunocytochemistry , in situ hybridization , and other methods ( reviewed in [66]–[69] ) . The bulk of this work was conducted with brain tissues from untreated patients , as brain autopsies after introduction of ART have become less frequent [70] . Consistent with these observations , the untreated HAND profile defined here reveals a broad and extensive dysregulation of cellular gene expression in brain tissues , with 1470 HAND-associated aberrantly expressed genes , up-regulation of immune activation , antiviral responses , and inflammation , and down-modulation of neuronal functions , neuronal repair , and cell cycle . It should be noted that our untreated HAND profile included transcriptomes from patients with and without HIVE ( Figure 2 and 3 ) . HIVE is a characteristic histopathology associated with high HIV-1 burdens in the brain [71] , [72] , and previous array studies documented differentially expressed transcripts and biological pathways potentially attributable to the extensive infection in the brain [43]–[45] , [73] . We confirmed some of these differences in our untreated group with and without HIVE in gene ontology analysis but it is noteworthy that they differed mainly in the degree of dysregulation and not the biological pathways affected ( Figure 1 ) . Thus , at the level of a transcriptome analysis , the altered expression of some transcripts attributed to high HIV-1 burdens in HIVE [43] , [44] contributes to but does not change the overall statistical characteristics of the molecular phenotype of HAND . Our results are consistent with reports of limited correlation between clinical manifestations of HAND and HIVE histopathology [58] , [74] , [75] . Rather , untreated HAND appears to correlate better with presence of inflammatory mediators and diffusely activated macrophages and microglial cells in the brain than with virus burdens in the tissue per se [17] , [58] , [75]–[78] . Independent microarray studies in other patient populations are needed to determine whether these gene expression profiles are fully representative of untreated HAND . In general , transcriptome profiles of disease can serve as a platform for verification of results obtained in studies of individual physiological processes [33] , [79] , [80] and as a tool for discovery of new ones . In this context , a number of “novel” ( i . e . , relatively new to the HAND literature ) dysregulated genes in our untreated HAND database may shed light on the process of HAND pathogenesis and merit further investigation . For example , transcripts encoding apolipoprotein C-I and C-II ( APOC-I and APOC-II ) were among the most up-regulated in the untreated HAND dataset , with FC of 5 . 71 ( p = 9 . 7×10−5 ) and 3 . 35 ( p = 0 . 009 ) , respectively ( Supplementary Table S2 ) . Dysregulation of lipid metabolism linked to APOE polymorphism is a marker of AD [81] and was indicated in HIV-1 dementia [82] , [83] . While we could not find reports on APOC-I in HAND , this lipoprotein was found in association with beta-amyloid plaques in AD brains and expression of human APOC-I allele in native APOC-I null mice was shown to impair learning and memory [84] . Conversely , hemoglobin α-2 and hemoglobin β ( HBA-2 and HBB ) were among the most down-regulated transcripts in our untreated HAND dataset ( FC of −5 . 31 and −4 . 22; Supplementary Tables S2 and S5 ) . Neuronal hemoglobins are members of the globin superfamily which are predominantly expressed in neurons and may play an important role in neuroprotection [85] . Consistent with our findings , expression of neuronal hemoglobin is reduced or absent in disease affected brain regions in patients with several neurodegenerative conditions including AD and Parkinson's disease [86] . The major finding of this work is the profound difference between the extensive global gene dysregulation observed in brain tissues of untreated patients with HAND and muted gene changes in their treated counterparts . The magnitude of ART effects in the brain suggested by our results was surprising given the limited clinical outcomes of ART on HAND [7] , [87] , including in the cohort evaluated here ( Table 1 ) , and variable findings in the CSF of treated patients [12]–[16] . The ART effects on the brain were inferred because we could not test brain RNA in the same individuals before and after initiation of therapy , as is possible in microarray studies with peripheral tissues [24] , [27] , [29] , [31] . However , the overall effects of ART on altered gene expression were remarkably similar in the periphery and brain . This was evident in markedly fewer dysregulated genes in treated compared to untreated patients , in our case 253 versus 1470; and in a global shift in gene expression patterns from aberrant in the absence of treatment to muted dysregulation under treatment , for example in peripheral CD4+ T cells [27] , [29] , lymphoid tissue [26] , and brain here ( Figure 3A ) . Importantly , both in the CD4+ T lymphocyte study [27] and in the present work , gene expression profiles of treated patients were statistically similar to those of HIV-1-negative controls , suggesting a trend toward normalization of gene expression under ART . We confirmed this trend for selected gene products in the present work by real-time PCR and immunocytochemistry ( Figure 2 ) . These results suggest that ART regimens , which generally include at least one brain penetrant antiviral compound ( Table 1 ) , are similarly effective in mitigating global molecular changes in the brain and in peripheral tissues . We noted two reciprocal effects of ART on gene dysregulation in brain tissues illuminating the systemic and brain-specific aspects of HAND pathogenesis . One is a significant and broad moderation of up-regulated genes linked to HIV-1 induced antiviral and inflammatory responses thought to drive HAND pathogenesis , including interferon-related ISG15 and IFIT3 , macrophage markers CD68 , CD163 , and CD14 , and chemokines and chemokine receptors CCL8 , CCR1 , and CXCR4 [67]–[69] ( Figure 2 and 3 ) . Interestingly , this effect of ART was common to diverse tissues examined by microarrays including brain ( this study ) and CD4+ T cells , macrophages , lymph nodes , and intestinal mucosa tested by others [24] , [27] , [29] , [31] , and thus it likely represents a system-wide response to suppression of HIV-1 replication . Although gene ontology pathways containing these genes in patients under ART were still up-regulated compared to controls ( Figure 2B and Figure 5 ) , our results suggest that ART can alleviate a surprisingly large number of deleterious responses in the brain that have been linked previously to HIV-1 infection in model systems [88] . This causal link is further strengthened by an apparent correlation in this work between treatment compliance and extent of gene dysregulation in the brain ( Figure 2 ) . The other effect of ART in the brain we observed was specific to the nervous system and it involved normalization of a large number of down-regulated genes and biological pathways linked to nervous system functions and by extension to neurocognitive disease . For example , the bioinformatics tool STRING [89] identified nervous system development , synaptic transmission , and GABA-neurotransmission pathway as the three major predicted interaction networks of genes that approached normal expression in treated patients ( Figure 4 ) . On a smaller scale , our confirmatory tests showed that products such as MAP2 and complexin-1 ( CPLX1 ) were significantly down-regulated in untreated patients at RNA and protein levels and they were expressed at control-like levels in tissues from treated patients ( Figure 2 ) . It is conceivable that restoration of normal expression of at least some of these genes under ART would restore some aspects of normal brain physiology [7] . Although treated patients in our cohort still manifested HAND prior to death , our results suggest that they may have already shifted to a milder molecular profile in the brain that had more in common with HIV-1-negative controls than untreated patients . Of interest , the global changes in brain cell gene expression seen by microarrays correlated positively with plasma but not brain virus burdens ( Figure 6 ) . This association may be analogous to the clustering of brain transcriptomes independently of encephalitis and brain HIV-1 burden by commonly dysregulated biological pathways . In any case , the single measurement of brain virus burden at autopsy may not capture the chronic insult to brain function suffered by patients living years with HIV-1 infection , albeit with some control exerted by ART . Perhaps the most intriguing contribution of the present array analysis lies in the ability to discern patterns of abnormality that persist in cognitively-impaired patients who are on central nervous system penetrant ART , and to distinguish these patterns from HAND in the untreated state . In the bioinformatics sense , we used ART as a biological filter to reduce the overall gene expression disturbance in brain transcriptomes of patients with HAND and through that determine whether continuing dysregulation of gene expression could play a role in continuing brain disease . The results indicate that continuing up-regulation of innate and adaptive immune responses are an important part of brain abnormalities in ART-treated dementia . Over-expression of Class II MHC in the brain persists despite therapy in our study; it has been associated with many neurodegenerative diseases [90] . Defects in myelin metabolism , shown for HAND at the gene expression level here and indicated previously by neuropathological observation of myelin pallor in brains of patients with HAD [58] , are also common to many other neurodegenerative diseases [38] , [91] . Activation of interferon-related genes often found in symptomatic HIV-1 infection may underlie abnormalities in cell function [28] . Such changes , coupled with cell cycle perturbations , may support emerging magnetic resonance spectroscopy studies that have demonstrated persistent white matter inflammation in patients with HIV-1-related cognitive impairment [92] . However , it is unclear what particular aspects of immune activation or response are relevant to nervous system dysfunction , and how to distinguish deleterious gene products from those that may function in a neuroprotective manner . In simian immunodeficiency virus infection , innate immunity , IL-6 , and interferon responses are important elements in brain viral control [56] , [93] , but in a recent study expression of interferon-α in the brain was conclusively linked to neuronal dysfunction in a mouse model of HIVE [94] . Careful analysis of individual genes identified here may begin to clarify the mechanism of HAND persistence under treatment .
Human brain samples and clinical data were obtained from the Manhattan HIV Brain Bank ( MHBB ) , a member of the National NeuroAIDS Tissue Consortium , under an Institutional Review Board-approved protocol at the Mount Sinai School of Medicine . Written informed consent was obtained from all subjects in this study or their primary next-of-kin . HIV-1 and gene expression analyses were conducted on de-identified brain samples under an “exempt” status approved by an Institutional Review Board of St . Luke's-Roosevelt Hospital Center . Study subject information is listed in Table 1 . Fifteen HIV-1-positive subjects used in this study were chosen on the basis of having HAND and the presence or absence of HIVE as determined by neuromedical or neuropsychological evaluation and postmortem neuropathology [95] , and then were further categorized as either dying on or off ART . Fourteen of HIV-1-positive patients in this study died with HAD; one patient designated ART2 in Table 1 was classified as MND based on his lack of emotive/behavioral criteria [96] . Patients who displayed HAND without HIVE histopathology at autopsy met American Academy of Neurology criteria for HAND regardless of ART status . This definition requires demonstration of cognitive and functional impairments , and the presence of emotive or motoric phenomena [96] . Except for patient ART2 , patients with HIVE were similarly impaired , regardless of ART status , or had histories of HAND on medical record review . Patients felt to have cognitive impairments ) due to non-HIV-1 causes ( neuropsychological impairment – other , as described in [95] were excluded . All patients dying on ART had substantive treatment histories , ranging 1 to 7 years prior to demise . The ART regimens at death had a mean duration of 16 months ( range , 3 months to 3 years ) . For five of seven patients on ART , treatment compliance estimates were made by self-reported 4 day recall [97] . HIV-1-negative subjects were chosen on the basis of normal neurological function ( as determined by chart review ) and normal neurohistology . All neurohistologic diagnoses were rendered by a board-certified neuropathologist ( SM ) and a minimum of 50 sections were examined for each brain . The following brain pathology definitions were used in the present work ( Table 1 ) : Normal: no brain pathology; HIVE: HIV encephalitis; HIVE*: HIVE was limited to the basal ganglia ( pid 30015 ) ; Minimal: minimal histopathological changes including trivial microscopic abnormalities such as an isolated vermal scar ( pid 10119 ) , a venous ectasia ( pid 10063 ) , atherosclerosis and minimal perivascular inflammation sub-threshold for diagnosis ( pid 10001 ) , and minimal perivascular inflammation sub-threshold for diagnosis ( pid 10015 ) . At the time of autopsy , coronal sections of brain were snap-frozen and maintained in −85°C until sub-dissection . Effort was made to keep the post mortem interval ( PMI ) to a minimum; the PMI for subjects in this study are listed in Table 1 . Brain samples for this analysis were obtained from the centrum semiovale ( deep white matter ) at the coronal level of the genu of the corpus callosum . Multiple samples from the same region were dissected for gene expression profiling , real-time PCR ( QPCR ) , and protein assays . Equivalent regions from the contralateral hemisphere were formalin fixed and utilized for immunohistochemistry . Total DNA and RNA were isolated from human brain tissue by , respectively , DNeasy Blood and Tissue Kit and RNeasy Mini Kit ( Qiagen , Valencia , CA ) according to the manufacturer's protocol . RNA was quantified by spectrophotometry and RNA quality was verified by spectrophotometry and agarose gel electrophoresis . RNA was then treated with DNAse I ( Fisher Healthcare , Houston , TX ) . cDNA was synthesized using the Superscript First-Strand Kit ( Invitrogen , Carlsbad , CA ) for quantitative analysis and the WT-Ovation™ RNA Amplification System ( NuGEN Technologies , Inc . , San Carlos , CA ) for relative analysis according to the manufacturer's protocol . Viral RNA and DNA burdens in human brain were determined by quantitative real-time PCR using primers designed based on HIV-1 consensus sequences for the group of patients in this study . We first screened patient brain samples for the presence of HIV-1 content by a standard nested PCR for viral DNA or RNA ( cDNA ) as previously described [98] . To achieve broad detection of diverse Clade B HIV-1 species , first-round PCR was performed using custom designed primers for a conserved Clade B gag consensus sequences: SQ5 ( + ) 5′-CAA ATG GTA CAT CAG GCC ATA TCA CC-3′ and SQ3′ ( − ) 5′-CCC TGA CAT GCT GTC ATC ATT TCT TC-3′ . For nested PCR step we used primers SQ5′ ( above ) and SK39 [99]; the PCR products were resolved on an agarose gel and detected by Southern Blot hybridization with ( 32P ) -labeled probe SK19 [99] ( Supplementary Figure S1 ) . The HIV-1 gag nested PCR amplicons from individual HIV-1 DNA positive patients were sequenced and used to design patient consensus primers ( 5′ ( + ) QSQ5 5′-ACC CAT GTT T ( T/A ) C AGC ATT ATC AGA-3′ and 3′ ( − ) QSQ3 5′-GAT GTC CCC CCA CTG TGT TT-3′ ) and Taqman probe ( HSQP 6FAM-AGC CAC CCC ACA AGA-MGBNFQ ) . For real-time PCR amplification , 2 µl of brain tissue DNA or 5 µl of cDNA were combined with 2× Universal Master Mix ( Applied Biosystems , Carlsbad , CA ) , 900 nM consensus primer ( custom synthesized by Invitrogen ) and 200 nM probe ( synthesized by Applied Biosystems; QPCR conditions were essentially as described [100] . Standard curve for viral DNA and cDNA quantification was constructed from graded amounts of HIV-1 NL4-3 plasmid DNA . Viral DNA burdens were normalized to total cellular DNA content by β-globin amplification and expressed as HIV-1 DNA copies/number of cells calculated from a β-globin DNA standard curve , with 2 copies of β-globin gene equaling one cell . Viral RNA burdens were normalized by tissue glyceraldehydes-3-phosphate dehydrogenase ( GAPDH ) content and expressed as number of viral copies in 1 µg tissue RNA [100] . Microarray experiments were conducted at the Bionomics Research and Technology Center in EOSHI University of Medicine and Dentistry of New Jersey . Total RNA were extracted from tissue samples using the RNeasy Mini Kit ( Qiagen ) followed by DNase I treatment . RNA qualities were assessed by electrophoresis using the Agilent Bioanalyzer 2100 and spectrophotometric analysis prior to cDNA synthesis . Fifty nanograms of total RNA from each sample were used to generate a high fidelity cDNA for array hybridization using NuGen WT-Ovation Pico RNA Amplification . Detailed protocols for sample preparation can be found at http://www . nugeninc . com . After fragmentation and biotin labeling using NuGen Encore Biotin Module , the samples were hybridized to Affymetrix Human Genome 133 plus 2 . 0 arrays . Washing and staining of all arrays were carried out in the Affymetrix fluidics module as per the manufacturer's protocol . The detection and quantitation of target hybridization was performed with an Affymetrix GeneChip Scanner . Data were assessed for array performance prior to analysis . The majority of patient samples were analyzed in duplicates starting from the cDNA synthesis step; in some cases second analysis was on an adjoining brain sample . The . cel data files generated by the Affymetrix microarray hybridization platform were analyzed by the ArrayAssist software ( Stratagene , Santa Clara , CA ) . Probe level analysis was performed using the RMA algorithm . After verification of data quality by Affymetrix internal controls and signal distribution analysis as described in Stratagene ArrayAssist Protocol , data was transformed using variance stabilization and logarithm transformation with a base of 2 . Fluorescence values were normalized by mean intensities of all chip samples . Means of normalized expression values were calculated for duplicate samples from each individual , and these values were either used directly in some analytical programs ( see below ) or employed to calculate fold change ( FC ) in the transcript compared to HIV-1-negative controls . Genes showing FC values above 1 . 5 or below −1 . 5 and unpaired t-test p-values of <0 . 05 were defined as significantly changed . In some analyses we applied a t-test cutoff of p<0 . 01 . To test for effect of antiretroviral treatment , we calculated the FC for significantly modulated transcripts separately from untreated and treated patients , the latter subdivided into all-treated and a subset without two known low-compliant patients , versus uninfected controls . We also compared directly normalized expression values of selected transcripts from treated and untreated patients to generate the treated versus untreated FC values that were not filtered through HIV-1-negative controls . To compare brain samples of all the individuals included in the study we clustered the expression profiles using unsupervised Hierarchical Clustering ( Average Linkage Clustering ) and heatmap visualization software from Genesis [101] available at http://genome . tugraz . at/ . This type of analysis allows us to identify the similarities and differences in the expression patterns of groups of patients and/or transcripts . For broad characterization of gene expression changes in untreated and treated patients we conducted gene set and gene ontology analysis using GAzer [50] ( http://expressome . kobic . re . kr/GAzer/index . faces ) . GAzer is a web-based tool that identifies , by a parametric statistical analysis of complete primary normalized microarray data , over-represented sets of genes ( functionally related genes ) rather than individual genes . This type of analysis compensates for the fact that small changes not seen at the gene level are often detected when the gene set as a whole is examined [50] , [51] . When analysis was limited to smaller sets of genes defined as differentially expressed , we performed functional categorization of gene families by EASE [102] , available at the NIH web site ( http://david . abcc . ncifcrf . gov/ ) . The predicted biological pathway and network relationships among differentially expressed genes in our array datasets were identified using the Search Tool for the Retrieval if Interacting Genes/Proteins ( STRING ) ( http://string-db . org/ ) [89] . Changes in expression of selected genes identified by microarrays analysis were validated in the same or adjoining brain samples by QPCR using Taqman chemistry and probes from the Universal Probe Library ( Roche , Indianapolis , IN ) . Primers were designed using the online ProbeFinder software available at the Roche Universal Probe Library Assay Design Center ( http://www . roche-applied-science . com ) . The QPCR reactions contained 2 µl of cDNA generated from tissue RNA obtained as described above , 10 µl of 2× Universal Master Mix ( Applied Biosystems-ABI ) , 0 . 2 µl of each forward and reverse primers at 200 nM , 0 . 2 µl of probe at 100 nM , and RNAse/DNAase-free water . All reactions were performed in duplicate and were run in a 7500 real-time PCR system ( ABI ) . Raw data was analyzed using the 7500 System SDS Software ( ABI ) . Data was normalized using 2 housekeeping genes , GAPDH and ribosomal protein S18 ( RPS18 ) to assure reproducibility . Relative quantification employed the comparative threshold cycle method ( Applied Biosystems Technical Bulletin n°2 ) . For immunohistochemical analysis , formalin fixed blocks were taken from the frontal white matter of the autopsy brains of 4 normal , 6 HIV-1-non-treated , and 4 HIV-1-ART-treated individuals . A microarray block consisting of 3 tissue punches diameter of 1 mm ) from each block was constructed . 5 µM serial sections were cut and immunohistochemistry performed with an array of antibodies listed in the Table 2 . Formalin-fixed , paraffin-embedded sections were deparaffinized with xylenes , hydrated in graded alcohols , and incubated with 3% H2O2 in methanol . Following washing , sections were boiled in Target retrieval solution ( DAKO Corp . , Carpinteria , CA ) , subsequently incubated in a serum free protein blocking solution and then incubated with the primary antibodies indicated in Table 2 . CD45 ( PTPRC ) was incubated overnight at 4°C and the all the remaining antibodies for 1 h at room temperature . Primary antibodies were detected with peroxidase anti rabbit or mouse IgG ImmPRESS ( DAKO Corp . ) reagent and counterstained with hematoxylin . Slides were visualized in a light microscope and either one , three or six 0 . 03 mm2 areas of white matter staining from each case were photographed using a 40× objective and a Nikon Coolpix II digital camera attached to the microscope by a Coolpix MDC lens . The intensity of illumination and position of sub-stage condenser on the microscope were constant for all images . The number of areas taken depended on the type of quantitative analysis that followed . The percentage area occupied by cells immunoreactive for CD68 , CD45 and HLA DP , DQ , DR was quantitated by analysis of six 0 . 03 mm2 images using a proprietary automated morphometric analysis software [103] . The mean intensity of MAP2 and STAT1 staining was analyzed on one 0 . 03 mm2 image , utilizing Image J software ( http://rsbweb . nih . gov/ij/ ) . Each image was converted to 8-bit grayscale and a mean gray value of the pixels in the full photographic image was measured after a background subtraction and inversion were performed . For C3c and PCNA , cell counts were done by eye using either three 0 . 03 mm2 ( C3c ) or the full punch ( PCNA ) . Statistical analysis was performed using StatView ( V . 5 . 0 . 1 ) ( Adept Scientific , Bethesda , MD ) . The principal statistical test used was the Analysis of Variance for single comparisons ( ANOVA ) . Follow-up post-hoc tests were conducted when required . Significance values were set at 0 . 05 and below . Correlation analysis between virus loads in plasma , CSF , or brain and gene expression in the brain was performed using the Pearson's correlation formula in Microsoft Excel software ( Microsoft Corporation , Redmond , WA ) . Transcripts showing positive ( >0 . 5 ) or negative correlation ( <−0 . 5 ) were categorized into biological functional pathways using EASE software . The microarray results presented here are available in the Gene Expression Omnibus database ( www . ncbi . nlm . nih . gov/geo ) in a MIAME compliant format under accession number GSE28160 . The Gene ID numbers for the genes mentioned in the text are listed below . The ID number corresponds to the ‘National Center for Biotechnology Information database’ ( http://www . ncbi . nlm . nih . gov/gene ) : CD4 ( 920 ) , C3 ( 718 ) , CD68 ( 968 ) , PTPRC ( 5788 ) , CDK5R2 ( 8941 ) , MAP2 ( 4133 ) , CPLX1 ( 10815 ) , SYNPR ( 132204 ) , NFEL ( 4747 ) , SYT4 ( 6860 ) , CD74 ( 972 ) , C1QC ( 714 ) , IFIT2 ( 3433 ) , CR1 ( 1378 ) , CXCL2 ( 2920 ) , HLA-DQB1 ( 3119 ) , IFIT1 ( 3434 ) , IFI44 ( 10561 ) , MX1 ( 4599 ) , OAS1 ( 4938 ) , STAT1 ( 6772 ) , PCNA ( 5111 ) , MOBP ( 4336 ) , MYT1 ( 4661 ) , MBD ( 4155 ) , IFI30 ( 10437 ) , CTSS ( 1520 ) , IFITM1 ( 8519 ) , B2M ( 567 ) , CD14 ( 929 ) , APOC-I ( 341 ) , APOC-II ( 344 ) , HBA-2 ( 3040 ) , HBB ( 3043 ) , IFI16 ( 3428 ) , TLR7 ( 51284 ) , SYN1 ( 6853 ) , SYN2 ( 6854 ) , GABRG2 ( 2566 ) , CDC42 ( 998 ) , CDK5R1 ( 8851 ) , GAPDH ( 2597 ) , RPS18 ( 6222 ) . | HAND is a common complication of HIV-1 infection in the nervous system presenting a varied spectrum of clinical manifestations with cognitive , motor and behavioral symptoms . Introduction of ART has greatly reduced morbidity and mortality in HIV-1 infection; however HAND persists and its overall prevalence appears to have increased despite treatment . The effects of the treatment on neurological disease are not well understood . Here , we used genomic analysis to compare gene expression profiles in brain tissues from treated and untreated patients who died with HAND . We identified a large number of genes and biological pathways dysregulated in untreated HAND compared with uninfected controls . ART appears to be effective in mitigating aberrant gene expression in brain tissues of patients with HAND but a fraction of genes remained dysregulated under ART and the patients continued to manifest HAND in the last evaluation prior to death . Our study provides new insights into the molecular changes in brain tissues of patients with HAND and the effect of the treatment on brain transcriptome . The identification of aberrantly expressed genes common to untreated and treated HAND may contribute to understand the neurocognitive impairment observed in patients under ART . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"medicine",
"neurobiology",
"of",
"disease",
"and",
"regeneration",
"functional",
"genomics",
"retrovirology",
"and",
"hiv",
"immunopathogenesis",
"microbiology",
"neuroscience",
"infectious",
"diseases",
"hiv",
"biology",
"dementia",
"virology",
"neurology",
"neurological",
"disorders",
"viral",
"diseases",
"genomics",
"computational",
"biology"
]
| 2011 | Significant Effects of Antiretroviral Therapy on Global Gene Expression in Brain Tissues of Patients with HIV-1-Associated Neurocognitive Disorders |
We identified a non-synonymous mutation in Oas2 ( I405N ) , a sensor of viral double-stranded RNA , from an ENU-mutagenesis screen designed to discover new genes involved in mammary development . The mutation caused post-partum failure of lactation in healthy mice with otherwise normally developed mammary glands , characterized by greatly reduced milk protein synthesis coupled with epithelial cell death , inhibition of proliferation and a robust interferon response . Expression of mutant but not wild type Oas2 in cultured HC-11 or T47D mammary cells recapitulated the phenotypic and transcriptional effects observed in the mouse . The mutation activates the OAS2 pathway , demonstrated by a 34-fold increase in RNase L activity , and its effects were dependent on expression of RNase L and IRF7 , proximal and distal pathway members . This is the first report of a viral recognition pathway regulating lactation .
The oligoadenylate synthetase ( OAS ) enzymes are activated by detection of double stranded RNA produced during the viral life cycle , and in response polymerize ATP into 2´-5´ linked oligoadenylates ( 2-5A ) of various lengths . The 2-5As then bind and activate latent RNase L , which degrades viral and host single stranded RNAs , so disrupting the viral life cycle [1] . It has been reported that OAS1 has antiviral activity independently of RNAse L [2] , that OAS2 binds to NOD2 [3] , and that OASL binds RIG-I [4] , pointing to additional mechanisms of action . Although mechanistic detail is lacking , it is proposed that OAS enzymes can activate anti viral responses via mechanisms independently of 2-5A production , by direct interactions within the viral signaling complex . For example , this complex is tethered to the mitochondrial outer membrane by the scaffold protein MAVS , and contains RIG-I , related helicase MDA5 , and possibly OAS family members [5] . OAS family members may also mediate apoptosis outside the context of viral infection [6 , 7] . Here we report a mutation of OAS2 that produces lactation failure in an otherwise normal mouse . This is the first demonstration that a viral recognition pathway can regulate lactation .
Using N-ethyl-N-nitrosourea ( ENU ) mutagenesis and a screen for failed lactation we established a mouse line in which heterozygous ( wt/mt ) dams showed partial penetrance of poor lactation , producing litters that failed to thrive , while homozygous ( mt/mt ) dams experienced complete failure of lactation ( Fig 1A ) , providing a dominant pattern of inheritance . Development of the mammary ductal network during puberty , and of the lobulo-alveolar units during pregnancy , was normal in mt/mt dams ( S1A and S1B Fig ) . The onset of milk protein synthesis also showed no defects during pregnancy by immunohistochemistry or western blot ( S1C Fig and S1D Fig ) . Lactation failure in mt/mt mice at 2 days post-partum ( 2dpp ) was seen as failure of alveolar expansion and retention of lipid droplets and colostrum ( Fig 1B–1G ) . Western blotting for milk ( Fig 1H and S1C and S1D Fig ) showed greatly reduced expression of all the major milk components at 2dpp relative to the level of the epithelial cell marker cytokeratin 18 . Quantitative PCR for the mRNAs for the milk proteins whey acidic protein ( WAP ) and β-casein ( β-Cas ) showed reduced levels in mt/mt dams ( mt ) compared to wt/wt dams ( wt ) at 18 days post-coitus ( dpc ) and especially at 2dpp ( Fig 1I and 1J ) . The number of cleaved Caspase-3 positive epithelial cells increased ( Fig 1K ) and BrdU incorporation by the epithelium was reduced , indicating increased cell death rate and decreased cell proliferation respectively ( Fig 1L ) . We used immunohistochemistry to examine STAT1 activation as it is an interferon regulated gene involved in mammary gland involution [8] . In wt/wt dams at d18 . 5 of pregnancy and 2 days post partum , we observed scattered regions of phosphorylated Stat1 staining in tightly packed areas of small and unexpanded alveoli ( Fig 1M ) . These regions were very rare at the other stages of development examined . In mt/mt animals Stat1 phosphorylation was again seen within regions of small unexpanded and tightly packed alveoli ( Fig 1N ) , but at day 18 . 5 of pregnancy , these regions of STAT1 phosphorylation occurred at a far greater frequency than in wt/wt glands , and instead of receding in the post partum period like wt/wt glands , the frequency of this pattern of staining increased further ( Fig 1O ) . We examined Stat1 phosphorylation in mammary glands formed by transplant of epithelium from mt/mt or wt/wt animals into the mammary fat pads of prepubescent wild type mice cleared of endogenous epithelium . We again observed a statistically significant increase in Stat1 phosphorylation in mt/mt transplants in the pre-partum period ( transplants can’t interrogate the post partum period ) , demonstrating that the ENU-mutation operates autonomously via the mammary epithelial cell ( Fig 1P ) . We used the Affymetrix Mouse Transcriptome Assay ( MTA ) 1 . 0 GeneChip to measure changes in gene expression underlying these events . We profiled RNA transcripts in the mammary glands at 18dpc and 2dpp from wt/wt and mt/mt mice . A Gene Set Enrichment Analysis ( GSEA ) of genes was carried out using the Limma t-statistic as a measure of ranked differential expression and visualized with the Enrichment Map plugin for Cytoscape . We compared gene expression changes between mt/mt and wt/wt mice at 18dpc or 2dpp ( Fig 2 , shown in detail S2 Fig ) . This identified a robust enrichment of a prominent cluster of gene sets involved in the interferon response in postpartum mt/mt but not wt/wt mammary glands , which increased in magnitude between 18dpc and 2dpp . Genes in these sets included the interferon-induced genes Isg15 , Mx1 , Rsad2 , Oas1 , Oas2 and OasL1 . Interferon-induced genes involved in the molecular pattern response pathway were also induced , such as Ddx58 ( RIG-1 ) , Dhx58 ( RIG-1 regulator ) , Mavs and Nlrc5 ( NOD5 ) . Additional downstream transcriptional regulators of the interferon response , such as Stat1 , Irf7 and Irf9 , were upregulated . In mt/mt glands this was accompanied by increased expression of a broad range of mitochondria-associated cell death genes such as Tnsfs10 ( TRAIL ) , Acin1 , Birc2 , Traf2 , Bcl2l1 ( BCL-XL ) , Bcl2l11 ( BIM ) , Apaf1 , Dffb , Xaf and Ripk1 . Very similar results were obtained using an independent analysis technique based on self-organizing maps ( S3 Fig ) . These results are also presented as a . txt table ( S1 Table ) of the 5000 probes showing most-changed expression . This transcriptional data indicates that a robust interferon response is induced by the mutation . PCR genotyping of polymorphic markers and their co-inheritance with lactation failure narrowed the mutation to a 4Mb region of chromosome 5 between rs3662655 and rs2020515 . We expected 4–8 ENU mutations per 4Mb and our strategy was to sequentially sequence exomes and then to experimentally validate when an exonic mutation was discovered . Sequencing revealed a T to A base change in Oas2 , resulting in a non-conservative isoleucine to asparagine amino acid substitution ( I405N; Fig 3A and S4A Fig ) in a conserved region of the OAS2 catalytic domain ( S4B and S4C Fig ) . In wt/wt animals Oas2 was expressed at a relatively low level until the establishment of lactation , when the level of Oas2 mRNA increased by 20 fold ( S5A Fig ) and subsequently fell during early involution . Changes in Oas2 expression in wt/wt animals compared to mt/mt animals are shown in S1E Fig . Using immunohistochemistry we observed corresponding changes in OAS2 levels in the mammary epithelium ( S5B Fig ) . We measured RNase L activity in the mammary glands of wt/wt and mt/mt mice at 18 days post coitus ( dpc ) and 2 days post partum ( dpp ) using a recently developed technique [9] . In wt/wt mice we observed a fall in RNase L activity from pre lactation at 18 dpp to lactation at 2 dpp despite the rise in OAS2 over this period ( Fig 3B top panel ) . In contrast mt/mt animals showed an increase in RNase L activity over this period , so that at 2 dpp , RNase L activity was 34 fold higher in mt/mt animals . PCR for RNase L-cleaved rRNA showed a six-fold increase RNase L activity ( Fig 3B lower panel ) , while non-RNase L generated cleavage was negligible . Bioanalyzer profiles of RNA ( S5C Fig ) showed increased RNA degradation in mt/mt animals , but not to the extent that appreciable loss of the 18S or 28S ribosomes was seen , and which may be a result of both RNase L dependent and independent mechanisms . Although robust activation of RNase L can cause the loss of the 18S and 28S ribosomes [10] , recent findings show that ribosomal degradation is not required for RNase L to stop protein synthesis [9] . To determine if the mutation altered OAS2 enzyme activity we purified the mutant and wild type forms of mouse OAS2 expressed in HeLa cells by FLAG-immunoprecipitation . Using a cell-free system we observed that both mutant and wild type forms of OAS2 showed induction of enzyme activity by the double-stranded RNA mimic poly ( I:C ) , seen as the formation of a series of 2-5A species resolved by denaturing PAGE . Both mutant and wild type forms of OAS2 showed similar sensitivity to increasing poly ( I:C ) concentrations ( Fig 3C , quantified in Fig 3D ) . Western blotting showed that the immunoprecipitates used in these experiments had similar OAS2 levels ( Fig 3E ) . These experiments show that the ENU-induced mutation in OAS2 does not change the size range of oligoadenylates that it produces , its capacity for 2-5A synthesis , or its sensitivity to activation by poly ( I:C ) . This assay uses a cell free system , so we cannot exclude a mechanism where mutant OAS2 activates RNase L activity via an indirect effect to increase the active 2-5A pool without altering its rate of synthesis , such as reduced 2-5A depletion or loss of 2-5A sequestration . Another possibility is that mutant OAS2 has an altered molecular interaction with a species that increases its enzymatic activity , but which is lost in the immunoprecipitation of OAS2 in this assay . Regardless , the mutation in Oas2 activates RNase L in mice and tissue culture models . We constructed a model of doxycycline ( Dox ) -inducible expression of mutant or wild type Oas2 in T47D human breast cancer cells ( Fig 4A ) . These models produced a 20-fold induction of Oas2 expression ( Fig 4B ) . Western analysis showed the appearance of mouse OAS2 protein following Dox administration just below endogenous human OAS2 , both above a non-specific band ( Fig 4C ) . Thus although PCR showed a small amount of leakiness in this system it seems negligible by western blot . Cells expressing either mutant or wild type Oas2 showed a similar sensitivity to poly ( I:C ) that was not changed significantly by induction with Dox ( Fig 4D ) . Induction of mutant , but not wild type Oas2 for 72 h reduced cell number ( Fig 4E ) . Increased cell detachment was observed , but the magnitude of this effect was highly variable between experiments using mutant cells and so did not reach statistical significance at p<0 . 05 ( Fig 4F ) . Reduced re-plating efficiency however following trypsinization was significant , indicting that cell surface re-expression of adhesion molecules following their trypsin digestion was impaired ( Fig 4G ) . Western analysis of two of these molecules , Beta-1 Integrin ( ß1 ) and E-Cadherin ( EC ) , showed reduced expression in response to Dox-induction of mutant Oas2 , especially for Beta-1 Integrin , shown in the far right hand side lane ( Fig 4H ) . We used flow cytometry to simultaneously measure cell viability by propidium iodide exclusion and cell death by cell surface expression of Annexin V , in response to Oas2 expression . While induction of wild type Oas2 expression produced no apoptotic response , induction of mutant Oas2 produced a doubling in the number early apoptotic cells within the cultures ( Fig 4I ) . Induction of wild type Oas2 did not alter the distribution of cells among the phases of the cell cycle , while induction of the mutant produced a shift of cells out of S-phase and into G1 ( Fig 4J ) . Thus the effects of mutant Oas2 expression in T47D cells reproduce the phenotypes of cell death and reduced cell proliferation seen in the mouse , and indicate that epithelial cell adhesion may also be affected . Mouse HC-11 cells express milk proteins in response to withdrawal of EGF and the addition of prolactin and dexamethasone , providing a way to examine the effects of mutant and wild type Oas2 expression on milk protein expression . The inducible vector system used successfully in T47D cells ( Fig 4A ) proved to be very leaky in HC-11 cells , resulting in high baseline expression of Oas2 in the pooled clones without DOX treatment . Cloning , in an attempt to find cells without leaky expression , was unsuccessful , but resulted in cell lines with similar levels of constitutive expression of mutant or wild type Oas2 that was many fold greater than seen in untransfected cells ( Fig 4K ) . Treatment with prolactin and dexamethasone induced beta casein levels in the cell line expressing wild type Oas2 , and this effect was comparable in magnitude to that seen in parental HC-11 cells , but in the two lines expressing mutant Oas2 the induction of beta casein was greatly reduced ( Fig 4L ) , reproducing the suppression of milk protein synthesis seen in the ENU-mutant mouse . Transient expression of wild type and mutant Oas2 in HC11 cells also showed an increase in the basal rate of cell death , reproducing the cell death phenotype ( Fig 4M ) . We used Affymetrix Human Transcriptome Assay 2 . 0 GeneChips to profile the changes in gene expression that occurred in T47D cells when either wild type or mutant Oas2 was induced for 72h , presented as GSEA/Cytoscape ( S6 Fig ) , self organizing maps ( S7 Fig ) and as table containing the top 500 differentially expressed genes ( S2 Table ) . We compared the transcriptional effects of mutant Oas2 in T47D cells to the effects in the ENU mouse shown in Fig 2 using Cytoscape ( S8 Fig ) , or self-organizing maps ( S7 Fig ) . The transcriptional effects of mutant OAS2 in T47D cells were very similar to those observed in the ENU-mutant mouse , with the interferon response most prominent . This demonstrates that expression of mutant but not wild type Oas2 in T47D cells reproduces the molecular phenotypes observed in the ENU mutant mice . While the phenotype in mice is likely to involve additional cells of the immune system , these effects in T47D cells show that the transcriptional phenotype can be elicited via the innate immune response of the mammary epithelial cell , in agreement with the findings made using transplanted ENU-mutant mammary epithelium into wild type mice ( Fig 1P ) . OAS2 activates RNaseL . In T47D cells we used siRNA against human RNASEL to knockdown its expression in the context of Dox-induction of mutant or wild type mouse Oas2 . In these experiments the induction of Oas2 in response to Dox was robust and knockdown of RNASEL was very effective , as demonstrated by qPCR ( Fig 5A and 5B ) and by western blot ( Fig 5C ) . Induction of wild type Oas2 had no effect on RNase L activity , cell death or cytokine levels and knockdown of RNASEL was without consequence to these endpoints . In contrast , induction of mutant Oas2 produced a large increase in RNase L activity , cell death , and interferon gamma and GM-CSF protein secretion , changes that were prevented by knockdown of RNASEL ( Fig 5D–5G ) . Expression of the IRF transcription factors , especially IRF7 , was increased by mutant OAS2 . We knocked down IRF7 ( Fig 5H ) and found a similar prevention of cell death ( Fig 5I ) , indicating that the signaling pathway activated by mutant OAS2 also involves IRF7 , a distal member of the viral-detection signaling pathway . Knockdown of IRF3 ( Fig 5J ) , which often acts together with IRF7 , had the opposite effect ( Fig 5K ) , suggesting IRF3 acts to oppose signaling via the OAS2 pathway .
These experiments show that the Oas2 mutation caused activation of OAS2 driven signaling to prevent the activation of lactation in the post partum period . The effect of the mutation could be detected via Stat1 activation from mid pregnancy and was most apparent in the post partum period , and was only required in the mammary epithelial cell for effect . The mutation increased RNase L activity but the enzymatic activity of mutant OAS2 was unaltered . Thus RNase L activation must occur via mechanisms that increase the effect of 2-5A without a change in its production , such as by reducing 2-5A degradation , increasing the efficiency of 2-5A interaction with RNase L or OAS2 interaction with dsRNA , or by causing relief of a mechanism that sequesters 2-5A . The activation of RNAse L is not sufficient to degrade the ribosomes , indicating that the loss of milk production does not occur via a generalized loss of translation . Thus while RNase L expression is required for activity of the mutation , the mutation may act via regulatory mechanisms that do not require 2-5A activation of RNase L . RNase L may be simply permissive of an alternative mechanism of action , such as altering interactions of OAS2 with its cellular binding partners , by changing its subcellular localization , or by decreasing the rate of OAS2 degradation . Thus it is possible that RNase L and OAS2 could also both be involved in as yet undiscovered molecular complexes that initiate activation of this pathway . For example OAS2 has been reported to bind NOD2 [3] , and the composition and mechanism of action of this mitochondrial-signaling complex is currently the subject of intense worldwide study , but its definition is proving to be elusive . The non catalytic OAS1b [11] and OASL1 [12] have mechanisms independent of 2-5A production involving molecular interactions . This is the first genetic demonstration that OAS2 can signal in ways other than by alterations in enzyme activity . This mutation may prove to be important for the discovery of the mechanisms signaling the detection of viral infection , which remain largely unknown , because it provides a single point of pathway activation , unlike the existing reagents used for this purpose . Like other family members , OAS2 may regulate apoptosis independently of its function to control viral replication [6 , 7] . Lactation failure and milk stasis characterize mastitis , raising an interesting new avenue of investigation opened by our findings . The major consequence of mastitis is reduced weight-gain of the infant , precipitating a switch to bottle-feeding where available , or reduced neonatal health where it is not . Our results raise the possibility that the OAS2 pathway may be involved in its pathogenesis . Bacterial infection is commonly thought to be the cause of mastitis but the evidence resoundingly shows that bacterial infection is the sequelae of an unknown primary cause of the disease . For example , in women the severity of symptoms of mastitis do not correlate with the level of bacterial infection , the disease is often observed in the absence of bacteria in the milk , bacteria are often found in the milk of healthy mothers , and meticulous hygiene or prophylactic antibiotics do not prevent mastitis ( reviewed [13] ) . Recent Cochrane Systematic Reviews concluded that there is insufficient evidence to support antibiotic use for the prevention [14] or treatment [15] of mastitis . The strongest risk factors for mastitis in women involve incomplete or interrupted milk flow from one or more galactophores [13] and the World Health Organization recognizes milk stasis as the cause of mastitis [16] . Thus bacterial infection most likely represents progression of mastitis to a more pathogenic form involving abscess formation , but it is not the primary cause . The concept of physiological inflammation as the primary cause of mastitis was proposed in 2001 , though no mechanism was proposed at the time [17] , and the unavailability of breast tissue from women with mastitis makes the study of mechanism near impossible . Using mice , Ingman and colleagues hypothesize that molecular pattern receptors like Tlr4 recognize molecules released by tissue damage caused by milk engorgement , which trigger an innate immune response and milk stasis [13 , 18] . Alleles of Tlr4 , a bacterial associated molecular pattern receptor , are linked with the occurrence of mastitis in cattle [19] . Tlr4 has also been linked to a number of the systemic symptoms of mastitis [13] . As we show , stimulation of the OAS2 pathway can produce the accepted cause of mastitis , milk stasis , opening a new avenue of investigation into human mastitis as a disease amenable to anti-inflammatory therapy . Our findings also open the question of the role of viruses in the initiation of mastitis . Even non-infectious forms could play a role . Fragments of the mouse mammary tumor virus are present in the genome of all laboratory mice and they continue to produce transcripts in response to the hormones of pregnancy , while homologous fragments exist in the human genome [20 , 21] , which may promote milk stasis and inflammation via OAS2 activation . This is the first time that a viral recognition pathway has been implicated in the regulation of lactation . Transmission of viruses via milk is a well-documented phenomenon and the evolution of mechanisms to prevent it would be expected . This would not necessarily be lethal for the neonate as all mammals have multiple and independent lactation systems . Mice , for example have 10 mammary glands each containing a single ductal system . Each human breast contains between 6 and 8 independent ductal systems , exiting at the nipple without joining . Viral infection in one ductal system , or one mammary gland , could initiate milk stasis in that system , leaving the others to continue lactation . Social systems in humans and mice allow the feeding of neonates by multiple mothers . There could be an intriguing evolutionary twist here resulting from the evolutionary arms-race between viruses and their hosts [22] . Since HIV transmission via the milk occurs far more frequently if mastitis is present [23] , could viruses have adapted to this defense and learned to induce a limited mastitis to aid viral transmission ? A molecular mechanism is suggested by our results ( S9 Fig ) that requires further investigation . STAT1 activation ( Fig 1 ) , presumably resulting from the production of interferon due to OAS2 pathway activation , would be expected to cause the induction of the SOCS proteins , which inhibit STAT phosphorylation via targeting the JAK kinases , including JAK2 which phosphorylates STAT5 in response to prolactin , the major hormone driving the onset of lactation . Many aspects of this pathway have been demonstrated in mice such as the regulation of lactation by prolactin via STAT5 [24 , 25] and the SOCS proteins [26–28] , the induction of STAT1 in conditions of sterile mastitis [29] and the ability of STAT1 to regulate prolactin signaling [30] . In the T47D transcript profiling ( S6–S8 Figs and S2 Table ) we observed increases in the levels of SOCS 1 , 4 , 5 and 6 . In the ENU mice we observed a decrease in STAT5 phosphorylation . So it is possible that OAS2 pathway stimulation , resulting from the natural rise in OAS2 at d18 . 5 of pregnancy ( S1 and S5 Figs ) , produces a persistent interferon response in OAS2 mutant animals , because mutant OAS2 activates RNase L which via the resulting interferon response maintains high OAS2 levels , establishing a positive feed-back loop which then persistently prevents prolactin from activating STAT5 ( maybe via SOCS ) to induce the activation of milk secretion during the post partum period .
All mice were housed in specific pathogen-free conditions at the Australian Phenomics Facility and the Garvan Institute , with all animal experiments carried out according to guidelines contained within the NSW ( Australia ) Animal Research Act 1985 , the NSW ( Australia ) Animal Research Regulation 2010 and the Australian code of practice for the care and use of animals for scientific purposes , ( 8th Edition 2013 , National Health and Medical Research Council ( Australia ) ) and approved by either the Australian National University or Garvan/St Vincent’s Animal Ethics and Experimentation Committees ( approval number 14/27 ) . ENU mutagenesis and pedigree construction was carried out as previously described [31] . The Oas2 mutation was discovered in a single G1 female and heritability of the phenotype confirmed by breeding with CBA CaJ male and cross fostering of pups . For quantification of lactation failure litters were standardized to 7 pups per dam . Pups were weighed , as a group , at the same time daily . Mice were injected with BrdU dissolved in H2O ( 100μg BrdU per gram body weight ) 2 h prior to sacrifice by CO2 asphyxiation , and mammary glands were collected . Mammary glands were either whole-mounted and stained with Carmine alum or snap frozen in liquid nitrogen for mRNA and protein analyses . All animals were housed with food and water ad libitum with a 12-h day/night cycle at 22°C and 80% relative humidity . A complete analysis of the histology and pathology of the Jersey strain was conducted by the Australian Phenomics Network ( APN ) Histopathology and Organ Pathology Service , University of Melbourne . Eight week and a 31 week female sibling pairs , comprised of mt/mt and wt/wt siblings , were examined macro and microscopically . Mammary tissue , ovaries , oviducts , bicornuate uterus , cervix , urinary bladder , liver/gall bladder , cecum , colon , spleen/pancreas , mesenteric lymph node , stomach , duodenum , jejunum , ileum , kidney/adrenal , salivary glands/lymph nodes , thymus , lungs , heart , skin , eyes , brain , spinal cord , skeletal muscle , skeletal tissue/hind leg were macro and microscopically examined . A pool of 15 affected N2 mice and a pool of 15 unaffected N2 backcrossed mice were screened with a set of ~130 markers polymorphic between C57BL/6 and CBA/CaJ mice that spanned the genome at 10–20 Mb intervals . Allele specific SNP primers were designed from a set of validated SNPs available at www . well . ox . ac . uk/mouse/INBREDS/ . SNP genotyping was performed using the Amplifluor kit ( Chemicon ) as per the manufacturers instructions . The confirmation and fine mapping were performed using Amplifluor markers designed to amplify C57BL/6 x CBA/CaJ SNPs within the linkage interval in individual affected and unaffected mice . Markers were designed at approximately 1–2 Mb distances within the initial map interval . More than 250 mice were screened from many successive cohorts of mice from backcrosses to CBA/CaJ to narrow the region to a 3 Mb interval . Sequencing of candidate genes was performed to locate the causal ENU base substitution using an affected mouse homozygous for the linkage region . Primers were designed for candidate genes to amplify all exons +/- 15 bp to cover splice junctions . Sanger sequencing was used to identify the causal mutation by comparing the sequence of the affected individual against a C57BL/6 mouse reference genome . Mutations were confirmed in a second affected individual and a C57BL/6wild type mouse . The mapping was performed with the gsMapper program , which is part of the 454 software suite . The two samples ( defined by Jersey_F4IC140 and Jersey_pool ) were mapped against the full region of the mouse genome on chromosome 5 . The sequence used as reference is from genbank build37/UCSC mm9 . Further analysis then focused on the reads mapped onto the target region: 118710087–123738720 on chromosome5 . Variation analysis detected where at least 2 reads differ either from the reference sequence or from other reads aligned at a specific location . SNPs , insertion-deletion pairs , multi-homopolymer insertion or deletion regions , and single-base overcalls and under calls are reported . Also , in order for a difference to be identified and reported , there have been at least two non-duplicate reads that ( 1 ) show the difference , ( 2 ) have at least 5 bases on both sides of the difference , and ( 3 ) have few other isolated sequence differences in the read . Variations were classified as high-confidence if they fulfilled the following rules: 1 . There must be at least 3 non-duplicate reads with the difference . 2 . There must be both forward and reverse reads showing the difference , unless there are at least 5 reads with quality scores over 20 ( or 30 if the difference involves a 5-mer or higher ) . 3 . If the difference is a single-base overcall or under call , then the reads with the difference must form the consensus of the sequenced reads ( i . e . , at that location , the overall consensus must differ from the reference ) . After identification of the causative mutation genotyping was performed using the following primers: The Oas2 mutant colony was maintained by breeding heterozygous males ( wt/mt ) with wt/wt females . For the generation of homozygous experimental animals wt/mt males were bred with wt/mt females and their offspring removed at 1dpp and fostered on a control mother . Total RNA was isolated using Trizol reagent ( mouse tissues; Gibco/Invitrogen Vic ) or RNeasy extraction kit ( cell pellets; Qiagen ) according to the manufacturer’s instructions . All total RNA samples were quantified with a Nanodrop 1000 Spectrophotometer ( ThermoFisher ) prior to loading 100 ng of total RNA on a 2100 Bioanalyzer ( Agilent ) total RNA analysis . Single stranded cDNA was produced by reverse transcription using 1 μg of RNA in a 20μl reaction and diluted 1:5 with H2O ( Promega ) . Quantitative PCR was performed using the Taqman probe-based system ( Table 1 ) on the ABI 7900HT as per the manufacturer’s instructions ( Applied Biosystems ) . tRNA cleavage by RNaseL was performed using the technique developed by JD and AK [9] . Briefly , RNA was ligated with 2’ , 3’-cyclic phosphate to an adaptor with RtcB as described in [9] . Reactions were stopped by adding EDTA and used as a template for reverse transcription with Multiscribe RT . A primer with a 3’-end complimentary to the adaptor and a 5’-overhang that serves as a universal priming site ( 5’-TCCCTATCAGTGATAGAGAGTTCAGAGTTCTACAGTCCG- 3’ ) was used for reverse transcription . SYBR green-based qPCR was conducted using a universal reverse primer that binds to the cDNA overhang ( underlined ) and cleavage-site specific forward primers designed for tRNA [9] . qPCR was carried out for 50 cycles using 62°C annealing/extension for 1 min . U6 , which has a naturally occurring 2’ , 3’-cyclic phosphate and an RNase L independent cleavage site in tRNA-His ( position 18 , transcript numbering; [9] ) was used for normalization . HeLa cells were maintained in MEM + 10% FBS in a humidified 5% CO2 atmosphere . Cells cultured in 10 cm dishes were transfected at 80–90% confluence with 10 μg empty pcDNA4/TO or N-terminally FLAG-tagged WT or I405N mouse OAS2 in pcDNA4/TO ( Life Technologies ) using Lipofectamine 2000 ( Life Technologies ) . Cells were harvested by trypsinization 24h post-transfection , resuspended in complete media , and washed 2 x 10 mL ice-cold PBS . Cell pellets were lysed in buffer A ( 20 mM HEPES pH 7 . 5 , 100 mM NaCl , 0 . 1% Triton X100 , and 1x complete protease inhibitor cocktail ( Roche ) for 10 min with end-over-end rotation at 4°C . Lysates were cleared by centrifugation at 16 , 000 x g , 15 min , 4°C and the supernatants subjected to immunoprecipitation with M2-α-FLAG magnetic beads ( Sigma ) for 2h at 4°C followed by 4 x 1mL washes , 10 min per wash , with buffer A without protease inhibitors . After the fourth wash the beads were washed with 2 x 1 mL storage buffer ( 20 mM HEPES pH 7 . 5 , 100 mM NaCl , 10% glycerol ) , the supernatant removed , storage buffer added to 100 μL . Inputs and IPs were blotted with α-FLAG to verify expression and IP of FLAG-OAS2 . 5% of each IP was incubated with 1 mM ATP , trace-labeled with 3 nM 32P-α-ATP , in the absence or presence poly ( I:C ) ( Sigma ) for 2h at 37°C . Reaction volumes were 20 μL and contained 20 mM HEPES pH 7 . 5 , 70 mM NaCl , 10 mM MgCl2 , 10% glycerol , and 4 mM DTT . After incubation , the reactions were quenched by adding 120 μL stop buffer ( 8M urea , 0 . 1% SDS , 1 mM EDTA , 0 . 02% bromophenol blue , 0 . 02% xylene cyanol ) . Equal portions of each reaction were resolved by 20% denaturing PAGE and visualized by phosphoimaging . Gels were quantitated using GelQuant . NET software . The N—terminal and C—terminal cDNAs of mouse Oas2 were obtained as a gift of Yoichiro Iwakura ( Institute of Medical Science , University Tokyo ) and subcloned into pcDNA3 . 1 and pBluescript vectors . Site directed mutagenesis was performed using Phusion Site-directed mutagenesis ( Thermo Scientific ) as per the manufacturer’s instructions using the following primers ( Forward Jer: TATATGTTCCTTCCTTAAAAATGTCTGC and Reverse AGGATTTCGTCTTGTTCCTTCGACAACTGTA ) . Wildtype and mutant ( I405N ) mouse Oas2 clones were then subcloned into pShuttle and finally into the pHUSH ProEx tetracycline inducible retroviral expression system [32] . Retrovirus was then packaged by transfecting Phoenix cells with pHUSH containing either mouse wildtype ( wt ) or Oas2 I405N mutant ( mt ) cDNAs using Fugene transfection reagent ( Promega ) . T47D breast cancer cells were then infected with filtered viral supernatants and stable cell lines selected using Puromycin . T47Ds were maintained sub-confluent in RPMI complete media ( Gibco ) containing 10% tetracycline free FCS and supplemented with 10μg/ml Insulin . Mouse Oas2 wt or mt expression was induced with 100 ng/ml Doxycline ( DOX ) or vehicle control daily in the media and cells were harvested at 72 hours after plating . Cell counts of viable and non-viable cells ( identified by the incorporation of 0 . 4% Trypan Blue at a 1:2 dilution ) were performed in triplicate from 3 independent experiments . Annexin V PI staining was performed using the Annexin V-FITC Apoptosis Kit ( Biovision , CA USA ) as per the manufacturers instructions . Human inflammatory cytokines were analyzed using the Multi-Analyte ELISArray ( Qiagen ) . Retrovirus was made packaged as above by transfecting Phoenix cells with pHUSH containing either mouse wildtype ( wt ) or Oas2 I405N mutant ( mt ) cDNAs using Fugene transfection reagent ( Promega ) . HC11 normal mouse mammary cell lines were then infected with filtered viral supernatants and stable constitutive cell lines selected using Puromycin and then clonal colonies established by titrating single cells into 96 well plates . HC11s were maintained in maintenance RPMI media containing 10% FCS and supplemented with 5μg/ml Insulin and 10ng/ml human recombinant epidermal growth factor ( EGF , Sigma-Aldrich ) . For differentiation assays , HC11 cells were plated sub confluent in maintenance media for 2 days until confluent and then media replaced with RPMI media containing 10% FCS and supplemented with 5μg/ml Insulin and supplemented with 5μg/ml sheep pituitary Prolactin ( Sigma Aldrich ) and 1μM Dexamethasone ( Sigma Aldrich ) daily for 3 days before RNA harvest and quantitative PCR analysis as above . For transient transfections apoptotic assays wildtype ( wt ) or Oas2 I405N mutant ( mt ) cDNAs were cloned into pIRES-EGFP vectors and transiently transfected with 2μg DNA/well using X-tremeGENE transfection reagent ( Roche ) in maintenance media as per the manufacturers instructions . After 24 hours the media replaced with RPMI media containing 10% FCS and supplemented with 5μg/ml Insulin and cells were harvested 96 hours after transfection for Annexin V PI staining as above . T47D Oas2 wt or mt cells were plated in 10cm dishes and treated daily for 48 hours with 100 ng/ml DOX or vehicle . At 48 hours the cells harvested by trypsinization , retreated with DOX or vehicle , counted , plated at a density of 5000 cells/well and simultaneously transfected with Poly ( I:C ) ( 11 point titration ) using RNAiMAX ( Invitrogen ) in opaque 96 well plates . 24 hours after transfection the plates were analysed using the CellTiter-Glo Luminescent Cell Viability Assay Protocol ( Promega ) . Inhibitory dose curves were plotted in Prism 6 statistical software and normalized data analyzed using the sigmoidal-dose response function . The mean was calculated from quadruplicate replicates . T47D Oas2 wt or mt cells were plated in T150 flasks and mouse Oas2 wt or mt expression was induced with 100 ng/ml DOX or vehicle for 48 hours prior to trypsinisation by 0 . 25% Trypsin ( no EDTA ) with phenol red ( Life Technologies ) for 3 mins . 1 x 106 cells from each cell line was treated with DOX or vehicle and then plated in 6-well plates and allowed to adhere for 4 hours . After 4 hours the cells were gently washed , trypsinised and the number of adherent cells counted and expressed as a proportion of the total number of cells plated . T47D Oas2 wt or mt cells were plated in 6-well plates and mouse Oas2 wt or mt expression was induced with 100 ng/ml DOX or vehicle for 72 hours . Cells were harvested by trypsinisation , washed in 1 ml of PBS and fixed by adding 10 mls of 100% cold ethanol drop wise onto 1ml re-suspended cells and incubated at 4°C overnight . Cells were then pelleted , washed and incubated at 90°C for 5 mins and then re-suspended in a FACS buffer containing 0 . 5ng/ml RNase ( Qiagen ) and 1μg/ml Propidium iodide . Flow cytometry was performed and G1 , S phase and G2/M phases for each experimental group estimated using the propidium iodide fluorescence intensity histograms . The mean of 5 independent experiments was calculated . T47D Oas2 wt or mt cells were reverse transfected with ON-TARGET plus SMARTPOOLS of siRNA against RNaseL ( L-005032-01-05 ) or Non Targeting controls ( D-001810-10-05 ) using RNAiMAX ( Invitrogen ) in 10cm dishes as per the manufacturers specification . 24 hours after transfection , cells were washed , media replaced and either treated with 100 ng/ml DOX or vehicle daily for 3 days after which they were harvested by trypsinisation . Annexin V PI staining was performed using the Annexin V-FITC Apoptosis Kit ( Biovision , CA USA ) as per the manufacturers instructions . Cell pellets were also collected for RNA isolation and western blotting and the supernatant collected , filtered with a 0 . 22μm filter and stored at -80°C . Human inflammatory cytokines were analyzed using the Multi-Analyte ELISArray™ Kits as per the manufacturers instructions . Wt/wt or mt/mt mice were time mated and mammary glands collected at day 18 of pregnancy or 2 days after partuition ( 2dpp ) and snap frozen in liquid N2 . Total RNA was isolated using Trizol reagent ( Gibco/Invitrogen , Vic ) and measured on the 2100 Bioanalyzer ( Agilent ) . From the cell experiments , total RNA was extracted using the RNeasy extraction kit ( Qiagen ) for cells with or without DOX induction of the wt or mt mOas2 gene . Total RNA from the mouse mammary glands was then labeled and hybridized to the Mouse Transcriptome Array ( MTA ) 1 . 0 as per the manufacturer’s instructions ( Affymetrix Ca , USA ) at Ramiaciotti Centre for Genomics ( UNSW Sydney , Australia ) . Likewise , total RNA from the T47D cells was labeled and hybridized to the Affymetrix Human Transcriptome Array ( HTA ) 2 . 0 as per the manufacturer’s instructions ( Affymetrix Ca , USA ) at Ramiaciotti Centre for Genomics ( UNSW Sydney , Australia ) . All mouse and T47D samples were prepared in biological triplicate for each experimental grouping , except for the T47D mt -DOX group where analysis was performed in duplicate due to one of the samples failing quality control . Microarray data are freely available from GEO: GSE69397 http://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? token=ohkleoespjotzoh&acc=GSE69397 Quality control was performed using the Affymetrix Expression Console . Normalisation and probe-set summarization was performed using the robust multichip average method of the Affymetrix Power Tools apt-probeset-summarize software ( version 1 . 16 . 1 ) ( using the -a rma option ) . The transcript clusters with official HGNC symbols were then extracted from the HTA 2 . 0 arrays , resulting in 23532 gene transcript clusters . Differential expression between experimental groups was assessed using Limma [33] via the limmaGP tool in GenePattern . Functionally associated gene-sets were identified using Gene Set Enrichment Analysis ( GSEA ) [34] on a ranked list of the limma moderated t-statistics , from each pair-wise comparison , against a combined set of 6947 gene-sets from v4 . 0 of the MSigDB [35] and custom gene-sets derived from the literature . Mouse gene-symbols were mapped to their human orthologs using the ensembl database . The Enrichment Map plugin [36] for Cytoscape [37] was used to build and visualize the resulting regulatory network of gene-signatures , with conservative parameters: p = 0 . 001; q = 0 . 05; overlap s = 0 . 5 . 10μg reduced protein was loaded in each well of 12% NuPAGE SDS polyacrylamide gels ( Life Technologies ) and separated using electrophoresis . Proteins were transferred to Immun Blot PVDF ( Biorad ) and Western blotted for mouse Oas2 ( M-105 , sc99098 Santa-Cruz ) , RNaseL ( H-300 , sc25798 Santa Cruz ) , E-cadherin ( 610182 BD Biosciences ) and beta-ACTIN ( AC-74 , A5316 , Santa Cruz ) The limma F-test statistic [33] , with a Benjamini-Hochberg adjusted p-value threshold of 0 . 05 , was used to identify differentially expressed transcripts across the four experimental groups in the mouse expression arrays ( wt/wt 2dpc , wt/wt 2dpp , mt/mt 2dpc , mt/mt 2dpp ) and T47D cell-line expression arrays ( wt–Dox , wt +Dox , mt–Dox , mt +Dox ) . This resulted in 660 and 135 significant transcript clusters from the mouse and T47D arrays , respectively . Self-organising maps ( SOMs ) , consisting of 6 nodes , were used to identify clusters of genes in both the mouse and T47D cells . The z-scores of the log2 normalised gene-expression values , for each transcript cluster , were used as input to the biopython SOM algorithm implementation [38] . The somcluster ( ) parameters used were: iterations = 50 , 000; nx = 2 , ny = 3 , inittau = 0 . 02 , dist = Euclidean . The db2db ( ) function from the BioDBNet database [39] was used to convert gene-symbols to Ensembl gene IDs for input into DAVID . Functional annotation clustering was carried out using the getTermClusterReport ( ) function from the DAVID web services interface ( Jiao et al . , 2012 ) , with the following parameters: overlap = 3 , initialSeed = 3 , finalSeed = 3 , linkage = 0 . 5 , kappa = 50 . DAVID databases used: ( BBID , GOTERM_CC_FAT , BIOCARTA , GOTERM_MF_FAT , SMART , COG_ONTOLOGY , SP_PIR_KEYWORDS , KEGG_PATHWAY , INTERPRO , UP_SEQ_FEATURE , OMIM_DISEASE , GOTERM_BP_FAT , PIR_SUPERFAMILY ) A Hypergeometric test was used to calculate the level of gene overlap between the genes identified in each SOM cluster and the MSigDB gene-set collections [35] and our custom functional signature gene-sets . A background set number , of 45956 , as described on the MSigDB website , was used . A Benjamini-Hochberg ( BH ) corrected p-value was calculated for each set and a threshold of BH<0 . 05 was considered a significant enrichment . Mouse gene-symbols were mapped to their human orthologs using the ensembl database . Mouse mammary glands were harvested from wt/wt and mt/mt mice and fixed in 4% buffered formalin for 4 hours . Glands were defatted in 3–4 changes of acetone before being dehydrated and stained in Carmine alum as previously described [40] . Glands were then dehydrated in a series of graded alcohols and embedded in Paraffin for sectioning . Sections were either stained with haematoxylin and eosin for routine histochemistry or stained with antibodies to the following antigens using immunohistochemistry protocols as detailed in Table 2 . | Using ENU-mutagenesis in mice we discovered a pedigree with lactation failure . Mammary development through puberty and pregnancy appeared normal in mutant animals , but the activation of lactation failed in the immediate post partum period and no milk reached the pups . Failure of lactation was accompanied by greatly diminished milk protein synthesis , decreased epithelial cell proliferation , increased epithelial cell death and a robust interferon response . A non-synonymous mutation in Oas2 ( I405N ) in the viral sensor Oas2 was found and expression of mutant Oas2 in mammary cells recapitulated these phenotypes . RNase L , the most proximal effector of OAS2 action , was activated in the mammary glands of mutant mice and in mammary cells expressing mutant Oas2 . Knockdown of RNase L , or the distal pathway member IRF7 , prevented these effects , indicating that the mutation in OAS2 caused activation of the viral signaling pathway . These results show that viral detection in the mammary gland can prevent lactation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"milk",
"reproductive",
"system",
"body",
"fluids",
"maternal",
"health",
"nucleases",
"enzymes",
"gene",
"regulation",
"dna-binding",
"proteins",
"enzymology",
"diet",
"reproductive",
"physiology",
"endocrine",
"physiology",
"epithelial",
"cells",
"nutrition",
"women's",
"health",
"mastitis",
"small",
"interfering",
"rnas",
"animal",
"cells",
"proteins",
"endocrinology",
"ribonucleases",
"gene",
"expression",
"lactation",
"exocrine",
"glands",
"biological",
"tissue",
"breast",
"tissue",
"biochemistry",
"rna",
"hydrolases",
"anatomy",
"nucleic",
"acids",
"cell",
"biology",
"physiology",
"beverages",
"genetics",
"epithelium",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"non-coding",
"rna",
"mammary",
"glands"
]
| 2017 | A mutation in the viral sensor 2’-5’-oligoadenylate synthetase 2 causes failure of lactation |
Sequencing of whole cancer genomes has revealed an abundance of recurrent mutations in gene-regulatory promoter regions , in particular in melanoma where strong mutation hotspots are observed adjacent to ETS-family transcription factor ( TF ) binding sites . While sometimes interpreted as functional driver events , these mutations are commonly believed to be due to locally inhibited DNA repair . Here , we first show that low-dose UV light induces mutations preferably at a known ETS promoter hotspot in cultured cells even in the absence of global or transcription-coupled nucleotide excision repair ( NER ) . Further , by genome-wide mapping of cyclobutane pyrimidine dimers ( CPDs ) shortly after UV exposure and thus before DNA repair , we find that ETS-related mutation hotspots exhibit strong increases in CPD formation efficacy in a manner consistent with tumor mutation data at the single-base level . Analysis of a large whole genome cohort illustrates the widespread contribution of this effect to recurrent mutations in melanoma . While inhibited NER underlies a general increase in somatic mutation burden in regulatory elements including ETS sites , our data supports that elevated DNA damage formation at specific genomic bases is at the core of the prominent promoter mutation hotspots seen in skin cancers , thus explaining a key phenomenon in whole-genome cancer analyses .
Whole genome analysis of cancer genomes has the potential to reveal non-coding somatic mutations that drive tumor development , but it remains a major challenge to separate these events from non-functional passengers . The main principle for identifying drivers is recurrence across independent tumors , suggestive of positive selection , which led to the recent identification of frequent oncogenic mutations in the promoter of telomere reverse transcriptase ( TERT ) that can activate its transcription [1 , 2] . However , mutation rates vary across the genome , and local elevations may give rise to “false” recurrent events that can be misinterpreted as signals of positive selection . While known covariates of mutation rate , such as replication timing and chromatin organization [3 , 4] , transcriptional activity [5] and local trinucleotide context [6] , can be accounted for to improve interpretation [7] , the non-coding genome may be particularly challenging . Mutational fidelity may be generally reduced in this vast and relatively unexplored space , as indicated by the presence of mechanisms directing DNA repair specifically to gene regions [8 , 9] , and yet-unexplained mutational phenomena may be at play . Indeed , recent studies have described a remarkable abundance of recurrent promoter mutations in melanoma and other skin cancers , often noted to overlap with sequences matching the recognition element of ETS family transcription factors ( TFs ) [10–16] . Strikingly , a large proportion of frequently recurring promoter mutations in melanoma occur at distinct cytosines one or two bases upstream of TTCCG elements bound by ETS factors as indicated by ChIP-seq , within a few hundred bases upstream of a transcription start site [17] . While often interpreted as driver events , we recently showed that these sites exhibit highly elevated vulnerability to UV mutagenesis , as evidenced by their rapid induction following low-dose UV light exposure in cultured cells [17] . The effect has often been attributed to locally impaired nucleotide excision repair ( NER ) caused by binding of ETS TFs [16 , 18 , 19] . However , our analysis of skin tumors lacking global NER ( XPC -/- ) contradicted this model [17] and the mechanism remains unclear . An understanding of this phenomenon , which may underlie a large part of all non-coding recurrent events in human tumors beyond TERT [10 , 12 , 16] , would resolve a key question that continues to confound whole cancer genome analyses . Here , through analysis of 221 whole tumor genomes , we first demonstrate the widespread impact of TTCCG-related mutagenesis on the mutational landscape of melanoma . Moreover , through UV exposure of a panel of repair-deficient human cell lines , we rule out inhibited DNA repair as core mechanism . Finally , we generate the highest resolution map of UV-induced cyclobutane pyrimidine dimers ( CPDs ) in the human genome to date , which provides clear evidence that ETS-related promoter hotspots are associated with strong local elevations in the efficacy of UV lesion formation at specific genomic bases .
To assess the impact of TTCCG-related mutagenesis on the landscape of recurrent mutations in melanoma in a more sensitive way than previously possible , we assembled a cohort of 221 melanomas characterized by whole genome sequencing by TCGA and ICGC [20 , 21] . These heavily mutated tumors averaged 110k somatic single nucleotide variants ( SNVs ) per sample , expectedly dominated by C>T transitions and a mutational signature characteristic of mutagenesis by UV light through formation of pyrimidine dimers ( S1 Fig ) . Notably , despite the genome-wide scope , nearly all highly recurrent mutations were found near annotated transcription start sites ( TSSs ) ( Fig 1a ) . For example , of the 22 most recurrent individual bases ( mutated in ≥18 patients ) , four were known drivers ( BRAF , NRAS or TERT promoter mutations ) while the rest were at most 524 bp away from a known TSS . Further , the vast majority of highly recurrent promoter sites were found in conjunction with TTCCG sequences ( Fig 1a and 1b ) , indicating a widespread influence from ETS elements to the mutational landscape of melanoma . Analysis of TTCCG-related recurrent mutations in relation to enhancers further supported that the effect is largely restricted to promoters ( S2 Fig ) . Of 51 recurrent promoter mutations ( +/- 500 bp from TSS ) mutated in ≥ 12 tumors , 42 ( 82% ) had a TTCCG element in the immediate ( +/- 10 bp ) sequence context , rising to 86% after excluding the known TERT C228T and C250T promoter mutations ( Fig 1b and S1 Table ) [1 , 2] . Most were within 200 bp upstream of a known TSS , as expected for functional ETS elements ( Fig 1b and 1c ) [22] . Among the few remaining sites , two ( upstream of AP3D1 and TMEM102 ) were instead flanked by TTCCT sequences likewise matching the ETS recognition motif ( Fig 1b ) [22] and the numbers are thus conservative . The fraction TTCCG-related sites increased as a function of recurrence , from 291/550 promoter sites ( 53% ) at n ≥ 5 to 7/8 ( 88% ) at n ≥ 20 , excluding the known TERT sites ( Fig 1d ) . For comparison , only 0 . 60% of C>T mutations in the dataset exhibited TTCCG patterns , underscoring their massive enrichment in recurrent positions . As noted previously [17] , there was a strong correlation between the number of mutated TTCCG hotspot sites and the total mutational burden in each tumor , compatible with these sites being passive passengers ( Spearman’s r = 0 . 94 , P = 1 . 5e-106; Fig 1e ) . Also confirming earlier observations , the TTCCG-related promoter hotspots were found preferably near highly expressed genes , as expected under a model where interaction with an ETS TF rather than sequence-intrinsic properties are responsible for elevated mutation rates in these sites ( Fig 1f ) . Taken together , these analyses clearly demonstrate that ETS-related mutations account for nearly all highly recurrent non-coding hotspots genome-wide in melanoma , as well as hundreds of less recurrent sites not detectable in previous analyses based on smaller cohorts . Recent studies have shown that NER , the main DNA repair pathway for UV damage , is attenuated in TF binding sites , leading to elevated somatic mutation rates [18 , 19] . While plausible as a mechanism for ETS-related mutation hotspots [16] , we recently showed that TTCCG elements were associated with elevated mutation rates also in cutaneous squamous cell carcinomas ( cSCCs ) lacking global NER ( XPC -/- ) [17] . We also established that mutations can be easily induced in TTCCG hotspot sites in cell culture by UV light , thus recreating in vitro the process leading to recurrent mutations in tumors [17] . We decided to use the RPL13A -116 bp hotspot site , notably more frequently mutated ( 58/221 tumors ) than both canonical TERT sites and on par with BRAF V600E at 60/221 ( Fig 1a and 1b ) , as a model to further investigate a possible role for impaired NER . To this end , we UV-exposed A375 cells with intact NER as well as fibroblasts with homozygous mutations in four key DNA repair components: XPC , required for global NER , ERCC8 ( CSA ) and ERCC6 ( CSB ) , required for transcription coupled NER ( TC-NER ) , and XPA which is required for lesion verification in both global and TC-NER ( S3 Fig ) . Correct genetic identity and complete homozygosity for the mutant allele was confirmed by whole-genome sequencing of all four mutant cell lines ( S2 Table ) . Even limited UV exposure led to high cell mortality in the mutant cell lines , forcing us to limit the exposure to a single low dose of UVB ( 20 J/m2 ) during approximately two seconds followed by three weeks of recovery , after which cells were assessed for RPL13A promoter mutations using error-corrected amplicon sequencing ( Fig 2a ) [23] . Between 7 , 332 and 13 , 774 error-corrected reads at ≥10x oversampling were obtained for each of 10 different libraries ( Fig 2b ) . Strikingly , even at this miniscule dose , subclonal somatic mutations appeared preferably at the known hotspot site in A375 cells ( Fig 2c ) as well as in all of the mutant cell lines ( Fig 2e–2g ) , despite abundant possibilities for UV lesion formation in flanking assayed positions . As expected , absolute mutation frequencies were low , less than 0 . 5% in all samples , bringing us close to the detection limit in some samples as indicated by noise in the untreated controls ( Fig 2d , 2e and 2g ) . In combination with earlier data from tumors lacking global NER [17] and the fact that the mutations are almost exclusively positioned upstream of TSSs where TC-NER should not be active ( Fig 1b and 1c ) , these results argue against impaired global NER as well as TC-NER as the basic mechanism behind TTCCG hotspot formation . It was established decades ago that DNA conformational changes induced by interactions with proteins can alter conditions for UV damage formation [24 , 25] , which prompted us to investigate whether ETS-related promoter hotspots may arise due to locally favorable conditions for UV lesion formation . For this , we adapted a protocol first established in yeast using IonTorrent sequencing [26] to the Illumina platform ( Fig 3a ) , to generate a genome-wide map of CPDs in A375 human melanoma cells immediately following UV exposure , before DNA repair processes have had a chance to act . CPDs were preferably detected at TT , TC , CT and CC dinucleotides as expected ( Fig 3b ) . An elevation at AT dinucleotides was consistent with an earlier report where this was attributed primarily to AT[T/C] sites , suggesting a contribution from CPDs at flanking dipyrimidines [26] . By comparing with median detection frequencies at non-dipyrimidines , we estimated the false positive rate for CPDs at dipyrimidines to vary from 5 . 5% for TT up to 17 . 0% for CC . The number of detected CPDs after removal of PCR duplicates scaled nearly linearly with simulated sequencing depth , indicating favorable random representation of CPDs ( Fig 3c ) . A total of 202 . 1 million CPDs were mapped to dipyrimidines throughout the cellular genome ( Fig 3c ) , constituting the highest resolution CPD map to date to our knowledge . Additionally , 95 . 3 million CPDs were mapped in UV-treated naked ( acellular ) A375 DNA lacking interacting proteins , while a non-UV-treated control , which expectedly yielded limited material , produced 18 . 5 million CPDs ( Fig 3c ) . We next investigated CPD formation patterns at TTCCG mutation hotspot positions identified above in melanoma ( Fig 1a and 1b ) . 291 recurrently mutated ( n ≥ 5/221 melanomas ) TTCCG promoter sites ( +/-500 bp from TSS ) were aligned centered on the mutated base such that CPD density in these regions could be determined . This revealed a striking peak in CPD formation that coincided with the hotspots , which was largely absent in naked DNA lacking bound proteins or in non-UV control DNA ( Fig 4a ) . Additionally , more recurrently mutated sites showed a stronger CPD signal , compatible with increased CPD formation being the key mechanism ( Fig 4b ) . For a more detailed understanding , we subcategorized the 291 melanoma ETS hotspot sites into four main groups based on sequence and mutated position . The strongest mutation hotspots , such as RPL13A and DPH3 , typically occurred at cytosines one or two bases upstream of the TTCCG element ( Fig 1b and S1 Table ) , which notably is outside of the core motif and therefore not expected to disrupt binding [22] . In CCTTCCG sites ( n = 82 unique loci ) , recurrent C>T transitions would typically appear at both 5ʹ cytosines ( underscored ) independently or , less frequently , as CC>TT double nucleotide substitutions . Aggregated CPD density over these sites , centered on the motif , revealed a strong peak bridging these two bases , which notably was absent in naked DNA ( Fig 4c ) . Thus , when the TF site is occupied , CPDs form efficiently between the two pyrimidines , leading to C>T mutations at either base although with a preference for the second position , in agreement with established models for UV mutagenesis [27] . The same pattern of strongly elevated CPD formation in cellular , but not naked , DNA was observed between the same positions in TCTTCCG and CTTTCCG sites ( n = 57 and 27 , respectively ) , with C>T mutations expectedly forming only at the first or second pyrimidine depending on the position of the cytosine ( Fig 4d and 4e ) . Many of the less recurrent bases in melanoma were often found at the first middle cytosine of a TTCCG motif ( S1 Table ) . Interestingly , a large fraction of these sites lacked a dipyrimidine at the two key positions identified above thus prohibiting CPD formation there , with ACTTCCG being the most common pattern ( 44/82 sites ) , which indeed matches the in vivo ETS consensus sequence [22] . Compatible with the mutation data , the strongest CPD peak was observed at the middle TC dinucleotide , and in agreement with the lower mutation recurrence , this signal was weaker compared to the other site types ( Fig 4f ) . Of note , elevated CPD formation between these bases could also be clearly seen in the other site categories ( Fig 4c–4e ) . Taken together , these analyses based on genome-wide CPD mapping provide strong evidence that locally elevated CPD formation efficacy shapes the formation of mutation hotspots at ETS binding sites . Earlier studies have described a general increase in mutation rate in promoter regions , attributed to reduced NER activity at sites of TF binding including ETS sites [18 , 19] . To investigate a possible contribution from increased CPD formation to this pattern , we first determined the overall mutation rate in melanoma near TSSs , which confirmed a sharp increase in upstream regions that coincided with reduced NER as determined by XR-Seq ( Fig 5a and 5b ) [28] . Further confirming earlier data [19] , this increase was abrogated in XPC -/- cSCCs lacking global NER , arguing against a major contribution from increased CPD formation ( Fig 5a ) . Consistent with this , CPDs were found to form at near-expected frequencies when aggregated over these regions ( Fig 5c ) . Interestingly , subtraction of TTCCG-related mutations revealed that these constitute a large proportion of promoter mutations in melanomas , but not in XPC -/- cSCCs , supporting a notable contribution from inhibited NER in ETS sites to the overall burden increase in promoters ( Fig 5a ) . While elevated UV-induced DNA damage is important in the formation of ETS-related recurrent mutation hotspots , we conclude that this effect has negligible impact on the general increase in mutation burden in regulatory regions . This is instead explained by repair inhibition including a prominent contribution from impaired NER in ETS sites , which can likely further add to elevated mutation frequencies at recurrent hotspot positions .
Proper analysis of recurrent non-coding mutations requires an understanding of how mutations arise and distribute across the genome in the absence of selective pressures . Here , we provide a mechanistic explanation for the passive emergence of recurrent mutations at specific positions in TTCCG/ETS sites in tumors in response to UV light , and also demonstrate the massive impact of such mutations on the mutational landscape of melanoma using a large whole genome cohort . Mutations at -116 bp in the RPL13A promoter were used here as a model to study mutation formation at ETS hotspot sites in vitro in repair-deficient cell lines , which ruled out inhibited repair as sole mechanism . Of note , this site is more recurrently mutated than the individual TERT C228T/C250T sites and nearly as frequent as chr7:140453136 mutations ( hg19 ) pertaining to BRAF V600E , thus representing the second most common mutation in melanoma and likely other skin cancers . Notably , mutations were detectable at this site in cultured cells following a UVB dose of 20 J/m2 UVB , equivalent to about 1/200th of the monthly absorbed UVB dose in July in Northern Europe [29] . This underscores the extreme UV sensitivity of ETS hotspots and explains their high recurrence in tumors . Genome-wide mapping of CPDs revealed that TTCCG-related mutation hotspots exhibit highly efficient CPD formation at the two bases immediately 5ʹ of the core TTCC ETS motif . The effect was lost in naked acellular DNA , showing that structural conditions for elevated CPD formation are induced when the TF binding site is in its protein-bound state . Interestingly , most functional ETS sites are expected to lack pyrimidines in the two key positions [22] thus prohibiting pyrimidine dimer formation , and conditions for forming a strong mutation hotspot are therefore only met in a subset of sites with CC , TC or CT preceding the TTCCG element . Additionally , CPDs form at lower but still elevated frequency at the middle TTCCG bases , consistent with weaker recurrence for mutations in these positions . CPD and cancer genomic data are thus in strong agreement , providing a credible mechanism for the formation of ETS-related mutations hotspots in UV-exposed cancers . As demonstrated here , frequent mutations at ETS-site hotspots are expected for purely biochemical reasons in UV-exposed cancers . Consequently , several observations are compatible with passenger roles for these mutations: The most recurrent sites arise at cytosines outside of the core TTCC ETS recognition element [22] where they are not expected to disrupt ETS binding . While mutations in the middle of the motif , common among the less frequent hotspots , should disrupt binding , ETS factors tend to be oncogenes that are activated in cancer [30] , and it can be noted that TERT promoter mutations instead enable ETS binding through formation of TTCC elements [1 , 2] . The mutations tend to arise near highly expressed housekeeping genes rather than cancer-related genes , and the particular set of sites that are mutated varies inconsistently in-between tumors . Moreover , as would be expected in the absence of selection and in contrast to known driver mutations [17] , the number of mutated ETS sites in a tumor is strongly determined by mutational burden . Our results complement a recent study by Mao et al . [31] , which was published during the preparation of this manuscript . This study likewise determined CPD formation patterns in ETS binding sites using whole genome CPD mapping obtaining results that are in full agreement with ours , and additionally proposed a structural basis based on available crystallography data for increased CPD formation at the center TC dinucleotide in the ETS-DNA complex , which was demonstrated to promote CPD formation also in vitro . Thus , data on CPD formation patterns from two independent studies , in combination with our data showing a sharply elevated mutation rate at the RPL13A TTCCG hotspot site in vitro in the absence of NER , support that base-specific elevations in CPD formation efficacy forms the foundation for prominent promoter mutation hotspots in skin cancers . At the same time , inhibited DNA repair explains a general increase in mutation burden in regulatory elements including ETS sites , which could act synergistically to further amplify elevated mutation rates at ETS-related hotspots . Future studies may want to better quantify the relative contributions of these effects , as well as define the exact subset of ETS factors or other proteins that interact with DNA at TTCCG-related mutation hotspot sites .
Whole genome somatic mutation calls from the Australian Melanoma Genome Project ( AMGP ) cohort [20] were downloaded from the International Cancer Genome Consortium’s ( ICGC ) database [32] . These samples were pooled with whole genome mutation calls from The Cancer Genome Atlas ( TCGA ) melanoma cohort [21] called as described previously [10] . Population variants ( dbSNP v138 ) and duplicate samples from the same patient were removed , resulting in a total of 221 tumors . Whole genome sequencing data from 5 XPC -/- cSCCs and matching peritumoral skin was obtained from Zheng et al . , 2014 [5] , and aligned with bwa ( v0 . 7 . 12 ) [33] followed by mutation calling using VarScan 2 ( v2 . 3 ) [34] and subtraction of population variants . Gene annotations from GENCODE [35] v19 were used to define TSS positions , encompassing 20 , 149 and 13 , 307 uniquely mapped coding genes and lncRNAs , respectively , considering the 5ʹ-most annotated transcripts while disregarding non-coding isoforms for coding genes . Processed RNA-seq data was derived from Ashouri et al . , 2016 [36] . Enhancer annotations were derived from ChromHMM segmentation ( Core 15-state model , E6 and E7 regions , representing enhancers and genic enhancers , respectively ) of epigenomic data from foreskin melanocytes ( Roadmap celltype E059 ) [37] . XP12 , GM16094 , GM16095 and GM15893 cells were a kind gift from Dr . Isabella Muyleart , University of Gothenburg . Cells were grown in DMEM + 10% FCS + Penicillin/streptomycin ( GIBCO ) . Cells were subjected to a single low dose UVB ( 20 J/m2 ) and left to recover for three weeks . DNA was extracted with Blood Mini kit ( Qiagen ) . To detect and quantify mutations we applied SiMSen-Seq ( Simple , Multiplexed , PCR-based barcoding of DNA for Sensitive mutation detection using Sequencing ) as described previously [17] . Sequencing was performed on an Illumina MiniSeq instrument in 150 bp single-end mode . Raw FastQ files were subsequently processed as described using Debarcer Version 0 . 3 . 1 ( https://github . com/oicr-gsi/debarcer/tree/master-old ) . For each amplicon , sequence reads containing the barcode were grouped into barcode families . Barcode families with at least 10 reads , where all of the reads were identical ( or ≥ 90% for families with >20 reads ) , were required to compute consensus reads . FastQ files were deposited in the Sequence Read Archive under BioProject ID SRP158874 . A375 cells were grown in DMEM + 10% FCS + Penicillin/streptomycin ( Gibco , Carlsbad , MA ) and were treated with 1000 J/m2 UVC following DNA extraction and DNA from untreated cells was isolated as a control , both in duplicates . Additionally , naked DNA from untreated cells was irradiated with the same dose , to provide an acellular DNA control sample . DNA was extracted with the Blood mini kit ( Qiagen , Hilden , Germany ) . Purified DNA ( 12 μg ) was sheared to 400 bp with a Covaris S220 in microtubes using the standard 400 bp shearing protocol . CPD-seq was modified from Mao , Smerdon [26] to adapt it to Illumina sequencing methods using primers described previously in Clausen et al . , 2015 [38] ( S3 Table ) . Briefly , sheared DNA was size selected with SPRI select beads ( 1 . 2 vol ) ( Life Technologies , Carlsbad , CA ) and the purified product ( approx . 4 μg ) subjected to NEBNext end repair and NEBNext dA-tailing modules ( New England Biolabs ( NEB ) , Ipswich , MA ) . ARC141/142 ( 8 μM ) was then ligated to the sheared and repaired ends O/N with NEBNext Quick Ligation module . DNA was purified with 0 . 8 vol CleanPCR beads and treated with 40 units Terminal Transferase ( TdT , NEB ) and 0 . 1mM dideoxy ATP ( Roche , Rotkreuz , Switzerland ) for 2h at 37 °C . DNA was purified and incubated with 30 units T4 endonuclease V ( NEB ) at 37 °C for 2 h , followed by purification and treatment with 15 units APE1 ( NEB ) at 37 °C for 1 . 5 h . DNA was purified and treated with 1 unit rSAP ( NEB ) 37 °C 1 h followed by deactivation at 65 °C for 15 minutes . DNA was purified , denatured at 95 °C for 5 min , cooled on ice and ligated with the biotin-tagged ARC143/144 ( 0 . 25 μM ) overnight at 16 °C with NEBNext quick ligation module . DNA fragments with the biotin tag were captured with 20 μl Streptavidin Dynabeads ( Invitrogen , Waltham , MA ) and the DNA strand without the biotin label was released with 2 x 40 μl 0 . 15 M NaOH and ethanol precipitated . This single-stranded DNA was resuspended in 14 . 9 μl H2O and used as the template to synthesize double-stranded products using ARC154 ( 0 . 25 μM ) by incubating with Phusion High-Fidelity DNA Polymerase ( Thermo Scientific , Waltham , MA ) at 98 °C for 1 min , 58 °C for 30 s and 72 °C for 1 min . The now double-stranded library was purified and amplified for 15 cycles with ARC49 and ARC78-82 ( 0 . 3 μM each ) to add Illumina barcodes and indexes . Two cellular UV-treated , two no-UV controls and one naked DNA control library were prepared , for a total of five libraries . The libraries were pooled with equal volumes of each of the libraries and sequenced using a NextSeq High Output kit ( Illumina , San Diego , CA ) . The data has been deposited in GEO under accession GSE119249 . FastQ files were aligned pairwise with Bowtie 2 version 2 . 3 . 1 [39] to hg19 , using standard parameters . For the -UV control and +UV cellular DNA samples , replicates were merged with Picard MergeSamFiles version 2 . 18 . 7 ( http://broadinstitute . github . io/picard ) . Duplicate reads were marked with Picard MarkDuplicates version 2 . 18 . 7 [40] with the parameter VALIDATION_STRINGENCY = LENIENT . Further analysis was performed in R with Bioconductor [41] , where CPD positions were extracted as the two bases upstream and on the opposite strand of the first mate in each read pair , removing those that mapped outside of the chromosome boundaries . Only biologically possible CPDs detected at dipyrimidines sites were considered in the CPD counts and downstream analyses . Data from duplicate libraries were pooled to achieve higher coverage , since downstream results were in close agreement when considering these libraries individually . To simulate lower coverage libraries , the bam files were subsampled with samtools view version 0 . 1 . 19-44428cd [42] with the parameter -s at 0 . 25 , 0 . 5 or 0 . 75 , and the subsequent bam files were reanalyzed as described above . For analyses of CPD formation patterns , C>T mutations and repair activity around TSSs , these regions were divided into 20 bp bins in which CPD counts or overlapping XR-seq reads were determined . XR-seq data from wild-type NHF1 skin fibroblasts was obtained from Hu et al . , 2015 [43] , and consisted of normalized read counts in 25 bp strand-specific bins . Background frequencies of dinucleotides and trinucleotides in hg19 were counted with EMBOSS’s fuzznuc [44] , using the parameters -auto T -complement T . Expected mutations were calculated by randomly introducing the same number of mutations as observed in the window based on observed probabilities for C>T mutations at different trinucleotides estimated from the complete mutation dataset . Expected CPDs were calculated in the same way , maintaining the number of CPDs in the observed data , but based instead on genomic dinucleotide counts . | Cancer is caused by somatic mutations that typically occur in protein-coding genes . However , the advent of whole genome sequencing has made it possible to venture beyond protein-coding DNA in search of non-coding mutations with putative cancer driver roles . Indeed , recent studies , in particular in skin cancers , describe individual positions in gene regulatory regions ( promoters ) that are recurrently mutated in many independent patients , suggestive of a contribution to carcinogenesis . In this paper , we show that recurrent promoter mutations arise at these sites due to an exceptional propensity to form UV-induced DNA damage lesions ( pyrimidine dimers ) at specific transcription factor binding sites . The effect is present in cellular DNA but not in naked acellular DNA , meaning that the sites need to be occupied by their transcription factor partners in order to induce favorable conditions for DNA damage formation . This explains an important confounding phenomenon in whole cancer genome analyses , and has implications for the interpretation of recurrent somatic mutation patterns in non-coding DNA . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"cancer",
"genomics",
"sequencing",
"techniques",
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"cancers",
"and",
"neoplasms",
"basic",
"cancer",
"research",
"organic",
"compounds",
"oncology",
"genome",
"sequencing",
"mutation",
"pyrimidines",
"sequence",
"motif",
"analysis",
"dna",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"bioinformatics",
"melanomas",
"chemistry",
"molecular",
"biology",
"somatic",
"mutation",
"biochemistry",
"organic",
"chemistry",
"nucleic",
"acids",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"repair",
"physical",
"sciences",
"genomics",
"genomic",
"medicine"
]
| 2018 | Elevated pyrimidine dimer formation at distinct genomic bases underlies promoter mutation hotspots in UV-exposed cancers |
Energy metabolism is central to cellular biology . Thus , genome-scale models of heterotrophic unicellular species must account appropriately for the utilization of external nutrients to synthesize energy metabolites such as ATP . However , metabolic models designed for flux-balance analysis ( FBA ) may contain thermodynamically impossible energy-generating cycles: without nutrient consumption , these models are still capable of charging energy metabolites ( such as ADP→ATP or NADP+→NADPH ) . Here , we show that energy-generating cycles occur in over 85% of metabolic models without extensive manual curation , such as those contained in the ModelSEED and MetaNetX databases; in contrast , such cycles are rare in the manually curated models of the BiGG database . Energy generating cycles may represent model errors , e . g . , erroneous assumptions on reaction reversibilities . Alternatively , part of the cycle may be thermodynamically feasible in one environment , while the remainder is thermodynamically feasible in another environment; as standard FBA does not account for thermodynamics , combining these into an FBA model allows erroneous energy generation . The presence of energy-generating cycles typically inflates maximal biomass production rates by 25% , and may lead to biases in evolutionary simulations . We present efficient computational methods ( i ) to identify energy generating cycles , using FBA , and ( ii ) to identify minimal sets of model changes that eliminate them , using a variant of the GlobalFit algorithm .
Constraint-based analysis , in particular flux-balance analysis ( FBA ) , is the current state of the art in genome-scale metabolic modeling [1] . Constraint-based modeling assumes a steady state ( i . e . , every internal metabolite that is produced must be consumed at the same rate ) and imposes lower and upper bounds on metabolic fluxes . However , constraint-based analyses typically do not explicitly consider thermodynamics . As a result , the mathematical solution of constraint-based problems is often thermodynamically infeasible [2 , 3] . Specifically , internal cycles ( sometimes called type-III pathways [4] ) , which consist only of internal reactions and do not exchange metabolites with the environment , violate the second law of thermodynamics . The thermodynamic driving forces around a biochemical reaction cycle must add up to zero; hence , there cannot be a flux in a closed cycle [5–7] . These thermodynamically infeasible type-III pathways [4] have to be distinguished from futile cycles ( type-II pathways; Fig 1 ) , which additionally consume cofactors to generate a driving force around the cycle [8 , 9] . Futile cycles are not an artifact of metabolic modeling , but have been experimentally observed [10]; e . g . , some prokaryotes that live in very energy-rich environments need to dissipate energy by converting ATP to ADP [11] . A futile cycle running in reverse would charge energy metabolites such as ATP without an external source of energy ( Fig 1 ) . Accordingly , we classify type-II pathways [12] into two subgroups by taking the directionality of cofactor utilization into account: ( a ) futile cycles , which consume energy and are thus thermodynamically feasible , and ( b ) energy generating cycles ( EGC ) , which charge energy metabolites without a source of energy . While such EGCs are thermodynamically impossible , they can—and , as we show below , do—occur in constraint-based models . Futile cycles will rarely occur in FBA solutions , as they dissipate energy and hence divert metabolic investment away from biomass production . EGCs , in contrast , can have a substantial effect on the predictions of constraint-based analyses , as they generate energy out of nothing that then supports in silico growth . A simple example illustrating a ( hypothetical ) EGC is shown in Fig 2 . Eliminating EGCs is crucial for the correct modeling of energy metabolism , as has been recognized earlier ( see , e . g . , [13–15] ) . While thermodynamically infeasible type-III pathways ( internal cycles ) can be easily removed through a simple post-processing step [5 , 16] , the same strategy cannot be used to suppress EGCs . In principle , EGCs could be excluded from the solution space by systematically assigning sufficiently detailed thermodynamic constraints . Thermodynamics-Based Metabolic Flux Analysis ( TMFA ) [17] , for example , searches for a set of feasible metabolite concentrations such that all reactions proceed in the direction of negative free energy change ( ΔG<0 ) or , equivalently , a ratio of product to substrate concentrations below the reaction’s equilibrium constant , Keq . However , it can be shown mathematically that for any flux distribution without type-III pathways , there exists a distribution of metabolite concentrations such that the flux distribution is thermodynamically feasible , i . e . , all fluxes proceed in the direction of negative free energy change ( see the theorem in [16] ) . This theoretical result respects the fact that metabolite concentrations must have a single value for all reactions they participate in; Supplementary S1 Text shows a small example network that illustrates the inability of ll-COBRA and TMFA to reliably exclude EGCs . The simplest EGC could be established through an ATP energy dissipation reaction ( ATP + H2O → ADP + Pi + H+ ) that is allowed to proceed in the backwards direction . An energy-generating backward flux can be achieved as long as the concentration ratio ( [ATP][H2O] ) / ( [ADP][Pi][H+] ) is smaller than the corresponding equilibrium constant Keq = 2×10-5M-1 . If we treat the concentration of H2O ( 55M ) as constant , this cannot occur within the physiological concentration bounds assumed by Henry et al . [17] , 10-5M and 0 . 02M , showing that TMFA’s metabolite concentration bounds avoid the utilization of at least some EGCs . Note , however , that reactions central to an EGC may have equilibrium constants compatible with the concentration bounds , especially if the total free energy change is spread over several individual reactions . Moreover , the TMFA strategy relies on the availability of equilibrium constants for all reactions central to the EGC . In contrast to TMFA , several alternative thermodynamically informed constraint-based methods only consider chemical potentials , which do not incorporate information on reaction specifics such as the equilibrium constant Keq [7 , 12 , 18 , 19] . For every flux distribution free of type-III pathways , it is possible to find a distribution of chemical potentials such that all fluxes proceed in the direction of chemical potential reduction [16] . This means that potentials capable of driving energy dissipation reactions towards the high-energy metabolite can always be found . Thus , constrained-based methods designed to ensure thermodynamic feasibility based on freely variable chemical potentials do not guarantee the elimination of EGCs . The detection and removal of EGCs is currently not part of established metabolic network reconstruction pipelines [2] . In particular , automatic reconstructions algorithms [20 , 21] currently do not test for EGCs . Sometimes , EGCs are identified in the manual reconstruction process , and parts of the cycles are constrained to zero flux as a makeshift correction [13] . Accordingly , as we demonstrate below , the problem of erroneous free energy generation occurs in a majority of automated and a subset of manual network reconstructions .
EGCs can be identified through a variant of FBA [14] . To efficiently identify the existence of diverse EGCs , we first add a dissipation reaction to the metabolic network for each metabolite used to transmit cellular energy; e . g . , for ATP , the irreversible reaction ATP + H2O → ADP + P + H+ is added . These dissipation reactions close any existing energy-generating cycles , thereby converting them to type-III pathways . Fluxes through any of the dissipation reactions at steady state indicate the generation of energy through the metabolic network . Second , all uptake reactions are constrained to zero . The sum of the fluxes through the energy dissipation reactions is now maximized using FBA . For a model without EGCs , these reactions cannot carry any flux without the uptake of nutrients . We used this approach to identify the presence of EGCs for 14 different energy metabolites ( ATP , CTP , GTP , UTP , ITP , NADH , NADPH , Flavin adenine dinucleotide , Flavin mononucleotide , Ubiquinol-8 , Ubiquinol-8 , 2-Demethylmenaquinol 8 , Acetyl-CoA , L-Glutamate ) and for proton exchange between periplasm and cytosol ( for simplicity counted as a 15th “energy metabolite” below ) ; see Suppl . S1 Table for the corresponding dissipation reactions . We did not require the energy dissipation reactions to be charge-balanced; e . g . , in the reaction NADH → NAD+ + H+ , we omitted the molecule that acts as the acceptor of the two electrons . Adding the electron acceptor to the dissipation reaction would not dissipate the energy stored in NADH , as this energy could then potentially be re-used by internal cofactor regeneration reactions; in this case , the dissipation reaction could be active even in the absence of EGCs . FBA models do not keep track of metabolite charges , and thus the general problem posed by charge unbalanced reactions is not that they affect constraint-based simulations directly; instead , they are a sign of incorrect reaction stoichiometry , which is especially severe in the case of electron imbalances . It is important that models are mass and electron balanced [2] before conducting the EGC analysis . While EGCs induced by mass or electron unbalanced reactions may be detected by our method , they cannot be removed properly without fixing the reaction stoichiometries . We analyzed all models in three large databases of constraint-based metabolic networks: BiGG [22] , ModelSeed [23] , and MetanetX [24] . Overall , we found that over two thirds ( 68% ) of tested models supported a non-zero flux through at least one of the 15 energy dissipation reactions , although this percentage differed drastically between databases ( Fig 3 ) . The BiGG database contains high-quality manual [2] and , in the case of 54 E . coli strains , semi-automated [25] genome scale metabolic reconstructions . We found EGCs in only 3 out of the 79 BiGG models ( 3 . 8%; Suppl . S2 Table ) . The ModelSEED database is connected to a service for high-throughput reconstruction and analysis of metabolic networks . A special feature is the fully automated reconstruction of genome-scale networks from genome sequences . Due to the fully automated reconstruction , models created by this service should be considered as draft models , and manual steps for model improvement are recommended [23] . Consistent with this recommendation , we identified EGCs in 95% of ModelSEED models ( 185 out of 195; Suppl . S2 Table ) . Finally , MetaNetX is a Meta-Database for metabolic network models , gathering metabolic networks from different databases ( including The ModelSEED and the BiGG database ) and mapping them to one common namespace . This allows easy meta-analysis , manipulation , and comparison of those models [24] . Our FBA strategy found EGCs in 66% of MetaNetX models ( 50 out of 76; Suppl . S2 Table ) . For each network with EGCs , we then used a slightly modified version of GlobalFit [26] to suggest a minimal number of reaction removals that eliminate all EGCs , allowing independent removals of forward and backward directions for reversible reactions . GlobalFit was originally designed to reconcile inconsistencies between FBA model predictions and measured growth/non-growth data , e . g . , from gene knockouts . GlobalFit uses a bi-level optimization method to identify the minimal set of network changes needed to correctly predict all experimentally observed growth and non-growth cases ( or a subset thereof ) simultaneously . We slightly altered the original algorithm , now simultaneously contrasting one growth case ( the network with the biomass reaction as the objective function , ensuring that the suggested modifications do not interfere with biomass production ) , and one non-growth case ( the network with the sum of energy dissipation reactions as the objective function , ensuring that the modified network contains no EGCs; for details see Materials and methods ) . It can be argued that some types of reactions should be preferentially removed; e . g . , reactions only weakly supported by genomic evidence may be removed first , and it may be more likely that one direction of a reaction labeled as reversible represents a network error than that an irreversible reaction is erroneous . While the modified GlobalFit algorithm allows such differential weighing of different reaction types , we considered all reaction removals as equally likely in the application detailed below . Moreover , reactions could be preferentially removed depending on the estimated equilibrium constant ( or standard Gibb’s free energy change ΔG0 ) . For 94% of metabolic models with EGCs ( 223 out of 238 ) , GlobalFit found a set of reaction removals that eliminated all EGCs while maintaining the ability to produce biomass . In many cases , GlobalFit suggested the removal of the ATP synthase reaction . While this will indeed remove most ATP-producing cycles , it will also abolish the model’s natural ability to produce ATP through respiration . To avoid this undesired side effect , we performed a second search for reaction removals that eliminated all EGCs , this time forcing the algorithm to retain the ATP synthase reaction . This step could be adapted to the physiology of the studied organism by selecting a different reaction set to be retained . In each case , we could identify an alternative set of reaction removals; below , we only consider these alternative sets of suggested network changes . Note that GlobalFit does not actually remove the offending reactions , but constrains their fluxes to zero . This allows their reactivation in conditions where they are deemed thermodynamically feasible , although alternative measures must then be taken to avoid EGCs . Most erroneous models can be corrected by making up to five originally reversible reactions irreversible ( Fig 4 ) . The removal of irreversible reactions was only rarely suggested by the algorithm ( Fig 4 ) , while the complete removal of reversible reactions was never observed . In the remaining unsolved models , EGCs could in principle be eliminated by adding reactions to the metabolic networks . The addition of reactions not directly connected to an EGC may be needed to restore biomass production in case no solution exists that preserves viability after EGC removal . While the modified version of GlobalFit is capable of suggesting such additions , the application of this strategy would require manual revision , as it might incorrectly add new metabolic capabilities . While bi-level mixed integer optimization algorithms such as the one used by GlobalFit typically require long computation times , GlobalFit resolved most solvable EGCs in under 10s , and all but one EGC within one minute ( Supplementary S1 Fig ) . The only calculation that required over one minute was for the yeastnet 7 . 6 model [27] , for which the CPLEX solver did not find an optimal solution within the set limit of 60 hours on 16 CPUs . The best set of changes found for this yeast model eliminated all EGCs by removing 76 reactions ( or reaction directions ) . According to CPLEX , there is no alternative elimination of EGCs with fewer than 33 reactions; thus , this model contains at least 33 EGCs . As many EGCs include transport reactions across cellular membranes , the large number of EGCs found in this eukaryotic model ( and the resulting increased computation time ) may be caused by the existence of several intracellular compartments and the associated transport processes . Freely available energy may boost biomass production . Accordingly , the elimination of EGCs through the reaction removals suggested by GlobalFit resulted in biomass reductions in 92% of cases ( 206 out of 223 ) , typically by more than 25% ( Fig 5 ) . This indicates that the in-silico biomass yield may be unrealistically high in a majority of automatically generated models . One of the simplest EGCs we identified is displayed in Fig 6 ( A ) . This cycle is contained in only two metabolic models from The ModelSEED database , Klebsiella pneumoniae MGH 78578 ( Seed272620 . 3 ) and Flavobacterium johnsonia johnsoniae UW101 ( Seed376686 . 6 ) . In this EGC , a malate symporter ( rxn10153 ) transports malate together with two protons out of the cell . The exported malate molecule is then re-imported together with a sodium ion via the malate/Na+ symporter ( rxn05207 ) . The sodium is in turn exported by a Na+/Proton antiporter ( rxn05209 ) in exchange for the import of only one of the protons of the first reaction . Thus , the second exported proton from the first reaction is free to drive an ATP-synthase reaction , generating ATP from ADP without access to an external energy source . To eliminate this EGC , the cost of either malate or sodium transport in terms of translocated protons must be corrected . This option was not given to GlobalFit , which instead suggests to remove the export direction of the malate symporter ( rxn10153 ) . In the manually curated model iJO1366 , six reactions ( SPODM , SPODMpp , SUCASPtpp , SUCFUMtpp , SUCMALtpp , and SUCTARTtpp ) were inactivated in the published model to avoid unrealistic energy generating loops by constraining their flux to zero [13] . We could identify two distinct EGCs ( Fig 6B and 6C ) by reactivating these reactions in the iJO1366 model and in the 54 other E . coli models derived from this reconstruction [25] . One of these EGCs is the rather simple cycle ( Fig 6B ) based on tartrate facilitated transport ( TARTRtpp ) , found in 45 of the 55 E . coli strain reconstructions in Ref . [25] . This reaction spontaneously imports tartrate from the periplasm into the cell , while the tartrate/succinate antiporter ( TARTRt7pp ) exports tartrate , but simultaneously imports succinate . The cycle continues with the succinate/aspartate antiporter and then the aspartate/proton symporter , so that eventually a proton gradient between periplasm and cytosol is established . GlobalFit suggests to remove the utilized direction of the tartrate/succinate antiporter . The other EGC found in the unconstrained E . coli models is a more complicated cycle ( Fig 6C ) that occurs in 46 of the 55 E . coli reconstructions [25] , including the manually curated iJO1366 model [13] . A proton gradient across the periplasmic membrane is established by a NADH:menaquinone oxidoreductase ( NADH17pp ) , which translocates protons in the process of transferring electrons from NADH to Menaquinone 8 , driven by a chain of four enzymes , including superoxide dismutase ( SPODM ) . In order to deactivate the cycle , GlobalFit removes the backward direction of the Malate oxidase ( MOX ) or the forward reaction of the Superoxide dismutase ( SPODM ) . In this case , removal of the Malate oxidase would also be suggested by an analysis of standard free energy changes , at it is highly energetically unfavourable . The EGC shown in Fig 6D was found in 99 out of 195 metabolic models from the ModelSEED database [20] . rxn00379 creates Adenosine 5'-phosphosulfate from ATP and sulfate . The sulfate adenyltransferase rxn09240 catalyses the backward reaction ( and has the same EC-number assigned ) , but charges not only an ATP , but additionally a GTP in the process . To eliminate this EGC , GlobalFit suggests removing either one of the participating reactions . EGCs are a major issue in FBA modeling—they are able to produce energy out of thin air , thereby severely affecting the appropriate representation of energy metabolism and of biomass yield . EGCs not only affect the accurate representation of existing metabolic systems . They will be particularly problematic in evolutionary simulations that involve the incorporation of foreign metabolic reactions from other species [28–30] . Such mixing of reactions from disparate model reconstructions may easily introduce EGCs , and may thus lead to erroneous phenotype predictions . We have recently suggested a protocol for evolutionary simulations that avoids this problem [31] . Here , we present an improved computational method for the high-throughput identification of EGCs . EGC identification is currently not a recognized step in model reconstruction , although some authors have eliminated EGCs from their manually curated models before publication . While constraint-based methods may avoid the utilization of EGCs based on thermodynamic considerations [17] , such methods are computationally expensive and require careful analysis of the EGCs and the bounds on metabolite concentrations to guarantee the absence of EGCs from the resulting flux distributions . Instead , we propose to correct the metabolic model itself , and present a modified version of the previously published GlobalFit algorithm to eliminate EGCs through the removal of minimal reaction sets . The resulting model can then be used with the full suite of standard constraint- based methods . We found EGCs in the majority of automatically generated models and in a small subset of manually curated networks . Many of the identified EGCs—in particular those that occurred most frequently—involved the erroneous maintenance of proton gradients across cellular membranes . The simplest EGCs would consist of two reversible reactions that catalyze the same biochemical conversion using different amounts of energy metabolites ( Figs 2 and 6D ) . Such trivial EGCs are easily recognizable and are consequently rarely included in published metabolic networks; most EGCs in published models are more complex , and not easily identified by eye . We note that automatically reconstructed models often contain other types of errors as well [2] . For example , charge and mass imbalanced reactions appear to be common in automatic reconstructions and can lead to erroneous FBA predictions . Such reactions can potentially introduce EGCs , and we thus suggest to correct them as a preprocessing step . The inclusion of reaction sets that are capable of forming an EGC into a metabolic network reconstruction is not necessarily erroneous . It is conceivable that one part of the cycle is thermodynamically feasible in one condition , whereas the other part is thermodynamically feasible in another condition , while both are not thermodynamically feasible simultaneously . Accordingly , modeling algorithms that respect thermodynamic constraints do not utilize potential EGCs [3 , 6 , 7 , 17 , 32–34] . FBA , however , does not consider thermodynamics; instead , optimization of its objective function ( e . g . , biomass production rate ) will usually lead to the exploitation of EGCs . One possible solution would be to constrain the fluxes through thermodynamically impossible sections of EGCs to zero in each simulated environment; this , however , would require a detailed understanding of environment-specific thermodynamics ( or , alternatively , environment-specific gene regulation ) . Our algorithms are suitable to guide a manual curation of draft networks , and should be included in the standard toolbox used for metabolic network reconstruction . GlobalFit can enumerate alternative solutions to eliminate EGCs , which can then be used as a basis for expert curation . In the context of automated network reconstruction pipelines such as ModelSEED or kBase , our methods could be applied without human interaction , albeit at the risk of removing reactions that might be thermodynamically feasible in particular environments .
We started from 350 genome-scale metabolic networks ( GSMs ) that were downloaded from three databases: BiGG [22]–mostly manually created GSMs ( accessed July 2015 ) ; ModelSeed [23]–GSMs created automatically from genome sequences ( accessed July 2015 ) ; and MetaNetX [24]–a meta-database containing metabolic models from various sources ( accessed January 2016 ) . We removed networks that were unable to produce biomass in a maximally rich environment . We checked the correct direction of exchange reactions , and set the lower bound of the ATP maintenance reaction ( ATPM ) to zero , i . e . , we did not require a non-growth-related production of ATP . To each GSM , we added 15 energy dissipation reactions ( Supplementary S1 Table ) , where the namespace for metabolite names had to match the source of the network , i . e . , BiGG , ModelSEED , or MetaNetX . Because not every metabolic network covers the full range of metabolites used in the energy dissipation reactions ( EDR ) , we checked the integration of the reactions in the network , defined as the fraction of the reaction’s metabolites also present in the remainder of the model ( i . e . , a reaction with an integration of 1 is completely integrated , whereas reactions with an integration < 1 cannot carry any flux ) . Because EGCs tend to run with maximal fluxes , all network reactions except the newly added ones ( those in energy dissipation reactions ) are restricted to fluxes in the range [–1 , 1] for reversible and [0 , 1] for irreversible reactions . To establish the presence of EGCs for different energy metabolites , we maximized one energy dissipation reaction flux vd at a time while prohibiting all influx into the model: max ( vd ) subject to: Sv=0∀i∉E:vimin≤vi≤vimax∀i∈E:vi=0 Here , S is the stoichiometric matrix , v the vector of fluxes , d the index of one of the energy dissipation reactions , vmin and vmax the vector of lower and upper reaction bounds , respectively , and E is the set of indices of all exchange reactions . An optimal value vd* for this optimization with vd*> 0 indicates the presence of at least one cycle that is able to generate a specific type of energy metabolite ( corresponding to the index d ) in the network . Because 0 ≤ ∣vi∣ ≤ 1 for all reactions other than dissipation reactions , the value of vd* is a lower bound for the number of non-overlapping EGCs for the tested energy metabolite in the network . Once a GSM was identified to contain at least one EGC , GlobalFit was used to eliminate all EGCs from the network . GlobalFit was developed to find globally minimal sets of model changes that simultaneously reconcile sets of experimental growth and non-growth observations with model predictions; a detailed description of the original GlobalFit algorithm can be found in [26] . We modified GlobalFit for the efficient removal of EGCs as outlined below . In this modified version , the only allowed type of model change is the removal of unidirectional reactions , where reversible reactions are treated as two independent unidirectional reactions . We contrast a single growth with a single non-growth case . The non-growth case reflects the removal of all EGCs: with no nutrient uptake allowed , the maximal sum of fluxes through the energy dissipation reactions must be zero , max ( ∑d∣vd∣ ) = 0 . To ensure that reaction removals do not abolish biomass production by eliminating EGCs , we set up a growth case with a minimal biomass production rate in a rich medium that allows uptake of all nutrients . Formally , we solve the following bi-level optimization problem , which is a variation of the original GlobalFit problem [26] ( Variable definitions are listed in Table 1 ) : minδ ( ∑y∈M ( δyRF+δyRB ) ) ( 1 ) subject to: Sg × vg= 0 ( 2 ) ∀y∈M vymin× ( 1−δyRB ) ≤ vyg≤ vymax× ( 1−δyRF ) ( 3 ) vBiog≥ Tg ( 4 ) Sng × vng = 0 ( 5 ) ∀y∈M vymin× ( 1−δyRB ) ≤ vyng≤ vymax× ( 1−δyRF ) ( 6 ) minvng ( ct× vng ) =0 ( 7 ) δyRB and δyRF are binary variables . Setting one of these variables to 1 will constrain the corresponding flux of the growth—Eq ( 3 ) —and non-growth case—Eq ( 6 ) –to zero . The total number of reaction removals is minimized , where the removal of forward and backward reaction is treated separately in Eq ( 1 ) –i . e . , δyRB and δyRF are independent . Both the growth and the non-growth case must be in steady state , Eqs ( 2 ) and ( 5 ) . The biomass production of the growth case has to be greater than a predefined threshold Tg , Eq ( 4 ) . All entries in ct are 0 , except for the positions of the energy dissipation reactions , which are 1 . The maximal summed flux through all energy dissipation reactions must be zero , Eq ( 7 ) . We convert this bi-level optimization problem into a single level optimization problem as described in [26] . All calculations were run in GNU R with the SyBiL library [35] and a modified GlobalFit library [26] under linux . We used IBM ILOG CPLEX as the solver for the mixed integer linear optimizations . Each calculation was run on 8 CPU cores and 50GB main memory . | Genome-scale metabolic models are routinely used to simulate the growth of unicellular organisms , and are likely to become an important tool in the medical sciences . The most popular method employed for this task is flux balance analysis ( FBA ) , a simplified mathematical description able to describe the simultaneous activity of hundreds of biochemical reactions . Cellular functions are often dependent on the availability of sufficient energy , and thus a correct representation of energy metabolism appears crucial to metabolic modeling . However , we found that the majority of FBA models generated directly from genome sequences , as well as a minority of carefully curated models , are capable of generating energy out of thin air . These models charge energy metabolites such as ATP without any nutrient uptake . We named the corresponding sets of reactions “erroneous energy generating cycles” ( EGCs ) and developed a high-throughput algorithm for their identification . We found EGCs in 238 ( 68% ) of 350 metabolic models from three different databases . We developed a second , fully automated method for EGC removal . Simulations on the corrected models typically showed growth rates that were 25% slower than in the original models , demonstrating the importance of checking metabolic model reconstructions for EGCs . | [
"Abstract",
"Introduction",
"Results",
"and",
"discussions",
"Materials",
"and",
"methods"
]
| [
"protons",
"medicine",
"and",
"health",
"sciences",
"metabolic",
"networks",
"physiological",
"processes",
"metabolites",
"network",
"analysis",
"genome",
"analysis",
"bioenergetics",
"thermodynamics",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"genomics",
"biological",
"databases",
"nucleons",
"energy",
"metabolism",
"free",
"energy",
"physics",
"biochemistry",
"nuclear",
"physics",
"physiology",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"metabolism",
"genomic",
"databases"
]
| 2017 | Erroneous energy-generating cycles in published genome scale metabolic networks: Identification and removal |
Recent studies found that mutations in the human SLC30A10 gene , which encodes a manganese ( Mn ) efflux transporter , are associated with hypermanganesemia with dystonia , polycythemia , and cirrhosis ( HMDPC ) . However , the relationship between Mn metabolism and HMDPC is poorly understood , and no specific treatments are available for this disorder . Here , we generated two zebrafish slc30a10 mutant lines using the CRISPR/Cas9 system . Compared to wild-type animals , mutant adult animals developed significantly higher systemic Mn levels , and Mn accumulated in the brain and liver of mutant embryos in response to exogenous Mn . Interestingly , slc30a10 mutants developed neurological deficits in adulthood , as well as environmental Mn-induced manganism in the embryonic stage; moreover , mutant animals had impaired dopaminergic and GABAergic signaling . Finally , mutant animals developed steatosis , liver fibrosis , and polycythemia accompanied by increased epo expression . This phenotype was rescued partially by EDTA- CaNa2 chelation therapy and iron supplementation . Interestingly , prior to the onset of slc30a10 expression , expressing ATP2C1 ( ATPase secretory pathway Ca2+ transporting 1 ) protected mutant embryos from Mn exposure , suggesting a compensatory role for Atp2c1 in the absence of Slc30a10 . Notably , expressing either wild-type or mutant forms of SLC30A10 was sufficient to inhibit the effect of ATP2C1 in response to Mn challenge in both zebrafish embryos and HeLa cells . These findings suggest that either activating ATP2C1 or restoring the Mn-induced trafficking of ATP2C1 can reduce Mn accumulation , providing a possible target for treating HMDPC .
Manganese ( Mn ) is an essential element required for the catalytic activity of numerous cellular enzymes [1] . The mammalian divalent metal ion influx transporter SLC39A8 has high affinity for Mn [2 , 3] , and mutations are associated with low blood Mn levels and impaired function of Mn-dependent enzymes , including β-1 , 4-galactosyltransferase , thus leading to congenital glycosylation defects and impaired growth in infants [4 , 5] . On the other hand , high levels of Mn are toxic due to increased cellular oxidative stress , reduced mitochondrial function , and cell death [6] . In humans , the major source of Mn is dietary Mn absorbed through the intestinal wall; excess Mn is removed from the body via the hepatobiliary route [7 , 8] . Under chronic conditions of high Mn , Mn either overloads the liver ( causing hepatic dysfunction ) or increases systemic toxicity , inducing an irreversible and incurable parkinsonian-like syndrome in which the vulnerable basal ganglia are affected by accumulated Mn [8–10] . This Mn-induced parkinsonian-like syndrome can occur due to increased environmental exposure or in certain occupational settings [11 , 12] . In addition , high Mn levels due to parenteral nutrition and/or drug addiction can also lead to disease [13 , 14] . Furthermore , patients who have impaired liver function and are unable to adequately excrete Mn are sensitive to high Mn and can develop Mn-induced parkinsonian-like symptoms [15 , 16] . For example , reduced function of the transporter SLC39A14 , which mediates the hepatic uptake of Mn for subsequent biliary excretion , causes cerebral Mn accumulation due to increased systemic Mn levels [17] . Several transporters play an important role in Mn metabolism , including importers such as transferrin , SLC39A8 , and SLC39A14 , and exporters such as ATP13A2 , ATP2C1 , and SLC30A10 [1] . However , how these proteins coordinate in order to maintain Mn homeostasis is poorly understood . Hypermanganesemia with dystonia , polycythemia , and cirrhosis ( HMDPC , OMIM# 613280 ) was the first Mn metabolism disorder to be linked to a genetic defect [18–20] . Specifically , autosomal recessive mutations in the SLC30A10 gene , which encodes a key Mn exporter [21] , have been linked to HMDPC [22–24] . To date , 28 affected individuals in 14 families have been described [25 , 26] . The majority of these 28 patients presented with dystonia ( 25 patients ) or spastic paraparesis ( one patient ) ; the remaining two patients developed late-onset parkinsonism [26] . However , the mechanism by which SLC30A10 contributes to Mn homeostasis , and the underlying pathological processes associated with SLC30A10 mutations , are unknown . Currently , the preferred treatment for HMDPC is chelation therapy with intravenous disodium calcium edetate ( EDTA-CaNa2 ) , a treatment commonly used for environmental Mn intoxication [22 , 27] . Although chelation therapy improves the neurological symptoms , reduces serum Mn accumulation , and prevents further hepatic pathology [18 , 28] , it has the disadvantage of requiring strict monitoring of other essential minerals , including zinc , copper , and selenium [22] . Moreover , intravenous EDTA-CaNa2 and close patient monitoring are generally not available in most regions with inadequate medical resources [26] . Chelation therapy also requires lifelong treatments of at least two 5-day courses per month , and once treatment has begun , reducing and/or stopping treatment can cause the patient’s condition to worsen [29] . In addition to chelation therapy , oral iron supplementation can reduce intestinal Mn absorption by competing for common transporter proteins; however , the patient’s iron levels must be monitored closely in order to prevent iron toxicity [29 , 30] . Therefore , safer , more efficacious treatments and/or specific therapeutic strategies are urgently needed for patients with HMDPC and other Mn-associated conditions . Here , we generated two slc30a10 mutant zebrafish models using the CRISPR/Cas9 system . We found that mutant animals develop a phenotype strikingly similar to the typical clinical symptoms in patients with HMDPC . EDTA-CaNa2 chelation therapy and treatment with ferrous fumarate partially rescued the phenotype , suggesting that these models might serve as a valuable tool for screening new pharmacological approaches . Importantly , we also found that SLC30A10 inhibits the Mn-clearing function of ATP2C1 , revealing a novel mechanism with respect to Mn metabolism and creating new opportunities for treating HMDPC and related conditions .
As a first step toward determining whether zebrafish are a suitable model for studying Mn metabolism , we examined whether Mn is metabolized in the same organs in zebrafish as in humans . Wild-type zebrafish embryos were exposed to various concentrations of MnCl2 for 24 hours at different developmental stages , and their behavior was monitored the following day . We found that embryos are most sensitive to Mn exposure at 5–6 dpf ( days post-fertilization ) ; concentrations lower than 0 . 3 mM had no effect on swimming ability , whereas concentrations higher than 1 mM caused complete immobility in 100% of embryos ( S1A Fig ) . Based on these preliminary data , we consider Mn concentrations lower than 0 . 3 mM and higher than 1 mM to be relatively low and relatively high , respectively . Importantly , embryos exposed to 1 mM MnCl2 developed dark-colored brain and liver tissues ( S1B and S1C Fig ) , suggesting that these two organs are sensitive to Mn exposure in zebrafish just like humans [22 , 23 , 26] . Next , we measured Mn accumulation in the head , middle trunk , and tail of Mn-exposed embryos . We found higher levels of Mn accumulated in the head and middle trunk , and lower levels in the tail ( S1D Fig ) . After transferring Mn-exposed embryos to fresh Mn-free Holt buffer , the color of both the liver and brain returned to baseline ( S1E Fig ) , and Mn levels were reduced in these embryos ( S1F Fig ) , suggesting that the accumulation of Mn in the brain and liver of wild-type embryos is reversible . Next , we examined whether zebrafish models are suitable for studying HMDPC , a genetic condition associated with impaired function of the Mn transporter SLC30A10 . By searching the zebrafish genomic database , we identified the zebrafish slc30a10 gene as an ortholog of the human SLC30A10 gene , with 42% identity and 55% similarity at the protein level ( S2 Fig ) . Because SLC30A10 is highly expressed in the human brain and liver , where it functions as a Mn exporter , we examined the expression pattern of slc30a10 in zebrafish . We found that slc30a10 is expressed in the brain , liver , and yolk syncytial layer ( YSL ) of zebrafish embryos ( Fig 1A ) , with robust expression beginning at approximately 5 dpf ( Fig 1B ) . These similarities between humans and zebrafish with respect to their Mn-sensitive organs and SLC30A10 expression patterns ( i . e . , in the brain and liver ) indicate that slc30a10 mutant zebrafish might serve as a suitable model for HMDPC . To create such a model , we generated two separate slc30a10 knockout zebrafish lines using the CRISPR/Cas9 system . One line contains a 10-bp frameshift deletion at codon 95 , which introduces a premature stop at codon 97 . The other line contains a 1-bp frameshift insertion at codon 96 , which introduces a premature stop at codon 173 ( Fig 1C ) . Both mutant lines have significantly reduced slc30a10 expression ( Fig 1D; S1G Fig ) , and the residual transcripts correspond to the predicted mutant forms , which were confirmed by sequencing ( S1H Fig ) . Because these two lines develop the same phenotype , all subsequent experiments were performed using the line carrying the 10-bp deletion . To confirm that the zebrafish slc30a10 gene encodes a Mn exporter , we measured the levels of several minerals in mutant animals . Severe hypermanganesemia—the primary metabolic abnormality in HMDPC—is associated with subtle perturbations in iron metabolism in some patients [22 , 23] . Moreover , the ZnT ( SLC30 ) family of proteins primarily facilitates the efflux of zinc [31–33] . Therefore , we measured Mn , Fe , and Zn levels in embryonic , larval , and adult zebrafish using ICP-MS . Surprisingly , Mn levels were lower in 1-week-old mutant embryos compared to wild-type embryos; in contrast , both Fe and Zn levels were similar between wild-type and mutant embryos ( Fig 1E ) . Because embryos are nourished exclusively by the yolk in this stage of development , the levels of trace minerals in these embryos reflect their absorption from the yolk . In 3-week-old larvae , on the other hand , we found higher Mn levels in the mutants compared to wild-type animals; as in the embryonic stage , both Fe and Zn were similar between groups ( Fig 1F ) . These data indicate that Mn accumulates rapidly in mutant animals during the feeding stage . Finally , in 4-month-old adults , systemic Mn levels were significantly higher in both male and female mutant animals compared to wild-type animals , and Fe levels were similar between wild-type and mutant animals . Interestingly , Zn levels were higher in mutant animals compared to wild-type animals , but only in male animals ( Fig 1G ) . We measured the levels of several other metals in 4-month-old adults , but found no additional differences between mutants and wild-type animals ( S1 Table ) . Taken together , these results indicate that Slc30a10 functions primarily as a Mn exporter during zebrafish development . We measured slc30a10 expression in wild-type embryos exposed to various concentrations of MnCl2 , ferric ammonium citrate ( FAC ) , and ZnCl2 . Even at a relatively high concentration ( 1 mM ) , none of the minerals tested caused a change in slc30a10 expression . Next , we measured the effect of mineral exposure on slc30a10 expression in heterozygous embryos . Unlike wild-type animals , the heterozygous developed defects when exposed to Mn2+ or Fe3+ at concentrations higher than 500 μM and Zn2+ at concentrations higher than 250 μM . ; therefore , we treated embryos with Mn2+ , Fe3+ , and Zn2+ concentrations below these levels . We found that slc30a10 expression increased in heterozygous with increasing concentrations of Mn2+ ( Fig 1H ) and was increased only with the highest concentration of Fe3+ ( Fig 1I ) ; in contrast , slc30a10 expression was not affected by Zn2+ at any concentration tested ( Fig 1J ) . These data indicate that slc30a10 expression is regulated primarily by Mn2+ in heterozygous embryos . We also found that slc30a10 expression increased in mutant adults following Mn exposure ( S3A Fig ) . In addition , the expression of slc30a10 in mutant adults was higher than in embryos; in contrast , slc30a10 expression was unchanged between the embryonic stage and adulthood in wild-type animals ( S3B Fig ) , suggests that the chronic accumulation of dietary Mn upregulates slc30a10 expression in adult slc30a10 mutants . Neurological disorders , including impaired gait and speech , are the most common clinical features among patients with HMDPC . Similarly , our slc30a10 mutant zebrafish begin to develop neurological defects at approximately 4 months of age . Within the first year , approximately 40% and 20% of the mutant males and females , respectively , are affected; by the second year , these percentages increase to approximately 70% and 40% , respectively . In general , the affected mutant animals are smaller and thinner than their unaffected siblings ( Fig 2A and 2B ) . In addition , our ICP-MS analysis revealed that the affected mutants have higher levels of Mn compared to unaffected siblings ( Fig 2C ) . In addition to morphological changes , affected mutants have a distinct bradykinesia-like swimming pattern ( S1 and S2 Movies ) . Thus , the affected mutants move slowly , less frequently , and tend to stay at the bottom of the tank . To better characterize the swimming pattern of these affected mutants , we used behavior-tracking software , which confirmed that the affected mutants stayed in a relatively limited area of the tank ( Fig 2D ) and swam more slowly with shorter distances ( Fig 2E ) than their unaffected siblings . In addition , reproductive capacity was significantly lower in affected mutants compared to unaffected mutants , leading to an overall decrease in reproduction in the entire group of mutant adults compared with age-matched wild-type animals ( Fig 2F ) . Interestingly , however , no defects in locomotion were observed in the mutant embryos during the hatching stage ( i . e . , within 7 dpf ) . Both wild-type and mutant embryos responded rapidly when touched on the tail ( S3 and S4 Movies ) . We hypothesized that mutant embryos do not exhibit behavioral defects because they have not yet developed excess Mn levels , as shown by our ICP-MS results ( Fig 1E ) ; moreover , during this stage , excess metals are not imported , as they have not begun feeding . In humans , manganism gives rise to a well-defined movement disorder with symptoms resembling Parkinson's disease [34 , 35] . Therefore , we exposed mutant embryos to various concentrations of Mn2+ at specific developmental stages in order to induce manganism . Wild-type embryos have robust slc30a10 expression starting at 5 dpf ( Fig 1B ) , which suggests the start of a Mn-sensitive stage in development . We therefore exposed both wild-type and mutant embryos to various concentrations of Mn2+ at 5 dpf and analyzed the embryos at 6 dpf . At 300 μM , Mn caused a clear color change in the brain and liver of mutant embryos , but had no effect in wild-type embryos ( Fig 2G ) . We then measured Mn levels in both Mn-treated and untreated wild-type and mutant embryos and found more accumulated Mn in mutants compared to wild-type ( Fig 2H ) . Moreover , this accumulation in the mutant embryos persisted after the animals were transferred to Mn-free Holt buffer , suggesting that slc30a10 plays an important role in the clearance of accumulated Mn . Finally , following exposure to relatively low concentrations of Mn2+ ( 150–200 μM ) , mutant embryos—but not wild-type embryos—had impaired locomotion and other defects , including tremor , postural instability , disorientation , and impaired balance ( Fig 2I and 2J; S5 and S6 Movies ) . At higher concentrations of Mn2+ ( 250–300 μM ) , the mutants developed more severe defects , including body distortion , rigidity , bradykinesia , and a lack of escape response ( Fig 2I and 2J; S7–S10 Movies ) . In addition , mutant embryos exposed to 300 μM Mn swam significantly shorter distances than both wild-type embryos and untreated mutant controls ( Fig 2K ) . At even higher doses ( >300 μM ) , Mn was lethal to some of the mutant animals . GABAergic neurons are the primary target affected by Mn toxicity [21] . We therefore measured the expression of GABA transporters , GABA receptors , and the glutamic acid decarboxylases Gad65 and Gad67 in Mn-treated and untreated embryos . Exposure to Mn significantly reduced the expression of the GABA transporter slc32a1 in both wild-type and mutant embryos ( Fig 2L ) ; in contrast , expression of the GABA receptors gabbr1a and gabbr1b was reduced only in the Mn-treated mutants ( Fig 2M and 2N ) . Finally , Gad65/67 protein levels were significantly increased in Mn-treated wild-type embryos , as well as in both untreated and treated mutants ( Fig 2O ) . Taken together , the reduced expression of the GABA transporter and receptors indicates reduced GABA uptake , while increased glutamic acid decarboxylases suggest more GABA production . These results suggest that extracellular GABA could be increased in both mutant embryos and Mn-treated embryos . Impaired dopaminergic signaling has been attributed to the parkinsonism-like effects associated with Mn accumulation in the basal ganglia of both HMDPC patients and workers with occupational Mn exposure [8 , 12 , 23 , 35] . Zebrafish models of Parkinson's disease develop two classic pathological features , namely locomotor deficits and a loss of dopaminergic neurons [36] . Moreover , the escape response in zebrafish embryos requires functional dopaminergic circuits [37 , 38] . We therefore investigated whether the Mn-induced parkinsonism-like phenotype observed in Mn-exposed mutant embryos is due to a defect in dopaminergic signaling by measuring the expression of the dopamine transporter ( dat ) and tyrosine hydroxylase ( th ) in embryos with and without Mn exposure; both dat and th mRNA are commonly measured as molecular markers of dopaminergic neurons in zebrafish [39 , 40] . We found that under basal conditions , both dat and th gene expression patterns were similar between mutant and wild-type embryos . However , following exposure to Mn , mutant embryos had significantly lower levels of both dat and th mRNA ( Fig 2P and 2Q; S3C and S3D Fig ) , suggesting decreased dopaminergic signaling . In addition , Mn-treated mutants—but not wild-type or untreated mutant embryos—had increased apoptosis markers in the brain ( Fig 2R; S3E and S3F Fig ) , consistent with neuronal death specifically in mutant embryos following Mn exposure . Taken together , these data suggest that Mn accumulation leads to impaired GABAergic transmission and loss of dopaminergic neurons , possibly explaining the impaired locomotion observed in mutant zebrafish with Mn accumulation . Patients with HMDPC develop a wide range of liver complications , including steatosis , hepatomegaly , hepatic fibrosis , and—ultimately—cirrhosis [23 , 26] . We therefore examined whether our mutant zebrafish develop a similar hepatic phenotype . Under basal conditions , mutant embryos develop a mild fatty liver , and this effect was increased upon Mn exposure ( Fig 3A ) . HE staining and Oil red O staining of frozen sections confirmed hepatic steatosis in Mn-exposed mutant embryos ( Fig 3B and 3C ) . In addition , Mn exposure caused a dose-dependent darkening of the liver in mutant embryos ( Fig 3D and 3E ) , consistent with hepatic Mn accumulation . At 4 months of age , Sirius red staining revealed hepatic fibrosis in mutant animals ( Fig 3F ) . Moreover , the mRNA levels of the fibrosis biomarkers col1a1a and ctgfa were higher in mutant animals compared to wild-type animals ( Fig 3G and 3H ) . These data indicate that slc30a10 mutant zebrafish develop hepatic pathology similar to the symptoms associated with HMDPC . Polycythemia develops early in all patients with HMDPC and is clinically distinguishable from patients with environmentally induced manganism [22 , 23 , 26] . We therefore measured the relative number of erythrocytes in both wild-type and mutant zebrafish . To visualize the erythrocytes , the animals were crossed to the Tg ( globinLCR:eGFP ) background , in which erythrocytes are labeled with GFP [41]; we then measured erythrocytes using FACS analysis . Our analysis revealed an increase in GFP-positive cells in 3-week-old mutants compared with wild-type animals ( Fig 4A ) , consistent with polycythemia . Increased systemic Mn may drive erythropoiesis by increasing the expression of erythropoietin ( Epo ) , similar to hypoxia-stimulated erythropoietin expression [42] . We therefore measured epo expression in 3-week-old animals and found significantly increased expression in mutant animals ( Fig 4B ) , and increased hepatic epo in 4-month-old mutants ( Fig 4C ) . Interestingly , however , at the 1-week-old embryonic stage , neither the number of erythrocytes nor the expression of epo differed between mutant and wild-type embryos ( Fig 4A and 4B ) . These results suggest that the polycythemia in mutant embryos develops secondary to the chronic accumulation of Mn and does not arise directly from slc30a10 deficiency . Chelation therapy with EDTA-CaNa2 is commonly used to treat manganism [18 , 23] . We therefore tested whether chelation therapy can prevent the Mn-induced phenotype in our mutant zebrafish model by exposing mutant embryos to Mn in the presence or absence of EDTA-CaNa2 . Treating mutant animals with EDTA-CaNa2 restored both locomotion and the escape response ( Fig 4D and S11 Movies ) , and it reduced the Mn-induced hepatic color change ( Fig 4E–4G ) , suggesting that EDTA-CaNa2 chelation is an effective treatment for slc30a10 mutant zebrafish and that this model may be used to screen other chelation therapies . To test whether chelation therapy can reverse the effects of Mn ( thereby mimicking better the therapeutic approach in patients ) , we first exposed mutant embryos to Mn; affected embryos with impaired locomotion and hepatic pathology were then transferred to Mn-free Holt buffer with or without EDTA-CaNa2 ( 2 mg/ml ) . Treatment with EDTA-CaNa2 improved the locomotion of Mn-exposed mutant embryos ( Fig 4H ) , but had no effect on hepatic color change ( S3G Fig ) , indicating that this therapy partially reverses the Mn-induced phenotype in slc30a10 mutant zebrafish . Ferrous fumarate has been used to reduce intestinal Mn absorption and relieve hypermanganesemia in patients with HMDPC [22 , 28 , 29] . Therefore , we exposed mutant embryos to Mn in the presence or absence of ferrous fumarate at 5 dpf and examined their phenotype the following day . Ferrous fumarate treatment caused the gut to turn brown in color , which suggest increased iron uptake in these embryos ( S3H Fig ) . Importantly , however , treating mutant embryos with ferrous fumarate significantly improved locomotion ( S3I Fig ) and significantly reduced the rate of hepatic color change ( S3J Fig ) , consistent with the notion that iron competes with Mn for intestinal import and can partially rescue Mn-induced hypermanganesemia in zebrafish embryos . In preliminary experiments , we found that exposing mutant embryos to 300 μM Mn at 2–3 dpf ( prior to the onset of slc30a10 expression ) had no effect , even when the embryos were exposed for several days; in contrast , the same concentration of Mn is toxic to slightly older ( i . e . , 5–6 dpf ) mutant embryos . We therefore exposed mutant embryos to 300 μM Mn for 72 hours beginning at 3 dpf ( “3–6 dpf mutants” ) or for 24 hours beginning at 5 dpf ( “5–6 dpf mutants” ) ; we then examined the phenotype of both groups at 6 dpf . We found that embryos exposed beginning at 5 dpf had a clear phenotype , including reduced survival rate , locomotion defects and a change in liver color ( Fig 5A–5C; S12 Movies ) . Strikingly , however , embryos there were exposed to Mn for a longer period ( beginning at 3 dpf ) were unaffected ( Fig 5A–5C; S4A–S4C Fig ) . Using ICP-MS , we detected slightly lower Mn levels in 3–6 dpf mutants compared to 5–6 dpf mutants , despite a longer period of Mn exposure ( Fig 5D ) . These data suggest that prior to the onset of slc30a10 expression , another exporter protein may be expressed , thereby preventing the toxic effects of Mn exposure . To test this idea , we measured the expression of several other exporters , including atp2c1 , slc40a1 , and atp13a2 [43] , and examined whether these genes are upregulated in slc30a10 mutants . Our analysis revealed that only one gene—atp2c1—is upregulated in mutant embryos compared to wild-type animals ( Fig 5E ) , suggesting that this gene may compensate for the dysfunction of slc30a10 . We also examined the effect of Mn exposure in 3 dpf wild-type and mutant embryos and found increased expression of atp2c1 ( Fig 5F ) —but not slc40a1 or atp13a2 ( S4D and S4E Fig ) —in both wild-type and mutant embryos , suggesting that atp2c1 is selectively upregulated in response to excess Mn at this early developmental stage . Interestingly , however , atp2c1 expression was unaffected by Mn exposure in either wild-type or mutant embryos at 5 dpf ( Fig 5G; S4F and S4G Fig ) , suggesting the presence of a relatively narrow developmental window in which atp2c1 expression can respond to Mn exposure . Taken together , these data indicate that although the expression of atp2c1 is higher in slc30a10 mutant embryos compared to wild-type embryos under basal Mn conditions , atp2c1 expression responds to Mn exposure only in an early developmental stage , suggesting that Atp2c1 may play a compensatory role as an Mn exporter when slc30a10 expression is low . To test the notion that Atp2c1 can functionally replace Slc30a10 in early development , we measured expression patterns using in situ hybridization and found that atp2c1 is expressed in the brain and abdominal viscera of zebrafish embryos ( S4H Fig ) , which overlaps partially with the expression pattern of slc30a10 ( i . e . , in the brain and liver ) , indicating that atp2c1 and slc30a10 are expressed in similar organs . Next , we used a morpholino to block atp2c1 expression in wild-type embryos and examined the effects of 72 hours of Mn exposure beginning at 3 dpf . Although Mn-treated WT-MO embryos survived ( S4I Fig ) , they had impaired locomotion and developed a dark-colored liver; in contrast , wild-type embryos injected with a control ( scrambled ) morpholino were unaffected by Mn exposure ( S4J and S4K Fig ) . These data suggest that Atp2c1 plays a protective role against Mn toxicity in early zebrafish development . Next , we examined the effect of blocking atp2c1 expression in slc30a10 mutant embryos exposed to Mn for 72 hours beginning at 3 dpf . We found that MO-treated mutant embryos were more sensitive to Mn exposure and had reduced survival and more severely impaired locomotion compared to control mutants injected with a scrambled morpholino ( Fig 5H and 5I ) . In addition , MO-injected mutant embryos had a darker colored liver ( Fig 5J ) and higher levels of Mn compared to control mutants ( Fig 5K ) . Taken together , these data support the notion that Atp2c1 can function as a Mn exporter under systemic Mn stress and can protect slc30a10 mutant embryos from Mn-induced toxicity selectively in early developmental stages . Recently , Mukhopadhyay and Linstedt suggested that increasing the Mn-pumping capacity of ATP2C1 may be a feasible strategy for treating Mn overload in patients [44] . To test this strategy using our zebrafish model , we injected single-cell mutant embryos with human ATP2C1 mRNA containing the gain-of-function Q747A mutation ( ATP2C1-Q747A ) and measured Mn clearance at 3 dpf . Compared with control mutant embryos , ATP2C1-Q747A‒injected embryos had higher Mn resistance and lower Mn levels ( Fig 5L and 5M ) , confirming that overexpressing ATP2C1-Q747A improves Mn clearance in zebrafish embryos early in development . Interestingly , however , ATP2C1-Q747A‒injected mutants were not rescued when exposed to Mn at 5 dpf , possibly due to a degradation of mRNA by this stage . We found that although both 3–6 dpf and 5–6 dpf mutant embryos ( i . e . , mutant embryos exposed to Mn for 72 or 24 hours , respectively ) developed a color change in the liver ( Fig 5C ) , only the 5–6 dpf mutants had decreased survival and impaired locomotion ( Fig 5A and 5B ) . We hypothesized that this difference could be due to a difference in the subcellular distribution of Mn between 3–6 dpf and 5–6 dpf mutants . Due to a lack of adequate antibodies against the zebrafish proteins , we tested this hypothesis in HeLa cells [21 , 44 , 45] . HeLa cells have barely detectable levels of SLC30A10 expression and robust levels of ATP2C1 expression ( S5A Fig ) , which mimics the expression pattern of slc30a10 and atp2c1 in 3 dpf mutant zebrafish embryos ( S5B Fig , 3 dpf ) . In contrast , expressing mutated SLC30A10 in HeLa cells mimics the expression pattern of 5 dpf mutant zebrafish embryos ( S5B Fig , 5 dpf ) . We then used GPP130 , a Mn sensor , to measure intra-Golgi Mn levels [21 , 44] . We found that exposing control-transfected HeLa cells to Mn led to increased GPP130 degradation ( i . e . , increased Mn transport to the Golgi apparatus ) ( Fig 6A ) ; in contrast , Mn transport to the Golgi apparatus was significantly decreased in HeLa cells transfected with the mutant SLC30A10 ( Fig 6B and 6C ) . Thus , by association we conclude that the subcellular distribution of Mn differs between 3–6 dpf mutants and 5–6 dpf mutants , possibly explaining their striking difference in phenotypes . Next , we measured the subcellular localization of ATP2C1 in HeLa cells transfected with wild-type or mutant SLC30A10 . We found that ATP2C1 was localized to the Golgi apparatus and throughout the cytoplasm under control conditions; however , following Mn treatment , the majority of ATP2C1 protein was localized to the Golgi apparatus ( Fig 6A and 6D; S5C Fig ) . Wild-type SLC30A10 was located at the cell membrane ( S5D Fig ) , whereas the mutant SLC30A10 was localized to the endoplasmic reticulum ( S5D Fig ) . However , despite their differences in subcellular localization , both wild-type and mutant SLC30A10 disrupted the Mn-induced trafficking of ATP2C1 to the Golgi apparatus ( Fig 6E–6G ) . Next , we measured the levels of ATP2C1 protein in HeLa cells . We found that Mn exposure increased ATP2C1 expression in untransfected cells ( S5F and S5G Fig ) , similar to our results in zebrafish embryos ( Fig 5F ) . However , Mn exposure had no effect on ATP2C1 levels in HeLa cells transfected with wild-type SLC30A10 and slightly decreased ATP2C1 levels in HeLa cells transfected with mutant SLC30A10 ( S5F and S5H Fig ) , similar to the expression pattern measured in zebrafish embryos at 5 dpf ( Fig 5G ) . Taken together , these data indicate that ATP2C1 expression increases in response to Mn only when SLC30A10 expression is low . Lastly , we measured Mn levels in HeLa cells following Mn exposure . We found that cells expressing mutant SLC30A10 accumulated significantly more Mn compared to control-transfected cells and cells expressing wild-type SLC30A10 ( Fig 6H ) .
Here , we report the generation and characterization of two slc30a10 mutant zebrafish lines . Our analysis shows that these mutant animals develop a phenotype strikingly similar to HMDPC in patients with a mutation in the SLC30A10 gene , showing that the functional effects of these mutations are conserved between humans and zebrafish . In wild-type zebrafish embryos , slc30a10 is expressed in the brain and liver , and its expression is sensitive to Mn exposure , confirming that this transporter protein functions as a Mn exporter in vertebrates . Bakthavatsalam et al . recently used wild-type zebrafish embryos to study manganism [46]; however , their goal was to model locomotor defects induced by environmental and occupational exposure to high concentrations of Mn . Using a mutant zebrafish model , we found that slc30a10 mutants have increased Mn sensitivity in the brain and liver . This approach enabled us to induce a phenotype using a relatively low concentration of Mn , thereby avoiding systemic Mn‒induced toxicity in wild-type animals . The manganism-like phenotype was reversed in wild-type embryos by removing the embryos from Mn-containing medium; however , Mn-exposed slc30a10 mutant embryos were only partially rescued , thereby enabling us to investigate the underlying mechanisms . Recently , Tuschl et al . reported that slc39a14 mutant zebrafish are sensitive to Mn exposure and develop abnormal activity [17]; however , unlike the human ortholog , zebrafish slc39a14 is not expressed in the liver [17] , suggesting that slc39a14 functions differently between zebrafish and humans . Although both slc39a14 and slc30a10 mutants accumulate Mn , only our slc30a10 mutants have morphological and developmental defects , suggesting that slc30a10 has a distinct function in regulating Mn homeostasis during zebrafish development . Interestingly , Hutchens et al . recently reported that Slc30a10-knockout mice have elevated systemic Mn levels [47] . It is important to note , however , that the severe hypothyroidism induced by Mn accumulation in Slc30a10-knockout mice differs from the symptoms present in patients with HMDPC . Therefore , we believe that our slc30a10 mutant zebrafish , which fully recapitulate the symptoms of HMDPC in patients , may serve as a more suitable model . It is interesting to note that in the embryonic stage , the mutant zebrafish have less Mn compared to wild-type embryos . During the pre-feeding embryonic stage , the embryo absorbs nutrients—including metals—primarily from the yolk via a process mediated by the yolk syncytial layer ( YSL ) , an extra-embryonic syncytium that functions throughout early development to hydrolyze yolk materials and transport these materials from the yolk to the embryo’s cells and tissues [48 , 49] . Blocking the transport of nutrients from the YSL to the embryo causes nutritional deficiency . For example , Slc40a1 ( ferroportin ) is an iron exporter expressed in the YSL , and slc40a1 mutant zebrafish embryos develop hypochromic anemia and iron deficiency [50] . Similar to slc40a1 , slc30a10 is also expressed in the YSL during early development ( see Fig 1A ) , and mutating slc30a10 could block the absorption of Mn from the yolk to the embryo , thus causing a Mn-deficient phenotype in mutant embryos . However , we found increased Mn accumulation in mutant animals during development from embryo to adult , suggesting that Slc30a10 has an extremely high affinity for Mn . The increase of Zn was comparable between the wild-type and mutants . Interestingly , Zn levels were slightly higher in mutant males than in wild-type males . Based on previous reports [33 , 51–53] , it is possible that Slc30a10 might have an affinity for Zn in males . With respect to Fe , Slc30a10-knockout mice have increased Fe levels in the brain and blood in early developmental stages [47]; in contrast , we found no difference in Fe levels between wild-type and mutant zebrafish . Up until 1 week of age , the slc30a10 mutant embryos showed no behavioral deficits; however , upon exposure to relatively low concentrations of Mn ( i . e . , 150–300 μM ) , these animals develop a series of locomotor deficits , indicating that Mn readily accumulates in mutant embryos . Moreover , similar to patients with HMDPC , mutant zebrafish develop locomotor deficits and impaired liver function without environmental Mn exposure; these effects begin to manifest at 4 months of age and are likely due to the accumulation of normal dietary Mn , which is similar to patients [26] . Given that some mutants developed a neurological phenotype earlier than others , and given that affected mutants have higher Mn levels than their unaffected siblings , this variability in phenotypes may be due largely to individual differences in Mn absorption . In the brain , the primary target for Mn is the basal ganglia , in which a tight balance between inhibitory and excitatory pathways maintains normal function of the network [46] . The majority of neurons in the basal ganglia are GABAergic; however , studies of the effect of Mn on the GABAergic system have yielded contradictory results; some groups reported increased GABA levels , whereas others reported decreased GABA levels [54] . Here , we found increased protein levels of GAD65/67 in mutant zebrafish; thus , increased extracellular GABA might reflect increased neuronal inhibition and impaired dopamine release . In addition , dopamine synthesis at dopaminergic terminals requires the enzyme tyrosine hydroxylase ( TH ) , and Mn accumulation in dopaminergic terminals is mediated by the dopamine transporter ( DAT ) [55] . We found that the expression of both dat and th is more sensitive to Mn exposure in slc30a10 mutants than in wild-type animals , suggesting that impaired dopaminergic transmission may underlie the locomotion defects in Mn-exposed mutants . A member of the P-type family of ATPases , ATP2C1 pumps Ca2+ and Mn2+ ions from the cytosol into the Golgi apparatus and therefore plays an important role in regulating cellular ion homeostasis [56 , 57] . Knocking down the expression of atp2c1 in zebrafish embryos reduced the embryo’s tolerance to Mn exposure in early development . In addition , we found that ATP2C1 is localized throughout the cytoplasm in HeLa cells under basal conditions , whereas ATP2C1 is localized primarily in the Golgi apparatus upon Mn exposure . Taken together , these in vivo and in vitro data provide compelling evidence that ATP2C1 can function as a Mn exporter . The increased expression of atp2c1 in slc30a10 mutant zebrafish suggests a compensatory role between these two genes . In addition , we found significantly higher levels of atp2c1 expression in the brain and liver of female mutants compared to male mutants ( Fig 5E ) , suggesting that female mutants may be more tolerant of Mn . Interestingly , the neurological phenotype is less prevalent in female mutants than in age-matched male mutants; however , the mechanisms that underlie this gender difference in Mn sensitivity are currently unknown . Leyva-Illades et al . previously reported that HeLa cells expressing wild-type SLC30A10 have reduced Mn in the Golgi apparatus , as Mn efflux through SLC30A10 is both rapid and efficient [21] . Moreover , using DT40 cells ( a chicken B cell line ) , Nishito et al . found that expressing SLC30A10 either alone or together with ATP2C1 induces similar levels of Mn resistance , suggesting that these two proteins are functionally redundant [58] . Our in vivo and in vitro results obtained with zebrafish embryos and HeLa cells , respectively , support the notion that ATP2C1 and SLC30A10 are functionally redundant with respect to exporting Mn from the cytoplasm; in addition , our results indicate that the effects of mutating SLC30A10 are mediated by regulating the expression of ATP2C1 . Moreover , the subcellular redistribution of ATP2C1 upon Mn exposure is blocked by expressing either wild-type or mutant SLC30A10 , and expressing mutant SLC30A10 leads to the accumulation of Mn in the cytoplasm ( Fig 7 ) . Interestingly , the mutant form of SLC30A10 ( with both the first and second exons deleted ) caused more accumulation of Mn in HeLa cells than the truncated form ( with a premature stop codon at position 308 ) ( Fig 6H ) , possibly due to the specific loss of the conserved amino acid that mediates Mn transport [23 , 45] . Taken together , these results suggest that upregulating ATP2C1 and/or restoring the Mn-responsive trafficking of ATP2C1 may be a viable strategy for increasing Mn clearance in cells in which SLC30A10 is dysfunctional , thereby providing new therapeutic options for patients with HMDPC . Nevertheless , precisely how SLC30A10 affects the Mn response and alters the trafficking of ATP2C1 is unknown and warrants investigation . Patients with HMDPC develop impaired neurological function and liver damage , which can become life-threatening conditions . Therefore , given that our slc30a10 mutant zebrafish develop impaired locomotion and liver damage due to systemic Mn accumulation following exposure to a relatively low dose of Mn , they represent a robust , reproducible model for studying these clinically severe complications . Because their skin is highly permeable , allowing the rapid absorption of compounds from the medium , zebrafish larvae are widely used for screening drugs . Using our slc30a10 mutant embryos , we tested the therapeutic potential of EDTA-CaNa2 chelation and supplementation with ferrous fumarate , both of which are currently used in a clinical setting . Although EDTA-CaNa2 chelation had no effect on liver damage , it improved locomotor function , indicating that these zebrafish models are suitable for screening new chelating agents . Furthermore , ferrous fumarate supplementation partially alleviated the Mn-induced phenotype in our models . In addition , although chelators and ferrous fumarate are currently used for treating HMDPC patients , both treatments have side effects; EDTA chelation can induce changes in several other metals , including manganese , zinc , copper , and selenium , whereas ferrous fumarate can induce iron toxicity . Therefore , future research should focus on identifying chemical chaperones and/or other small compounds that specifically target Mn metabolism at the molecular level; in this respect , a drug that restores the Mn-responsive trafficking of ATP2C1 in cells with mutant SLC30A10 expression would be particularly relevant . In conclusion , we developed and functionally characterized two independent lines of slc30a10 mutant zebrafish and found that these models recapitulate the symptoms of patients with HMDPC , thereby providing an invaluable model for studying this genetic disorder . Mechanistically , we found that ATP2C1 plays a compensatory role in the context of SLC30A10 dysfunction , thereby maintaining Mn metabolism . These findings shed light on our understanding of the role that the Mn exporter SLC30A10 plays in the pathophysiology of HMDPC , and they provide experimental evidence for targeting ATP2C1 in managing HMDPC and related disorders .
Zebrafish embryos were exposed to various concentrations of MnCl2 or ZnCl2 , and their phenotypes and survival were recorded at different time points , depending on the specific experimental aims . Unless otherwise specified , the embryos were exposed to the indicated MnCl2 solution for 24 hours starting at 5 dpf ( days post-fertilization ) . Total mRNA was isolated from the embryos and tissues using Trizol ( Invitrogen ) . First-strand cDNA was synthesized using the PrimeScript RT reagent kit and Reverse Transcriptase M-MLV ( M1701 , Promega ) . Real-time PCR was performed using the two-step quantitative RT-PCR method with a kit ( RR037A , Takara Bio ) , and target gene expression was normalized to β-actin mRNA levels . Whole-mount in situ hybridization was performed as described previously [59] . Mutant zebrafish lines were generated using previously reported methods [60] . In brief , a target site ( GCTCTGCTTCTCCATCAGCATGG ) in exon 1 of the slc30a10 gene was selected , and the guide RNA ( gRNA ) template was amplified from the pMD-gata5-gRNA scaffold vector . In vitro transcription was performed using 1 μg of template DNA and T7 RNA polymerase . Zebrafish codon-optimized Cas9 plasmid [61] was linearized with XbaI , and Cas9-capped mRNA was transcribed using the mMESSAGE mMACHINE T7 kit ( Ambion ) . The size and quality of the capped mRNA and gRNA were confirmed by electrophoresis through a 2% ( w/v ) agarose gel . Subsequently , 100 pg of gRNA and 400 pg of Cas9 mRNA were injected directly into single-cell embryos . The following primer pairs and the restriction enzyme BstXI were used to assess the efficiency of genetic disruption: forward 5’-ACTCCTTCAACATGCTGTCGGACA-3’ , reverse 5’-AGCACCTGTTTCCACCTGCT-3’ . The injected embryos were raised to adulthood . F0 fish were crossed with wild-type fish to generate heterozygous F1 progeny , which were then genotyped using BstXI digestion and DNA sequencing . Heterozygous F1 fish were crossed in order to generate homozygous mutant fish . The mutant and wild-type animals were obtained from heterozygous crosses . At least 1000 1-week-old embryos or 50 3-week-old larvae were collected in order to obtain the required amount of tissue ( 100 mg ) for analysis . Single adult zebrafish were measured separately . At least three groups were measured for each stage . HeLa cells were seeded in 6-well dishes and transfected the next day . Twenty-four hours after transfection , the cells were exposed to 500 μM MnCl2 in serum-containing growth medium for 12 hours . After Mn exposure , the cells were washed with phosphate-buffered saline ( PBS ) and incubated in 500 μl trypsin-EDTA at 37°C for 3 min . The trypsin was then neutralized by adding 1 ml of serum-containing growth medium , and the cells were collected by centrifugation for 3 min in a microfuge . The supernatants were discarded , and the cell pellets were washed twice in 1 ml PBS containing 10 mM EDTA , followed by centrifugation as described above . After a final wash in PBS ( without EDTA ) , the cell pellets were re-suspended in 200 μl RIPA buffer in order to extract the proteins; 1–2 μl of protein extract was used to measure protein concentration , and the remainder was used to measure the concentration of Mn . All samples were transferred to acid-washed cans with 5 ml metal-free HNO3 and 1 ml HF , then digested overnight at room temperature; the following day , 1 ml HClO4 was added , and the samples were incubated at 180°C for 10 hours . The cans were then transferred to 120°C and shaken for 2 hours in order to evaporate the residual HF and HClO4 . Acid-washed vials were prepared by incubation in 20% HNO3 for 1 day , followed by extensive washing in ultrapure water . The digested samples were then diluted to 50 ml with Milli-Q water in the acid-washed vials and measured using an Agilent 7500 series inductively coupled plasma mass spectrometer . Whole embryos were lysed in modified RIPA buffer containing 150 mM sodium chloride , 50 mM Tris ( pH 7 . 4 ) , 1% Triton X-100 , 1% sodium deoxycholate , 1% SDS , and proteasome inhibitor with PMSF , then centrifuged at 4°C for 15 minutes . Equal amounts of protein in each sample were resolved by SDS-PAGE on a 10% Mini-PROTEAN TGX Precast Gel ( Bio-Rad ) and transferred to an Immuno-Blot PVDF membrane ( Bio-Rad ) . The membrane was blocked with 5% ( w/v ) non-fat milk and incubated with anti-GAD65/67 ( ab11070 , Abcam ) and anti-actin ( R1207-1 , Huabio ) antibodies . After washing three times for 10 minutes each with Tris-buffered saline containing Tween-20 , the membranes were incubated with HRP-conjugated secondary antibodies for 1 hour at room temperature . Finally , the membranes were incubated in chemiluminescence reagent ( 34080 , Pierce ) and exposed to X-ray film . Embryos were incubated in 6-well plates , and their morphological changes and behavioral patterns were observed and recorded using a Nikon SMZ18 stereo microscope . For measuring the escape response , a soft tip was used to touch the tail of embryo , and locomotion was recorded . For behavioral analysis , a digital video tracking system running the ANY-maze software program was used . The experiments were performed in a temperature-controlled room between 1:00 pm and 5:00 pm . The fish were allowed to acclimate to the system for 1 h before the start of data acquisition . Each adult fish was placed in a 20 cm × 10 cm tank containing 1 L of water , and the total distance traveled was recorded for 30 s . Each embryo was then placed in a well of a 24-well plate containing 2 ml Holt buffer with or without Mn , and the total distance traveled was recorded for 10 min . Apoptosis were determined using the in situ Cell Death Detection Kit with TMR red ( 12156792910 , Roche ) in accordance with the manufacturer’s instructions . The fluorescence signal was analyzed using ImageJ software . After fixation with 4% paraformaldehyde in PBS overnight , 6 dpf embryos were washed with PBS and incubated in a 30% sucrose solution overnight . The embryos were then incubated in a 1:1 mixture of 30% ( w/v ) sucrose and OCT compound ( Tissue-Tek ) for 2 hours , followed by overnight incubation in OCT . The samples were frozen in OCT compound at -80°C , and 10-μm cryo-sections were cut using a CM1950 cryostat ( Leica ) . Cells were fixed with 4% paraformaldehyde in PBS for 10 minutes , washed three times with PBS , and then permeabilized in 0 . 2% Triton X-100 in PBS for 5 minutes . After washing three times with PBS and blocking with PBS containing 5% fetal bovine serum ( FBS ) , the cells were immunostained with rabbit anti-flag ( 2368s , Cell Signaling ) , rabbit anti-giantin ( ab80864 , Abcam ) , mouse anti-pan-Cadherin ( ab22744 , Abcam ) , or mouse anti-atp2c1 ( WH0027032M1 , Sigma ) in PBS with 1% FBS at 4°C overnight . The cells were then washed three times with PBS and incubated with mouse anti-calnexin Alexa Fluor 647 ( ab202572 , Abcam ) , goat anti-rabbit Alexa Fluor 488 ( A11008 , Molecular Probes ) , donkey anti-mouse Alexa Fluor 488 ( ab150105 , Abcam ) , donkey anti-rabbit Alexa Fluor 555 ( A31572 , Molecular probes ) , and/or goat anti-rabbit Alexa Fluor 350 ( A0408 , Beyotime ) in PBS with 1% FBS at 37°C for 1 . 5 hours . The slides were mounted under coverslips using ProLong Gold Antifade reagent ( Thermo Fisher Scientific ) , and fluorescence images were obtained using an LSM-710 microscope ( Zeiss , Germany ) equipped with a 63X/1 . 4 numerical aperture oil-immersion objective . HeLa cells were grown in Dulbecco's modified Eagle's medium ( DMEM , Gibco ) supplemented with 10% FBS ( Gibco ) and 1X penicillin-streptomycin ( Gibco ) . The cells were cultured at 37°C in humidified air containing 5% CO2 . The cells were transfected using the X-tremeGENE HP DNA Transfection Reagent ( Roche ) in accordance with the manufacturer’s instructions . The cells were generally transfected 24 hours after plating and used 36 hours after transfection . Where indicated , freshly prepared MnCl2 solution was added to the medium at a final concentration of 300 μM for the indicated times . Embryos were fixed in 4% formaldehyde overnight at 4°C , washed with PBS , and stained with filtered 0 . 3% oil red O in 60% 2-propanol for 2 hours . After a final wash in PBS , the embryos were placed in depression wells and photographed . One-week-old and 3-week-old larvae on the Tg ( globinLCR:eGFP ) background were crushed in Hank's buffered salt solution ( HBSS , Invitrogen ) , followed by complete dissociation to a single-cell suspension by trituration . The cells were washed and harvested in HBSS and then examined using a Cytomics FC 500 MCL flow cytometer ( Beckman Coulter , Inc . ) . The full-length wild-type human ATP2C1 cDNA was cloned using PCR , and the Q747A mutation was introduced using two-step site-directed mutagenesis . The cDNA sequence containing the mutation was inserted into the pCS2 expression plasmid , linearized , and transcribed using the mMESSAGE mMACHINE SP6 kit ( Ambion ) . The resulting mRNA was injected into one-cell stage zebrafish embryos at 200 pg/embryo . The atp2c1 morpholino ( 5’-TGCATCTTCCGTCAGCTTTCGTTGC-3’ ) or a scrambled sequence was injected into one-cell stage embryos at 0 . 1 nmol/embryo . A final concentration of 2 mg/ml EDTA-CaNa2 was prepared and either mixed with MnCl2 ( at a final concentration of 250–300 μM ) or used alone . Saturated ferrous fumarate was added to 5 ml 200 μM MnCl2 to assess rescue efficiency . For controls , treated embryos were washed several times with fresh Holt buffer and then incubated in Holt buffer . All summary data are presented as the mean ± the standard error of the mean ( SEM ) . Differences between two groups were analyzed using the two-tailed Student’s t-test , and multiple groups and time points were analyzed using a two-way ANOVA . Differences with a p-value <0 . 05 were considered significant . Approval for the study was granted by the Institutional Animal Care and Use Committee of the Laboratory Animal Center , Zhejiang University . We followed the relevant guidelines from the Laboratory Animal Center of Zhejiang University . We anaesthetized or sacrificed the zebrafish by freezing them in ice before performing the experiments . All relevant data are provided within the paper and the Supporting Information . | Impaired function of the manganese transporter SLC30A10 has been implicated in HMDPC ( hypermanganesemia with dystonia , polycythemia , and cirrhosis ) , an early-onset metabolic disorder clinically characterized by increased systemic Mn levels , neurological impairment , polycythemia , and hepatic injury . No specific treatment is currently available for HMDPC . Moreover , the mechanisms that underlie Mn metabolism are poorly understood , thereby hindering the development of effective treatments . To investigate the physiological processes underlying Mn metabolism and to develop new disease models of HMDPC , we generated two zebrafish slc30a10 mutant lines using the CRISPR/Cas9 system and found that these mutants develop clinical deficits typically associated with HMDPC . Furthermore , we identified a putative compensatory role for ATP2C1 in the absence of SLC30A10 with respect to modulating Mn metabolism . These findings provide a valuable tool for investigating the role of manganese dysregulation in neurological degenerative diseases and which can be used to develop new pharmacological approaches for managing Mn accumulation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"chemical",
"bonding",
"medicine",
"and",
"health",
"sciences",
"manganese",
"hela",
"cells",
"biological",
"cultures",
"cell",
"processes",
"vertebrates",
"animals",
"biological",
"locomotion",
"animal",
"models",
"osteichthyes",
"organisms",
"developmental",
"biology",
"model",
"organisms",
"experimental",
"organism",
"systems",
"cell",
"cultures",
"golgi",
"apparatus",
"embryos",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"physical",
"chemistry",
"embryology",
"fishes",
"chemistry",
"cell",
"lines",
"zebrafish",
"chelation",
"cell",
"biology",
"secretory",
"pathway",
"phenotypes",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"cultured",
"tumor",
"cells",
"chemical",
"elements"
]
| 2017 | Zebrafish slc30a10 deficiency revealed a novel compensatory mechanism of Atp2c1 in maintaining manganese homeostasis |
For more than a century , the origin of metazoan animals has been debated . One aspect of this debate has been centered on what the hypothetical “urmetazoon” bauplan might have been . The morphologically most simply organized metazoan animal , the placozoan Trichoplax adhaerens , resembles an intriguing model for one of several “urmetazoon” hypotheses: the placula hypothesis . Clear support for a basal position of Placozoa would aid in resolving several key issues of metazoan-specific inventions ( including , for example , head–foot axis , symmetry , and coelom ) and would determine a root for unraveling their evolution . Unfortunately , the phylogenetic relationships at the base of Metazoa have been controversial because of conflicting phylogenetic scenarios generated while addressing the question . Here , we analyze the sum of morphological evidence , the secondary structure of mitochondrial ribosomal genes , and molecular sequence data from mitochondrial and nuclear genes that amass over 9 , 400 phylogenetically informative characters from 24 to 73 taxa . Together with mitochondrial DNA genome structure and sequence analyses and Hox-like gene expression patterns , these data ( 1 ) provide evidence that Placozoa are basal relative to all other diploblast phyla and ( 2 ) spark a modernized “urmetazoon” hypothesis .
Mainly because of misinterpretation of life cycle stages between Trichoplax adhaerens and the hydrozoan Eleutheria dichotoma , Placozoa lost their predominant role as the key model system for studying the origin of metazoan life [5 , 17] . This outcome was nourished by molecular studies based on a variety of character sources , which created a series of conflicting phylogenetic scenarios in which most often Porifera came out basal [18–24] . Figure 1 shows six plausible scenarios for the relationships of five taxonomic groups ( Bilateria , Cnidaria , Ctenophora , Porifera , and Placozoa ) and two plausible arrangements for four taxa when Placozoa are left out that are critical in assessing the early relationships of metazoans . For five taxa and one outgroup , there are 105 ways to arrange these taxa in dichotomous branching trees . Nearly 95% of these possible trees can be eliminated as not plausible based on existing data . All six of the hypotheses in Figure 1 have been suggested as viable in the literature over the past two decades ( see Table S1 for a summary of papers in the last decade addressing the phylogenetics of these taxa ) . All six hypotheses have been suggested in publications in the last year alone . For instance , Srivastava et al . ( 2008 ) [23] hypothesize Placozoa as the sister group to both Cnidaria and Bilateria , with sponges branching off earlier ( arrow b in Figure 1 ) . Another recent study , which suggests a basal position for Ctenophora and Anthozoa ( arrow E in Figure 1 ) , unfortunately does not add to the issue , since it does not include Placozoa in the analysis [25] . However , this study does suggest that Cnidaria are not sister to Bilateria , but rather to Porifera [25] . A study that does include Placozoa [26] also suggests that Bilateria and Placozoa are basal metazoans ( arrow a in Figure 1 ) . Striking examples of the diversity of hypotheses generated on these taxa are recent analyses of mitochondrial genome sequence data [27–29] that place Bilateria as sister to all non-Bilateria , with Placozoa as the most basal diploblast ( arrow e in Figure 1 ) . In the following , we use the term “diploblasts” for all nonbilaterian metazoans; we do not intend to contribute to the discussion of whether diploblastic animals may have a mesoderm , however [1 , 30–33] .
Given that both nonphylogenetic interpretation of morphological data as well as molecular analyses of sequence data have failed to resolve the issue , a more comprehensive , systematic analysis of morphological data and new molecular markers are now a requisite for identifying the root of the metazoan tree of life . To approach this goal , we conducted concatenated analyses for 24 metazoan taxa from all of the major organismal lineages in this part of the tree of life that included morphological characters ( 17 characters ) , both mitochondrial and nuclear ribosomal gene sequences ( five gene partitions for 6 , 111 nucleotide positions ) and molecular morphology [8] ( ten characters ) , as well as nuclear coding genes ( 16 gene partitions derived from our database searches and another 18 gene partitions derived from the Dunn et al . ( 2008 ) study [25]; see Materials and Methods ) for 8 , 307 amino acid positions and protein coding genes ( 16 gene partitions for 3 , 004 amino acid characters ) to resolve phylogenetic relationships between recent diploblast groups . The total number of characters included was 17 , 664 from 51 partitions , giving 7 , 822 phylogenetically informative characters . We also constructed a matrix with a larger number of taxa based on the Dunn et al . ( 2008 ) [25] study with 73 taxa for the same gene partitions ( see Materials and Methods and Tables S2 and S4 ) . This matrix had 17 , 637 total characters and 9 , 421 phylogenetically informative characters . In addition , Hox gene expression was compared for a placozoan and a cnidarian bauplan to test predictions from the placula hypothesis [5] . Parsimony , likelihood ( with morphological characters removed ) , and mixed Bayesian analysis of the smaller concatenated matrix using a variety of approaches , weighting schemes , and models is generally consistent with the view that Bilateria and diploblasts ( Porifera , Ctenophora , Placozoa , and Cnidaria ) are sister groups . In addition , Placozoa are robustly observed as the most basal diploblast group ( Figure 2 and Figure 3 ) . Figure 3 shows the support for several hypotheses of monophyly obtained from diverse methods of analysis . Porifera , Bilateria , and Fungi all form strong monophyletic groups ( Figure 3 ) . The four cnidarian classes ( Anthozoa , Hydrozoa , Scyphozoa , and Cubozoa ) together with the Ctenophora form a monophyletic group , the “Coelenterata . ” Within the Cnidaria , the generally accepted basal position of the anthozoans is also recovered by this analysis [34 , 35] . Both choanoflagellates and Placozoa are strongly excluded from a Porifera–Coelenterata monophyletic group . The basal position of Placozoa is also strongly supported by comparing the phylogeny in Figure 2 with hypotheses that place it more derived , using the statistical approach of Shimodaira and Hasegawa [36 , 37] . This battery of tests ( Table 1 ) demonstrates that the basal position of the Placozoa is significantly better than other hypotheses . The 95% confidence tree includes the Maximum Likelihood ( ML ) and Bayesian trees ( both with Placozoa as basal in the diploblasts ) with a cumulative expected likelihood weight ( ELW ) of 0 . 960763 . The tree topology shown in Figure 2 summarizes the best supported phylogenetic hypothesis obtained by using Maximum Parsimony , ML , and Bayesian analyses of the concatenated dataset . Analysis of the larger matrix ( Figure S2 ) was less well resolved within the Bilateria , but showed the same general topology as the smaller analysis . Specifically , Bilateria are monophyletic and sister to the diploblasts , with the choanoflagellate Monosiga basal to these taxa with high jackknife values and Bayesian posteriors . Diploblasts are also monophyletic , and Placozoa are the most basal taxon in the diploblasts . In addition , within the diploblasts , Porifera and Coelenterata are monophyletic , and within Bilateria , Ecdysozoa and Deuterostomia are monophyletic; all groupings with high node support . The topology within the diploblasts is also robust when Bilateria are removed from the analysis . The full analysis seemingly misplaces the Bilateria clade as the sister to all diploblasts . The classical position of the Bilateria is in a highly derived position from within the diploblasts and usually sister to the Cnidaria . The seemingly “weird” prediction of a basal Bilateria from the present analysis has been observed before in other studies ( see Table S1 ) . Several studies have addressed phylogenetic problems specific to this region of the tree of life and have suggested that this region of the tree will be inherently difficult to resolve . These studies suggest that the compression of splitting events in this region renders the resolution of these nodes with high support difficult , if not impossible [38–42] . These studies have suggested that even large amounts of data might not resolve the problem . Other studies have pointed to taxon sampling and modeling as a potential problem in resolving this part of the tree of life [25 , 38–40] . Another problem is that the large number of molecular phylogenetic approaches creates multiple and possibly the most short-lived hypotheses in biology . The large repertoire of algorithms , models , and assumptions sometimes produces a forest of trees from the same dataset ( cf . [43] ) . Thus , tree-building procedures are highly crucial and deserve particular attention if this region of the tree of life is to be resolved [38] . Our analyses provide strong evidence for a basal position of Placozoa relative to other diploblasts , and thus agrees with the mitochondrial genome data analyses ( as indicated by arrow f in Figure 1; [27 , 28] ) . It is therefore important to examine whether the mitochondrial signal swamps out the nuclear data , to rule out the possibility that the topology we present in Figure 2 is biased by mitochondrial information . Figure S1 addresses this problem and demonstrates that nuclear information contributes positive support to 16 of the 21 nodes in the tree . Mitochondrial information contributes positive support to only 15 out of 21 nodes . In addition , examination of the amount of hidden support contributed by nuclear versus mitochondrial data ( not shown ) shows that the majority of the hidden support comes from nuclear information . Both of these results using partitioned support measures indicate that the addition of nuclear data does not conflict with mitochondrial information and is indeed contributing positively to the overall phylogenetic hypotheses Although the hypothesis in Figure 2 is in conflict with a recent analysis of coding genes from whole genomes [23] as well as is in conflict with other studies ( Table S1 ) , the scenario presented here is consistent with another set of studies and also with one of the major urmetazoon hypotheses , the placula hypothesis ( Figure 4 ) . This hypothesis fuels intriguing scenarios for the mechanisms and direction of anagenetic evolution in Metazoa , and in the form presented here , it can illustrate the derivation of Cnidaria and Bilateria from a placozoan-like ancestor . A basal position of Placozoa relative to Cnidaria , and diploblasts sister to Bilateria are cum grano salis consistent with several recent molecular phylogenetic analyses ( [23 , 27] and this study ) encouraging us to reconsider the placula hypothesis in a modern light . The comparison of Hox/ParaHox-like gene expression patterns in Placozoa and Cnidaria creates a new working hypothesis for the origin of the entoderm , a main body axis , and symmetry . Based on the undisputed evidence that Placozoa are basal relative at least to Cnidaria , the Trox-2 gene is likely ancestral to Hox/ParaHox-like genes from Cnidaria ( as formerly suggested [44 , 45] ) . Trox-2 is expressed at the gastrodermis/epidermis ( lower/upper epithelium ) boundary in Trichoplax [46] . Strikingly , we found similar expression patterns for two putative Trox-2 descendents in the hydrozoan Eleutheria dichotoma ( Figure 4 ) . These regulatory gene expression data mirror directly the beginning and ending stage of a modern interpretation of the placula hypothesis . The latter explains the origin of a symmetric bauplan with one or two defined body axes and an internal feeding cavity from a simple placuloid ( proto-placozoan–like ) bauplan that lacked all of the former characteristics . In the most parsimonious scenario , the expression of a single regulatory gene defines polarity in Placozoa , i . e . , the differentiation of a lower versus upper epithelium . According to the proposed “new placula hypothesis , ” the nonsymmetric placozoan bauplan transforms into a symmetric Cnidaria ( or also Bilateria ) bauplan by the former ring of epithelia boundary separation transforming into the new “oral” region of the derived symmetric bauplan ( Figure 4 ) . This transformation is simply the result of a placula lifting up its feeding epithelium in order to form an external feeding cavity , keeping function and morphology of the epithelium unchanged . In the final stage , the “oral” pole develops specialized organs , such as a mouth and tentacles for feeding ( cf . [47] ) . The latter could be driven by duplication of the regulatory gene , which originally defined polarity in the placula ( Figure 4; cf . [48] for review ) . Observations on extant Placozoa and Cnidaria mirror this scenario almost perfectly ( Figure 4 ) . Although prediction and observation match nicely , one has to note , however , that no gene or even gene family , no matter how important , can provide more than just indirect support for a working hypothesis on a hypothetical animal bauplan that can never be observed . It is important to note that multiple topologies can be consistent with the placula hypothesis and that the form of the extant earliest-branching lineage does not necessarily have to represent the form of the ancestor; we consider the latter , however , the more parsimonious alternative . We also point out that the regulatory gene family mentioned here , Hox/ParaHox-like genes , seems to be absent in sponges [49] . A secondary loss of Hox/ParaHox-like genes in sponges seems plausible , and the work by Peterson and Sperling , 2007 [50] provides some evidence for this assumption . Whether a possible loss of a Hox/ParaHox gene might be related to the reduction of epithelial organization in Porifera [3] remains an interesting speculation . The Hox/ParaHox loss scenario in sponges is just one of several crucial questions raised by the phylogeny in Figure 2 . According to this phylogeny , diploblasts and Bilateria both may have started from a placula-like bauplan as suggested in Figure 4 ( “new placula hypothesis” ) . The shown new placula hypothesis illustrates a potential transition from a nonsymmetric , axis-lacking placula into a radial symmetric and head–foot axis organized cnidarian . In a similar way , the placula could also be transformed into a Bilateria bauplan , i . e . , a bilaterally symmetric bauplan with an anterior–posterior body axis . One of the easiest models for adopting a bilateral symmetry suggests that the “urbilaterian” kept the benthic lifestyle of the placula but adopted directional movement . The latter almost automatically leads to an anterior–posterior and ventral–dorsal differentiation . The pole moving forward develops a head and becomes anterior , the body side facing the ground carries the mouth and thus by definition becomes ventral . According to the above scenario , the main body axes of diploblastic animals and Bilateria were independent inventions . Whereas an independent evolution of body axes in diploblastic animals and Bilateria seems easily plausible , the independent evolution of other characters ( e . g . , the nervous system; see below ) seems less plausible given our knowledge of the development and morphology of these characters . We will never observe the hypothetical placula , but we may draw some conclusions from Placozoa , which seem to have retained many of the characteristics of the placula if our interpretation is valid . This scenario draws into question several aspects of animal evolution that will require reinterpretation if this hypothesis is correct . Most notable of these aspects is the evolution of the nervous system , which in the hypothesis in Figure 2 , can only be explained by convergent evolution of Cnidaria and Bilateria nervous system organization . According to the placula hypothesis , we suggest that the placula already had the genetic capability and basic building blocks to build a nervous system , and that from here , the final build-up of the nervous system developed via independent , but parallel , pathways in diploblasts and Bilateria . The genome of the placozoan Trichoplax adhaerens indeed delivers some notable evidence that the genetic inventory may precede morphological manifestation of organs [23] . For example , the placozoan genome harbors representatives of all major genes that are involved in neurogenesis in higher animals , whereas placozoans show not the slightest morphological hint of nerve or sensory cells . Quite noteworthy , however , is that placozoans are quite capable of stimuli reception and perception used to coordinate behavioral responses . In this light , the generally accepted unlikely convergent evolution of a nervous system only looks unlikely from a morphological , but not from a genetic and physiological , point of view . Regardless of the need for reinterpretation of this and other anatomical characters , the findings presented here provide a viable hypothesis for the major cladogenetic events during the metazoan radiation . Given the basal position of Placozoa , we suggest that at least for diploblastic metazoan life , the body plan started with the following: an asymmetric body plan , a most simple morphology ( only two steps above basic definition [51] ) , a single ProtoHox gene , a large mitochondrial ( mtDNA ) genome , an outer feeding epithelium that gave rise to the entoderm , and the smallest of all known ( not secondarily reduced ) metazoan genomes . If the placula is also the ancestral state for metazoans ( i . e . , the common ancestor of Bilateria and diploblasts in Figure 2 ) , then the same could be said for the urmetazoon .
In order to extend the analyses of Rokas et al . [42] to basal metazoans also , we isolated 13 of the suggested target genes that were missing from the placozoan Trichoplax adhaerens . These genes could be amplified by using the primer sets that had worked in the previous study in sponges: TOA04 , 05 , 06 , 09 , 10 , 11 , 13 , 15 , 16 , 17 , 21 , 25 , 33 , 48 , 53 , 56 , 57 , 59 , 62 , 65 , 67 , and 68 . In order to obtain sequences of these genes for Placozoa and to characterize variation within Placozoa , we also isolated six of these genes from a second , distantly related placozoan species ( Placozoa sp . H2 , TunB clone , Tunisia ) . For cubozoans , we filled gaps in the matrix by isolating three target genes from Carybdea marsupialis ( Table S5 ) . We amplified target genes from cDNA . For both placozoan species , some 200 healthy growing vegetative animals of each species were used for the isolation of total RNA . Before extraction , animals were washed three times with sterile 3 . 5% artificial seawater ( ASW ) and starved overnight to prevent algae contamination . Animals were lysed in 500 μl of fresh homogenization buffer ( HOM: 50 mM Tris HCl , 10 mM EDTA , 100 mM NaCl , 2 . 5 mM DTT , 0 . 5% SDS , 0 . 1% DEPC in ultrapure water [Gibco]; pH 8 . 0 ) . After addition of 25 μg of DEPC-treated Proteinase K , samples were stored for 30 min at 65 °C . The homogenate was squeezed through a needle connected to a 2 . 5-ml syringe . This protocol significantly increased RNA yield compared to conventional RNA extraction kits . Nucleic acids were isolated by two rounds of phenol/chloroform/isoamyl alcohol ( 25:24:1 ) purification . Nucleic acids were dissolved in ultrapure water , and DNA was digested with DNase I ( Fermentas ) . Total RNA was used for cDNA transcription with poly-T primers following the manufacturer's protocol ( Invitrogen Superscript II Kit ) . Target genes were amplified after initial denaturation ( 3 min at 94 °C ) by 40 rounds of 94 °C for 30 s , 50 °C for 30 s , and 72 °C for 75 s , followed by a final elongation step ( 5 min at 72 °C ) using the Bioline Taq system following the manufacturer's recommendations ( Bioline ) . Amplified fragments of the predicted size were purified and cloned into pGEM-T ( Promega ) . Sequencing was performed on a Megabase 500 using the DYEnamic ET Terminator Cycle Sequencing Kit ( Amersham ) or by using the service provided by Macrogen . For further details , see Jakob et al . [46] and Table S5 . For a detailed explanation of the inclusion of sequences in the phylogenetic matrices used in this study , see Table S2 , which shows the source of sequences in this study . We constructed two matrices , a small one composed of 24 taxa ( see Figure 2 ) and a large one composed of 73 taxa . For the smaller matrix , we chose nine bilaterian taxa based on the availability of sequence information for a species . We chose three Lophotrochozoa , three Ecdysozoa , and three Deuterostomia as representatives of the Bilateria . Other ingroup taxa include representatives of the four classes of Cnidaria , the three major groups of Porifera ( Desmospongiae , Calcarea , and Hexactinellida ) , Placozoa , and Ctenophora . Since rooting of the tree is critical , we attempted to break up the root by including several outgroup species: two fungal species ( Saccharomyces and Cryptococcus ) , Tetrahymena , Trypanosoma , and Dictyostelium based on their relevance to the study and the availability of genome-level information . Trypanosoma was used as outgroup species in all aspects of the study , but the topology of resultant trees indicates that slime mold or Tetrahymena could also be used . To increase the number of placozoan and cubozoan sequences , we PCR amplified several genes as indicated in Table S5 . Morphological characters were scored for the taxa in this study as described in Schierwater and DeSalle ( 2007 [10]; see Table S3 ) . Molecular “morphology” characters were also included for the taxa in this study as scored by Ender and Schierwater , 2003 [8] ( see Figure S3 ) . The final partitioned matrices for the smaller ( 24 taxa ) and the larger ( 73 taxa ) can be found in Table S4 . In addition to genes already available from whole mitochondrial sequencing ( 15 genes ) and nuclear genes ( 16 genes ) , we included 18 genes from the Dunn et al . ( 2008 ) study [25] . These genes were chosen on the basis of taxonomic representation being over 50% in the Dunn et al . ( 2008 ) study . For the larger 73-taxon matrix , we included all of the taxa from the Dunn et al . ( 2008 ) study ( their smaller matrix in their Figure 2; [25] ) plus Cubozoa , Scyphozoa , Placozoa , Hexactinellida , Calcarea , Caenorhabditis , Tetrahymena , Trypanosoma , and Dictyostelium . For this larger matrix , we filled in character information for these taxa for the 18 Dunn et al . ( 2008 ) [25] genes from GenBank as completely as possible . We used Blast scores and existing annotations as criteria for assessing orthology for these added sequences . In this larger matrix , we used only genes from the Dunn et al . ( 2008 ) study [25] with greater than 50% taxon representation . RNA in situ hybridization studies were performed as described before [46 , 52] . For immunocytology studies , polyclonal antibodies were produced to oligopeptides near the C-terminal of the Trox-2 , Cnox-1 , and Cnox-3 proteins . For whole-mount analysis , live animals were fixed for 1 h in 5% formaldehyde in sterile seawater . Immunocytochemistry was performed with anti-Trox or anti-Cnox , respectively , antisera and goat anti-rabbit-AP ( Novagen ) or FITC-conjugated goat anti-rabbit antibody ( Sigma ) . Localization of antibody complexes was revealed by staining with NBT and X-phosphate ( Roche ) or fluorescent microscopy , respectively . Further details will be described elsewhere ( S . Sagasser et al . unpublished data ) . To generate static alignments , we used MAFFT [53] , initially with a gap opening penalty of 1 . 5 and gap extension penalty of 0 . 123 . We also examined the impact of varying gap opening penalties by obtaining alignments using opening penalties of 1 . 0 , 0 . 5 , and 0 . 1 . The alteration of gap penalty only served to alter the number of characters in our matrices and did not severely impact phylogenetic hypotheses . For our 24-taxon matrix , we conducted parsimony , Bayesian , and likelihood analyses as explained below . The 73-taxon matrix was analyzed with Bayesian inference . Phylogenetic trees using static alignment were generated using PAUP v4b10 [54] . Tree searches were accomplished using 1 , 000 random taxon additions and Tree Bisection Reconnection ( TBR ) . Jackknife measures for node support were obtained using PAUP with 30% character removal and 1 , 000 repetitions . To examine the effect of character weighting in phylogenetic analysis of this dataset , we implemented character weighting for nucleic acids and amino acid partitions as follows . First , we implemented three schemes for weighting transitions and transversions ( 100 , 10 , and 2 ) for nucleic acids . Second , we used four transformation matrices for amino acid weighting: Gonnet [55] , WAG [56] , LG [57] , and Genetic Identity ( GI ) . Bremer support measures ( decay indices ) [58] , partitioned Bremer and hidden support values [59 , 60] were generated using TreeRot v3 [61] . The parallel implementation of MrBayes v3 . 1 . 2 [62 , 63] was used for Bayesian inference of phylogeny . Two simultaneous runs with random starting trees were launched for two million generations , each with a 1 , 000-step thinning , a 10% burn-in , and a temperature parameter of 0 . 2 so as to lead to better mixing . All three data types ( DNA , protein , and morphology ) were accommodated in the Bayesian analysis . We employed ML inference in RAxML v7 . 0 . 4 [64] using the GTR substitution model for DNA [65 , 66] along with G-distributed rate heterogeneity [67 , 68] and the Whelan and Goldman ( WAG ) amino acid substitution matrix [55] with empirical residue frequencies coupled with G-distributed rate heterogeneity . Node support was evaluated with 1 , 000 rapid bootstrap replicates [69] . Alternative phylogenetic hypotheses were compared using the Shimodaira-Hasegawa test [37] and expected likelihood weights [70] , as implemented in RAxML . | Following one of the basic principles in evolutionary biology that complex life forms derive from more primitive ancestors , it has long been believed that the higher animals , the Bilateria , arose from simpler ( diploblastic ) organisms such as the cnidarians ( corals , polyps , and jellyfishes ) . A large number of studies , using different datasets and different methods , have tried to determine the most ancestral animal group as well as the ancestor of the higher animals . Here , we use “total evidence” analysis , which incorporates all available data ( including morphology , genome , and gene expression data ) and come to a surprising conclusion . The Bilateria and Cnidaria ( together with the other diploblastic animals ) are in fact sister groups: that is , they evolved in parallel from a very simple common ancestor . We conclude that the higher animals ( Bilateria ) and lower animals ( diploblasts ) , probably separated very early , at the very beginning of metazoan animal evolution and independently evolved their complex body plans , including body axes , nervous system , sensory organs , and other characteristics . The striking similarities in several complex characters ( such as the eyes ) resulted from both lineages using the same basic genetic tool kit , which was already present in the common ancestor . The study identifies Placozoa as the most basal diploblast group and thus a living fossil genome that nicely demonstrates , not only that complex genetic tool kits arise before morphological complexity , but also that these kits may form similar morphological structures in parallel . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"molecular",
"biology",
"evolutionary",
"biology"
]
| 2009 | Concatenated Analysis Sheds Light on Early Metazoan Evolution and Fuels a Modern “Urmetazoon” Hypothesis |
Implantation of a blastocyst in the uterus is a multistep process tightly controlled by an intricate regulatory network of interconnected ovarian , uterine , and embryonic factors . Bone morphogenetic protein ( BMP ) ligands and receptors are expressed in the uterus of pregnant mice , and BMP2 has been shown to be a key regulator of implantation . In this study , we investigated the roles of the BMP type 1 receptor , activin-like kinase 2 ( ALK2 ) , during mouse pregnancy by producing mice carrying a conditional ablation of Alk2 in the uterus ( Alk2 cKO mice ) . In the absence of ALK2 , embryos demonstrate delayed invasion into the uterine epithelium and stroma , and upon implantation , stromal cells fail to undergo uterine decidualization , resulting in sterility . Mechanistically , microarray analysis revealed that CCAAT/enhancer-binding protein β ( Cebpb ) expression is suppressed during decidualization in Alk2 cKO females . These findings and the similar phenotypes of Cebpb cKO and Alk2 cKO mice lead to the hypothesis that BMPs act upstream of CEBPB in the stroma to regulate decidualization . To test this hypothesis , we knocked down ALK2 in human uterine stromal cells ( hESC ) and discovered that ablation of ALK2 alters hESC decidualization and suppresses CEBPB mRNA and protein levels . Chromatin immunoprecipitation ( ChIP ) analysis of decidualizing hESC confirmed that BMP signaling proteins , SMAD1/5 , directly regulate expression of CEBPB by binding a distinct regulatory sequence in the 3′ UTR of this gene; CEBPB , in turn , regulates the expression of progesterone receptor ( PGR ) . Our work clarifies the conserved mechanisms through which BMPs regulate peri-implantation in rodents and primates and , for the first time , uncovers a linear pathway of BMP signaling through ALK2 to regulate CEBPB and , subsequently , PGR during decidualization .
The tunica mucosa of the uterus , called the endometrium , must undergo significant changes to become receptive to implantation of the blastocyst . The endometrium consists of glandular and luminal epithelium and stroma , and when the endometrium is receptive , the embryos can attach to the endometrial epithelium and invade into the stromal compartment . Stromal cells respond to the invasion of the embryo with a wave of proliferation followed by differentiation; this morphological and functional transformation is called decidualization [1] . These steps are fundamental to the implantation process and are dependent upon the action of ovarian progesterone ( P4 ) signaling through its cognate receptor ( reviewed in [2] ) . Several functions have been attributed to the decidua: 1 ) it provides the growing embryo with growth factors and cytokines; 2 ) it regulates the local immune response at the feto-maternal interface; 3 ) it maintains tissue homeostasis during trophoblast invasion; 4 ) it protects the blastocyst from inflammation and reactive oxygen species; and 5 ) it supports the angiogenic processes necessary to create new vessels for the perfusion and nourishment of the embryo ( reviewed in [3] ) . The development of the embryo and the cycling of the uterus must be synchronized for implantation to occur . This synchrony requires complex cell-specific crosstalk , and although many factors have been shown to be involved in implantation , it is still unclear how these factors function and interact . One group of signaling proteins that is expressed in the uterus during early pregnancy is the bone morphogenic protein ( BMP ) subfamily of the transforming growth factor β ( TGFβ ) superfamily . Unraveling the signaling processes regulated by BMPs during embryo implantation is important for the understanding of endometrial health . BMP2 was identified as one of the factors that failed to exhibit induction in the decidual uterus treated with the PR antagonist , RU486 [4] . Because BMP2 is essential for embryonic development of the mouse , the function of BMP2 in the postnatal uterus was initially unknown [5] . Conditional genetic deletion of Bmp2 in the uterus was facilitated by the generation of a mouse model that expresses cre recombinase under control of the progesterone receptor ( Pgr ) promoter [6] . This technology has been largely used to study the functions of multiple genes in the mouse uterus ( [7] ) . Mice lacking BMP2 in the uterus ( Pgrcre/+ Bmp2flox/flox ) are sterile due to the inability to undergo decidualization [8] . Furthermore , deletion of Bmp2 in the uterus causes the deregulation of downstream targets like the WNT signaling pathway , PGR signaling , and induction of Ptgs2 . Roles of BMP2 in the decidual reaction have been confirmed in humans [9] . Because BMP2 is one of the master regulators of implantation [8] , and other BMPs are expressed in the uterus during the peri-implantation period [10] , [11] , it is important to determine the receptors that transduce BMP signaling in the regulation of endometrial function . BMP ligands bind heterodimers of type 1 and type 2 serine/threonine kinase receptors . Upon ligand binding , the type 2 receptor phosphorylates and activates the type 1 receptor , which leads to phosphorylation of receptor-regulated SMAD proteins ( R-SMADs; i . e . , SMAD1 , 5 and 8 ) . Activated SMADs form a complex with the common SMAD ( C-SMAD ) , SMAD4 , and accumulate in the nucleus to regulate the expression of genes involved in cell growth , cell differentiation , apoptosis , cellular homeostasis , and other cellular functions . BMP ligands bind different receptors in different contexts: the physiological association with a specific receptor depends on both the binding affinity and the actual availability of the ligand and the receptor in a specific environment . While the type 2 receptors are essential for recognizing the ligands , the type 1 receptors determine the specificity of the intracellular response . Three type 1 receptors , activin-like kinase 2 ( ALK2; ACVR1 ) , activin-like kinase 3 ( ALK3; BMPR1A ) , and activin-like kinase 6 ( ALK6; BMPR1B ) , are known to mediate BMP signaling , and all three type 1 receptors are expressed in the pregnant uterus ( data not shown ) . Our previous studies showed that ALK6 was not required for decidualization [12] . In this paper , we demonstrate essential roles of ALK2 in the regulation of endometrial function and female fertility .
ALK2 is required during embryonic development , and Alk2 null mice die at ∼E7 . 5 because of their inability to complete gastrulation [13] , [14] . To study the physiological roles of Alk2 in female reproduction , we generated a conditional mouse model using Pgr-Cre and an allele of Alk2 in which exon 7 is flanked by loxP sites ( floxed allele ) [15] ( Figure S1A ) . Pgr-cre is expressed postnatally in the anterior pituitary , preovulatory granulosa cells of the ovaries , oviduct , and epithelial and stromal compartments of the uterus [6] . Cre-mediated recombination was confirmed by performing PCR on uterine DNA collected from Alk2 control and cKO females; we used primers specific for exon 7 of Alk2 ( Figure S1A ) and observed that recombination had efficiently occurred in mice carrying the Pgr-cre allele ( Figure S1C ) . To further confirm the efficiency of deletion of Alk2 in the uterus , we treated ovariectomized mice with 17β-estradiol ( E2 ) and progesterone ( P4 ) to mimic the hormonal levels that naturally occur during blastocyst implantation [16] . We then quantified the expression of Alk2 by qPCR in samples enriched in the epithelial or stromal component obtained by enzymatic and mechanical separation of the two uterine compartments . Alk2 mRNA is present in both uterine compartments and the expression in the epithelium is ∼4-fold higher than in the stroma ( Figure 1A ) . In Alk2 cKO mice , the mRNA level of the receptor was reduced by ∼94% in the epithelium and ∼70% in the stromal compartment ( Figure 1A ) . To further investigate the spatiotemporal localization of ALK2 in the pregnant uterus , we performed immunofluorescence analysis during sequential time points ( Figure 1B–F ) . While ALK2 is not detectable at 2 . 5 dpc , the receptor is highly expressed on the luminal side of epithelial cells at 3 . 5 and 4 . 5 dpc , the period during which the uterine epithelium prepares for and permits the embryo to implant . At these time points , the receptor is also detected in the glandular epithelium but at lower levels in the stromal cells . At 5 . 5 and 6 . 5 dpc , ALK2 is highly expressed in stromal cells , and the receptor is more prominently localized in the secondary decidual zone ( further from the embryo ) than in the primary decidual zone ( closer to the embryo ) . Thus , ALK2 is present in both compartments during pregnancy , and suggests that BMPs play a role both in the epithelium , during the first phases of pregnancy , and in the uterine stroma after implantation . We next investigated how ablation of ALK2 in the uterus affects female fertility by performing a 6-month breeding trial in which sexually mature control and cKO females ( n = 10 for each group ) were mated with wild type males of proven fertility . Alk2 cKO female mice are sterile , while control littermates display normal fertility and fecundity ( 9 . 1±0 . 2 pups/litter and 1 . 0±0 . 2 litter/month ) , comparable to wild type ( 8 . 1±0 . 3 pups/litter and 1 . 2±0 . 1 litter/month ) and Pgrcre/+ females ( 7 . 8±0 . 1 pups/litter and 1 . 1±0 . 4 litter/month ) [17] . The presence of vaginal plugs in Alk2 cKO females showed that the infertility was not due to abnormal mating behavior . We then performed timed mating experiments in which control and cKO females were mated with wild type males and sacrificed at sequential time points during pregnancy . The attachment of the embryo to the uterine epithelium causes a local increase in vasculature permeability , and implantation sites can be visualized at an early stage by injecting Chicago blue dye intra venous ( i . v . ) . On the day of attachment ( i . e . , day 4 of pregnancy ) , we did not observe any difference in the number of sites of attachment between control and cKO females ( data not shown ) . The presence of a normal number of attached embryos shows that the hypothalamic–pituitary–gonadal axis is intact , and ovulation is normal . This suggests that the infertility in Alk2 cKO females is likely caused by uterine defects . Upon embryo attachment , the embryo starts invading into the uterine stroma ( day 5 ) and stimulates the differentiation of the stromal cells into decidual tissue , around each implanted embryo . When we inspected the implantation sites at day 5 . 5 post coitum ( 5 . 5 dpc ) ( Figure 2A–B ) , we observed that embryos in Alk2 cKO mice were attached to the luminal epithelium , but the epithelial layer was still intact because the invasion into the stroma had not occurred ( Figure 2E–F ) . Moreover , the uteri of Alk2 cKO mice still showed similar numbers of implantation sites compared to controls ( Figure 2D ) , but their size and weight were significantly reduced compared to controls ( Figure 2C ) . To further characterize this defect , we collected uteri from pregnant females at 6 . 5 and 7 . 5 dpc and examined the histology of the implantation sites ( Figure 3 ) . While the embryos eventually invade into the stroma by 6 . 5 dpc in Alk2 cKO mice , the process is delayed compared to controls ( Figure 3G , L ) . In addition , embryos that implant into the Alk2 cKO uteri are consistently smaller compared to controls ( Figure 3G–I , L–N ) and start to be reabsorbed at 7 . 5 dpc ( Figure 3N ) . The implantation sites are also smaller in cKO mice compared to controls at all time points observed beginning at 5 . 5 dpc . To further study implantation in Alk2 cKO mice , we analyzed prostaglandin-endoperoxide synthase 2 ( PTGS2 ) protein levels and localization by immunohistochemistry . PTGS2 shows a dynamic localization that precisely correlates with different phases of the implantation process [18] . In particular , PTGS2 localizes to luminal epithelial and subepithelial stromal cells at the antimesometrial pole of the implantation site at the moment of embryo attachment ( ∼4 . 0 dpc ) , and at 5 . 5 dpc , PTGS2 is produced in decidualizing cells at the mesometrial pole . In control mice , PTGS2 correctly localized at the mesometrial pole at 5 . 5 dpc . In contrast , in Alk2 cKO mice , at 5 . 5 dpc PTGS2 displayed a pattern resembling the moment of attachment ( Figure 2G–H ) , confirming that Alk2 cKO mice have a delay in the implantation process . These results show that ALK2 is dispensable for the attachment of the embryo to the uterine epithelium , but is required at a later step during implantation . Ablation of BMP2 has also been shown to disrupt the implantation process soon after the embryo attachment , although the differences in the phenotypes observed between Alk2 and Bmp2 cKO mice suggest that BMP2-ALK2 is not the only ligand-receptor pair involved in the regulation of this stage of pregnancy . In fact , the presence of other BMP signaling components in the pregnant mouse uterus suggests that other BMP ligand ( s ) and receptor ( s ) can compensate with various degrees of efficiency for the loss of one component of the pathway . Invasion of an embryo into the uterine wall stimulates stromal cells to undergo a functional and morphological transformation called decidualization . During this process , stromal cells of the uterus extensively proliferate and differentiate to create a permissive environment for the implanting embryo . Because Alk2 cKO mice show smaller implantation sites , we decided to further investigate how ablation of ALK2 affects the proliferation and differentiation of the uterus when a decidualizing stimulus is given . The uterus is receptive only for a limited period of time during which the sex hormones prepare it to adequately respond to the attachment of the embryo . Because sex hormones are produced by the ovaries in cyclic waves , we ovariectomized and injected mice with sex hormones to artificially induce receptivity and exogenously control and synchronize the cycle of the mice in our experiments . We mimicked the implantation of the embryo by injecting one horn with oil . Consistent with the reduced size of implantation sites during natural pregnancy , we observed that Alk2 cKO females undergo a significantly reduced increase in uterine horn weight one day after oil injection compared to control littermates ( Figure 4A–C ) . At five days , the difference in size of the stimulated to untreated horn of Alk2 cKO mice to control mice was even greater ( Figure S2A–C ) . These results confirm that the uteri of cKO mice are not able to decidualize in response to the attachment of the embryo . We next investigated the ability of stromal cells to proliferate and differentiate in response to the decidual stimulus . Both control and cKO mice show extensive stromal proliferation in the uterus ( detected by MKi67 staining ) , but the amount of cells in the stromal compartment of cKO mice was reduced compared to controls , suggesting an impairment of the process ( Figure 4D–E ) . Moreover , cKO mice showed a significantly reduced expression of several molecular markers of the S phase of the cell cycle ( Ccne1 , Ccne2 , Cdk2 , and Cdc25 ) ( Figure 4F ) . To investigate whether the differentiation of stromal cells is also impaired by the ablation of ALK2 , we evaluated alkaline phosphatase activity ( ALP ) in stromal cells one day after decidualization was induced . While control mice showed a strong purple signal in the uterine stromal compartment ( indicating the presence of alkaline phosphatase activity , and thus differentiation ) , we observed a lack of differentiating cells in the uteri of cKO females ( Figure 4I–J ) . In addition , qPCR analysis of the two differentiation markers Prl8a2 and Prl3c1 confirmed a significant impairment of differentiation at a molecular level ( Figure 4K ) . These data demonstrate that signaling through ALK2 is required during decidualization to achieve stromal cell proliferation and differentiation . Although stromal cell proliferation and differentiation was reduced , Ki67 staining also revealed a pronounced proliferative activity in the epithelium of cKO mice that is not present in control mice ( Figure 4G–H ) . E2 stimulates the proliferation of uterine epithelial cells in preparation for implantation , and this proliferative action is inhibited by P4 acting through its receptor , PGR [16] . These observations led us to hypothesize that ALK2 acts to regulate stromal cell proliferation and repress epithelial cell proliferation by regulating the expression and/or activity of PGR . However , when we assayed the levels of PGR in the uterus one day after the induction of decidualization , we did not observe any difference in the total amount nor cellular localization of PGR protein in control and cKO mice , with PGR mainly localized in the nuclei of stromal cells ( Figure S4A–D ) . To investigate the downstream pathways controlled by BMPs during the decidualization process , we performed microarray analysis on control and cKO uteri collected one day after artificial induction of decidualization . We found that the expression of 909 unique genes is significantly affected by the ablation of ALK2 ( p<0 . 01 , fold change <0 . 7 , >1 . 3 ) ( Figure 5A ) . We used the Database for Annotation , Visualization and Integrated Discovery ( DAVID ) to identify the main pathways controlled by BMPs during decidualization . During decidualization , stromal cells undergo a wave of proliferation and differentiation that lead to the formation of a specialized tissue ( the decidua ) able to sustain the growth of the implanted embryo . Our data show that without ALK2 , stromal cells cannot respond to a decidual stimulus by normally proliferating and differentiating . Thus , the significant alteration in the nucleotide synthesis pathways and DNA metabolic process that we observe in Alk2 cKO mice is likely a consequence of the defect in proliferation; moreover , because there is a profound remodeling of the vasculature during decidualization that is tightly regulated by immune system factors , it was not surprising to see an effect of ALK2 ablation on pathways involved in vasculature development and immune response ( Table 1 ) . Another Gene Ontology ( GO ) term that we found to be significantly affected in the array analysis comprises genes involved in the regulation of in utero development ( Table 2 ) . Among these genes is CCAAT/enhancer-binding protein beta ( Cebpb ) , which has been shown to regulate the proliferation and differentiation of stromal cells during decidualization . Similar to Alk2 cKO females , mice with a null mutation in Cebpb are not able to decidualize [19] . We used qPCR and immunohistochemistry to confirm that the expression of Cebpb mRNA and CEBPB protein levels are significantly reduced in Alk2 cKO mice compared to controls during decidualization ( Figure 5B–D ) . Although both ALK2 and CEBPB are required for stromal cell proliferation during decidualization , the two factors seem to regulate different phases of the cell cycle; while Cebpb cKO mice show a disruption of expression of factors controlling the G2 to M phase transition of the cell cycle , Alk2 cKO mice display an alteration at the S phase . These data suggest that BMP signaling acts upstream to regulate the expression of the transcription factor Cebpb , and this regulatory mechanism is required for the progression of the cell cycle in uterine stromal cells during decidualization . In many species , endometrial stromal cell decidualization is induced upon implantation of the embryo . In humans , this process takes place during the mid-secretory phase of each menstrual cycle , independently of pregnancy . Similar to the mouse , decidualization in humans is a process of dramatic transformation of endometrial stromal fibroblasts into secretory , epithelioid-like decidual cells . To study decidualization in humans , this process can be recapitulated in primary cultures of human endometrial stromal cells ( hESC ) [20] . Treatment of hESC with a cyclic AMP analog triggers the expression of decidual marker genes like prolactin ( PRL ) and insulin-like growth factor binding protein 1 ( IGFBP1 ) within hours and , via activation of protein kinase A ( PKA ) , sensitizes hESC to progesterone action . When hESC are treated with both cAMP and P4 , the expression of decidual markers is enhanced and the decidual phenotype is maintained for a longer period . We utilized hESC to investigate the role of ALK2 during decidualization in human . hESC were first treated with either non-targeting ( NT ) small interfering RNA ( siRNA ) or siRNA targeted to ALK2 ( ALK2 knock down , KD ) , then stimulated to decidualize with a hormonal cocktail containing estradiol , medroxyprogesterone acetate ( MPA ) and cAMP ( EPC ) . After 3 days of treatment , we used qPCR to measure the expression of IGFBP1 and PRL . Both decidualization markers show a significant reduction in ALK2 KD cells compared to NT-treated cells ( Figure 6A–B ) , indicating that ALK2 is required during decidualization both in mouse and human . Upon differentiation of the stromal compartment , many genes first expressed in the epithelium , are induced in decidual cells . CEBPB has been shown to be required for human and murine decidualization and the physical interaction between CEBPB and PGR has been demonstrated using an in vitro system [21]–[23] . Because we found Cebpb to be downregulated during decidualization in the mouse , we were interested in testing whether BMPs control the expression of CEBPB in human endometrial stromal cells as well . Both CEBPB mRNA and protein level were significantly lower in ALK2 KD cells compared to NT-treated cells , indicating that the involvement of the BMP-CEBPB pathway in the regulation of decidualization is conserved in mice and humans ( Figure 6C , E ) . Progesterone acts through its cognate receptor , the nuclear progesterone receptor ( PGR ) , to regulate implantation and decidualization via the activation of several downstream pathways [24] . Although Alk2 cKO mice show a desensitization to progesterone , neither the PGR mRNA nor protein levels show a significant alteration in the cKO mice . It has been suggested that BMPs are required for the expression of FK-506-binding proteins ( FKBPs ) that , in turn , control the activity of PGR by binding and maintaining the receptor in a functional state . This hypothesis is confirmed by the observation that Fkbp4 and Fkbp5 are downregulated in Alk2 cKO mice during decidualization ( Figure S3A–B ) . To investigate whether the same mechanism was involved in the control of decidualization in hESC , we first measured the expression and translation of PGR in NT-treated and ALK2 KD cells . In contrast to the mouse , we found that ablation of ALK2 in human cells causes a downregulation of progesterone receptor at both the mRNA and protein levels . Because the levels of PGR are reduced in ALK2 KD cells even before decidualization was induced ( day 0 ) , the downregulation of PGR is not a consequence of the lack of decidualization but is a direct effect of the ablation of ALK2 . To confirm the hypothesis that ALK2 is required for functional PGR activity , we measured the expression of known PGR targets that are normally activated by PGR during decidualization . The expression of all the PGR targets that we quantified by qPCR [i . e . , wingless-type MMTV integration site family , member 4 ( WNT4 ) , forkhead box O1 ( FOXO1 ) , and heart and neural crest derivatives expressed transcript 2 ( HAND2 ) ] is significantly reduced in ALK2 KD cells ( Figure 6F–H ) . These findings show that ALK2 is required during the decidualization of hESC , and the mechanism through which it acts likely involves the regulation of CEBPB and PGR . The regulation of CEBPB , but not PGR levels , by BMPs during decidualization , is conserved between mice and humans . BMPs exert their action on target cells by causing the phosphorylation and activation of their intracellular effectors , SMAD1 , SMAD5 and SMAD8 . These proteins , once activated , bind SMAD4 and accumulate in the nucleus where they activate the expression of a variety of BMP target genes . To test whether the canonical BMP-SMAD pathway was active in hESC , we treated the cells with recombinant human BMP2 ( rhBMP2 ) and measured the phosphorylation of SMAD1 , 5 and 8 at sequential time points . The decision to use rhBMP2 was driven by the known central role of this factor during decidualization in both mouse and human [8] , [9] . As expected , we observed that rhBMP2 causes the phosphorylation of SMAD1 , 5 and 8 between 30 minutes and 1 hour of treatment ( Figure 7A ) . We then repeated the same experiment using hESC treated with either NT-siRNA or ALK2-siRNA . Because ALK2 KD cells displayed a strong decrease in SMAD1 , 5 , 8 phosphorylation ( Figure 7B ) , we propose that BMP2 signals through ALK2 in hESC in a SMAD1 , 5 , 8-dependent manner to regulate the stromal decidualization . Nevertheless , some residual phosphorylation can still be observed at 30 minutes in ALK2 KD cells , and we suggest that this is due to a redundant action of another BMP type 1 receptor . Because progesterone has been shown to regulate the expression of CEBPB in the uterus , we tested whether BMPs directly regulate CEBPB expression or if the downregulation of CEBPB in ALK2 KD cells is secondary to an alteration in PGR expression and/or activity . To test these possibilities , we performed ChIP analysis of SMAD1/5 on putative DNA binding regions in the proximity of the CEBPB gene in hESC induced to decidualize . We found a significant enrichment of one of the putative regions in the DNA immunoprecipitated using an antibody directed against SMAD1/5 , indicating the presence of SMAD1/5 on the CEBPB 3′ UTR during decidualization ( Figure 7C–D ) . These findings confirm that BMPs regulate the expression of CEBPB during hESC decidualization and show that this regulation occurs directly through binding of SMAD1/5 to the 3′UTR of CEBPB . Because we also observed a decrease of PGR in ALK2 KD hESC , and CEBPB has been shown to regulate the expression of PGR in the mammary gland during alveologenesis ( unpublished ) , we wanted to test whether CEBPB regulates the expression of PGR in hESC during decidualization . We performed ChIP analysis of CEBPB on putative binding sites in the proximity of PGR and found a 35-fold enrichment of one of the four sites that we tested ( Figure 7E–F and Table S1 ) . Our results show for the first time a direct regulatory connection between BMPs , CEBPB , and PGR , three of the main regulators of decidualization .
The data presented in this paper show evidence that BMPs play a central role during early pregnancy in both human and mouse . In particular , we discovered that the BMP type 1 receptor ALK2 is required in the uterus for normal embryo implantation and subsequent decidualization of stromal cells . To implant , the embryo first attaches to the uterine epithelium , then invades into the uterine stroma . The uterine stroma responds to the embryo by undergoing stromal cell decidualization . Without uterine ALK2 , embryos start invading into the stroma one day later than in control mice , show growth retardation , and die by 7 . 5 dpc . In contrast , the ablation of BMP2 in the uterus causes the arrest of pregnancy at the attachment phase [8] . We suggest that without ALK2 , other receptors partially compensate and mediate BMP2 signaling to allow the completion of implantation . Both Bmp2 and Alk2 cKO mice show a defect in stromal cell proliferation and differentiation during decidualization . By using microarray expression analysis , we investigated the molecular pathways affected by the ablation of ALK2 during the induction of decidualization . It was not surprising to see that Bmp2 cKO and Alk2 cKO mice show a significant number of common downstream targets . In particular , in both mouse models there is a downregulation of the immunophilins , Fkbp4 and Fkbp5 . These FKBP proteins can associate with heat shock proteins ( HSP70 and HSP90 ) and regulate the intracellular trafficking and function of steroid hormone receptors [25] . Because the expression of Fkbp4 and Fkbp5 is repressed in the absence of BMP2 or ALK2 , the function of PGR is impaired and decidualization is defective . The impairment of PGR activity caused by the ablation of ALK2 also causes a perturbation of the uterine epithelium , It is known that the proliferation of uterine epithelial cells is controlled by sex hormones: during each estrous cycle , in preparation for implantation , an increase in E2 levels causes a wave of cell proliferation; this is followed by a surge of P4 that inhibits the E2-dependent proliferation and stimulates cells to differentiate [26] . Alk2 cKO mice show retention in epithelial proliferation when decidualization is artificially induced and during natural pregnancy , suggesting that the absence of ALK2 causes desensitization to progesterone . In preparation for embryo attachment , PGR activates the expression of heart and neural crest derivatives expressed 2 ( Hand2 ) in the stroma , and this , in turn , represses the FGF pathway that acts in a paracrine fashion to inhibit the ER-mediated proliferation in the epithelium [27] . Our results show that Alk2 cKO mice display retention in epithelial proliferation without showing a significant change in the mRNA or protein levels of HAND2 ( Figure S7C–E ) . ALK2 can regulate the epithelial proliferation during pregnancy through two alternative mechanisms . First , ALK2 may regulate the inhibitory activity of stromal PGR through a HAND2 independent mechanism . Second , ALK2 may repress epithelial cell proliferation directly . It has already been demonstrate that epithelial PGR can directly regulate the expression of epithelial target genes [28] , and signaling through ALK2 in the epithelium may mediate this repression . When we analyzed the level of PGR in the uterus during decidualization , we could not detect any significant difference in the amount or histological localization of the protein in the uterine epithelium and stroma ( Figure S4A–D ) . This observation confirms that PGR activity , rather than its expression or translation , is affected by the ablation of ALK2 in the mouse . ALK2 is localized in both epithelium and stroma during pregnancy , and its ablation affects both compartments: embryos show delayed implantation , and the stromal cells of the uterus do not respond to the embryo by normally decidualizing . Based on these results , it is not possible to establish whether the two defects are caused by the impairment of the same mechanism or if ALK2 controls two independent processes in the two compartments . Because we investigated decidualization by using an artificial induction experiment , we can conclude that the decidualization defect is not due to the delayed invasion of the embryo . However , we cannot determine whether decidualization is affected by a lack of ALK2 in the epithelium or in the stroma . In fact , both paracrine and autocrine factors seem to be involved in regulating this process . Because the uterus requires an intact epithelium to decidualize [29] , it has been suggested that some signal from the epithelium is required to have a functional response in stromal cells . It is possible that the lack of ALK2 in the epithelium makes stromal cells unresponsive to the invasion of the embryo , likely through a deregulation of either epithelial or stromal progesterone receptor activity . A role of ALK2 in epithelial cells in the mouse is supported by the observation that , ablation of ALK2 only in the stromal compartment by using the Amhr2cre/+ mouse model , does not significantly affect the fertility of the mice ( Figure S8A–B ) . However , the efficiency of the KO and the residual presence of ALK2 in this model have not been investigated and the normal fertility could be explained by a partial retention of ALK2 . When siRNA targeted to ALK2 is used to silence the receptor in human endometrial stromal ( hES ) cells , the expression of the decidual markers IGFBP1 and PRL is significantly repressed compared to hES cells treated with non-targeting siRNA . These results suggest that ALK2 is required in the human uterine stroma during decidualization . Although this in vitro model of decidualization has been largely used by others [30] and is commonly accepted as a valid tool to recapitulate the decidualization process in cultured cells , the limitations of this simplified , artificial system must still be taken into account before drawing conclusions . However , the similarities of the findings made by using a combination of in vitro and in vivo systems strongly suggests a certain level of conservation . For instance , in both the mouse and in human cells , BMP2 and ALK2 have been shown to be required for implantation and decidualization and the pathways downstream to BMP2 in decidualization have been shown to be conserved in mouse and human cells [9] . Furthermore , in both mouse and human cells , CEBPB levels are decreased without ALK2 , suggesting that the same downstream pathways are affected . Our data show that when ALK2 is ablated in hES cells , several pathways are affected . In particular , we saw that the presence of ALK2 is required for the expression of two important regulators of decidualization: CEBPB and PGR . Although the role of these factors in decidualization has been already demonstrated , we found that BMP , CEBPB and PGR signaling are directly connected and can regulate each other's expression during decidualization . P4 activates Bmp2 and Cebpb expression in the mouse uterus [4] , [19] and , in hES cells , we show that BMPs signal through ALK2 and SMAD1/5 to promote CEBPB expression . Finally , CEBPB binds to intron 2 of PGR . In the mouse , we also show that BMP signaling is required for the expression of Cebpb , but the expression of Pgr is not reduced in mice lacking uterine ALK2 . The results we obtained in hESC suggest a mechanistic solution to the phenotype that we observe in the Alk2 cKO mouse . In total , these data support a fundamental role of ALK2 in both species . In humans , infertility is one of the most common reproductive disorders , with 10–15% of couples finding it difficult or impossible to conceive [3] . The implantation of the embryo in the uterus is one of the most critical steps of pregnancy . Approximately 75% of pregnancy failures are due to implantation defects [31]–[33] . The implantation rates in the assisted reproductive technology ( ART ) laboratory rapidly decrease with the age of the female patient , going from 30 . 8% in women under 35 year old ( y . o . ) to 8 . 2% in the upper age range ( 41–42 y . o . ) ( SART 2009 ) . Furthermore , early pregnancy loss occurs in up to ∼50% of patients [34] , [35] . Even when defective implantation does not cause pregnancy loss , it can still be the cause of several later complications [i . e . , spontaneous abortion , preeclampsia ( high blood pressure and protein in the urine ) , preterm labor , premature rupture of the membranes , intrauterine growth retardation , and disorders of placental attachment] [32] , [35] . Roles of BMP2 and activin A in the process of decidualization has been shown in vitro using human endometrial stromal cells and in vivo for BMP2 using a mouse model [8] , [9] , [36] . We now show that ALK2 is another signaling component of the TGFβ superfamily involved in the regulation of implantation and decidualization in both mouse and human . Because the ablation in the uterus of BMP2 compared to its receptor ALK2 affects implantation of the embryo at different stages , we suggest that other BMP ligands and/or receptors are also involved in the regulation of early stages of pregnancy . In our current study , we further clarify the molecular mechanism through which BMPs act in the uterus during decidualization , and we demonstrate for the first time a direct connection between BMP , CEBPB , and PGR pathways . A better knowledge of the molecular mechanisms involved in the regulation of implantation will be critical for developing drugs designed for improving pregnancy rates or , alternatively , for developing new forms of contraceptives that avoid the side effects of the current hormonal contraceptives , such as an increased risk of blood clots , high blood pressure , and breast cancer [37]–[39] . Furthermore , the involvement of TGFβ signaling in a broad array of developmental and physiological processes [40] makes the investigation of the pathway a topic that goes well beyond the scope of the field of reproductive biology . New drugs acting on the TGFβ pathway could be suitable for use in the treatment of a variety of human diseases ranging from cancer , fibrosis and cardiovascular diseases to metabolic disorders like diabetes and obesity [41]–[47] .
Mice carrying the Alk2 conditional allele ( Alk2flox/flox ) were created by flanking Alk2 exon 7 , encoding a critical portion of the kinase domain , with two loxP sites [15] . The Alk2flox/flox mice were bred to mice carrying progesterone receptor-cre knock-in ( Pgrcre/+ ) allele [6] to generate Alk2flox/floxPgrcre/+ mice , which were designated as Alk2 cKO . Alk2flox/flox female mice were used as controls . Mice were genotyped by PCR analyses of genomic tail DNA as described by others [48] . Analysis of DNA recombination in the uterus was performed using DNA isolated from uterine tissue . All mouse lines used in the present study were maintained on a hybrid C57BL/6J and 129S7/SvEvBrd genetic background . Animal handling and surgeries were performed according to the NIH Guide for the Care and Use of Laboratory Animals . All procedures have been approved by the Institutional Animal Care and Use Committee ( IACUC ) at Baylor College of Medicine . Physiological concentrations of E2 and P4 were given exogenously to mice to recapitulate the hormonal levels occurring during early pregnancy as previously described [16] . Briefly , 6–8 week old mice were ovariectomized via a dorsal incision under 2 , 2-tribromoethanol ( 2 . 5% Avertin ) anesthesia and left untreated for 2 weeks to achieve complete withdrawal of endogenous ovarian hormones; then , mice were “primed” with daily subcutaneous ( s . c . ) injections of 100 ng of 17β-estradiol ( E2 ) for 2 days . After 2 days of rest , mice were injected for 3 days with 1 mg of progesterone ( P4 ) daily and on the fourth day with 1 mg of P4 and 50 ng of E2 . Mice were sacrificed 6 hours after the last injection . The artificial induction of decidualization has been previously described [26] . Briefly , control and cKO mice were ovariectomized and , after 2 weeks , treated with s . c . injections of 100 ng of E2 once a day for 3 days . After 2 days rest , mice were injected daily with P4 ( 1 mg , s . c . ) and E2 ( 6 . 7 ng , s . c . ) once a day for 3 days . One uterine horn was then treated with a decidual stimulus by injecting 200 µl of sesame oil into the lumen 6 hour after the last hormone injection . The contralateral horn was not injected and served as a control . The next day , mice were given a final injection of P4 ( 1 mg , s . c . ) and E2 ( 6 . 7 ng , s . c . ) and euthanized 6 hours later; wet weight of the traumatized and control horns was recorded . Another group of mice was injected for 5 days after the oil injection and sacrificed 6 hours after the last injection . Upon euthanasia , the uterine tissue was collected and either fixed in 4% paraformaldehyde ( PFA ) or frozen for RNA extraction . To separate the epithelial and stromal components of the uterine tissue , uteri were collected , cut in 1–2 mm pieces and incubated in Hank's balanced salt solution ( HBSS , Invitrogen ) +1% trypsin ( Sigma ) for 3–5 hour at room temperature . Afterwards , tissues were washed in PBS and luminal epithelium was mechanically separated from the stromal compartment under a dissecting microscope using dissecting forceps and a mouth pipette . To ensure adequate purity of the isolated epithelia and stromal cell populations , the expression of epithelial ( cadherin1 , Cdh1 ) and stromal ( vimentin , Vim ) markers was evaluated by qPCR ( data not shown ) . To examine the fertility of female mice , wild type ( WT ) , Pgrcre/+ , control ( Alk2flox/flox ) , and Alk2 cKO females were mated independently with WT fertile male mice for a 6-month period ( n = 10 per genotype ) . Cages were monitored daily , and the numbers of litters and pups were recorded . To investigate the different phases of early pregnancy we performed timed mating experiments: adult female mice were mated to fertile wild type males and the presence of vaginal plugs was monitored every morning . The morning of the plug was designated as 0 . 5 day post coitus ( dpc ) ; to investigate embryo attachment , 4 . 5 dpc pregnant animals were injected intravenously ( i . v . ) in the tail vein [49] with Chicago blue dye and sacrificed 10 minutes after the injection . Uteri were dissected and fixed in 4% PFA for histology . Tissue processing and embedding in paraffin were performed in the Baylor College of Medicine Department of Pathology Core Laboratory . Paraffin sections ( 5 µm ) were deparaffinized , rehydrated and boiled in citrate buffer for antigen retrieval . Blocking of non-specific signal was performed by incubating the sections with 10% normal goat serum ( NGS ) or 3% BSA in TBS-T for 1–3 hours at room temperature . Tissue sections were incubated overnight at 4°C with primary antibodies specific to ALK2 ( 1∶200 , LifeSpan Biosciences , Inc . , LS-B6835 ) , CDH1 ( 1∶400 , Cell Signaling , 24E10 ) , PTGS2 ( 1∶1000 , Thermo Scientific , RB-9072 ) , MKI67 ( 1∶1000 , BD BioScience , 550609 ) , CEBPB ( 1∶200 , Santa Cruz Biotechnology , Inc . sc-150 ) , PGR ( 1∶500 , DAKO , IR068 ) or HAND2 ( 1∶100 , R&D Systems , AF3876 ) . After washing with TBS-T , sections were either incubated with biotin-labeled ( CDH1 , PTGS2 , MKI67 , CEBPB , PGR , HAND2 ) or fluorescent-labeled ( ALK2 , PGR ) secondary antibody for 1 hour at room temperature . Sections treated for immunohistochemistry were then incubated with VECTASTAIN ABC solution ( Vector Labs ) for 30 minutes followed by peroxidase substrate ( Vector Labs ) for signal development . Finally , sections were briefly counterstained with CAT hematoxylin ( Biocare Medical ) , dehydrated and mounted . Alkaline phosphatase staining was performed as previously described [50] Frozen sections were incubated with a 100 mM Tris buffer ( pH 9 . 5 ) containing 5-bromo-4-chloro-3-indolyl phosphate and Nitro blue tetrazolium chloride ( Roche Applied Science ) . Nuclear Fast Red ( Vector ) was used for counterstaining . Uterine tissues were collected from control and Alk2 cKO female mice and stored immediately at −20°C until RNA extraction . Total RNA for each sample was extracted using the RNeasy Mini kit ( Qiagen ) . Gene expression was analyzed by real-time quantitative polymerase chain reaction ( qPCR ) . Either SYBR Green detection system ( Life technologies ) or TaqMan Assays-On-Demand PCR primer and probe sets ( Life technologies ) were used for individual gene analysis . Melt curve analysis was performed when using Sybr Green to verify a single amplification peak . Relative mRNA levels of transcript were calculated by the 2−ΔΔC method after normalizing to the endogenous reference ( Rn18S ) , and plotted as mean ± SEM . For microarray analysis , we pooled RNA from three mice per genotype and performed the experiment in triplicate . Samples were tested for quality assurance on the Agilent 2100 Bioanalyzer by the Genomic and RNA Profiling Core at Baylor College of Medicine . The labeled cRNA was hybridized to Illumina Mouse WG-6 v . 2 . 0 by the Genomics and Proteomics Core at Texas Children Hospital ( Houston , TX ) . Array data were quantile normalized . Differentially expressed genes were defined , using two-sided t-test and fold change ( on log-transformed data ) . Array data have been deposited on the Gene expression omnibus ( GEO , accession number GSE46689 ) . Samples were collected from patients undergoing gynecological surgery at Ben Taub General Hospital or the Obstetrics and Gynecology Clinic of Baylor College of Medicine ( Houston , TX ) . The study was approved by the Institutional Review Board of Baylor College of Medicine . Undifferentiated stromal cells were obtained from endometrial biopsies from normally cycling fertile women with no history of uterine disease in the proliferative phase of the cell cycle . Written informed consent was obtained before sample collection . All samples were collected under Baylor College of Medicine Institutional Review Board ( IRB ) approval and IRB approval from each individual hospital . After collection , samples were transported to the laboratory and processed to obtain hESC . Cells were cultured at 37°C , 5% CO2 , 90% RH in growth medium ( DMEM/F12 +10% csFBS +1% Antibiotic-Antimycotic +1% HEPES ) . To investigate the role of ALK2 in controlling the decidualization of stromal cells in human we utilized siRNA specifically targeting ALK2 to knock down the receptor ( Figure S5 ) . The day of decidualization induction was designed as day 0 . During the two days prior to induction ( day −2 and −1 ) , cells in 6-well plates were treated for 6–8 hours with DMEM/F12 medium +2% csFBS ( Invitrogen ) containing a mix of Lipofectamine 2000 ( Invitrogen ) and 100 pmol of either non-targeting ( NT ) siRNA ( ON-TARGETplus Non-targeting Control Pool , Thermo Scientific ) or siRNA targeting to ALK2 ( ON-TARGETplus Mouse Acvr1 siRNA , Thermo Scientific ) dissolved in Opti-MEM ( Invitrogen ) . On day 0 , an aliquot of cells for each treatment ( NT or ALK2 siRNA ) was collected and the remaining cells were treated with decidual medium containing 10−8 M 17β-estradiol , 10−6 M 17α-hydroxy-6α-methylprogesterone acetate ( MPA ) and 50 µM cyclic AMP ( cAMP ) in Opti-MEM +2% csFBS +1% Antibiotic-Antimycotic ( Invitrogen ) . The medium was changed ( to freshly prepared decidual medium ) on day 2 , and on day 3 cells were harvested and stored at −80°C . To investigate the effects of rhBMP2 on SMAD1 , 5 , 8 phosphorylation in human stromal cells , we dissolved rhBMP2 ( R&D Systems , 355-BM ) in DMEM/F12 +10% csFBS at a concentration of 100 ng/ml and we treated the cells with this solution for 30 , 45 , 60 , 180 or 360 minutes . To test the effect of ALK2 ablation on SMAD1 , 5 , 8 phosphorylation , we performed the same experiment but transfected the cells with either NT or ALK2 siRNA during the two days before rhBMP2 treatment as described above . After the treatment , cells were trypsinized , washed in PBS , pelleted and stored at −80°C . Protein extraction was performed by incubating the pelleted cells with RIPA buffer ( Thermo Scientific ) in agitation for 30–45 minutes . Protein concentration was determined by Bradford's method using BSA as the standard . Samples containing 20 µg of total protein were subjected to sodium dodecyl sulfate ( SDS ) -polyacrylamide gel ( NuPAGE 4–12% Bis-Tris gel , Invitrogen ) electrophoresis . The separated proteins were transferred onto a nitrocellulose membrane and the membrane was blocked for 1 hour with 5% skim milk in Tris-buffered saline with 10 mM Tris-HCl +150 mM NaCl +0 . 1% Tween 20 ( TBS-T ) . Then , the membrane was probed overnight with antibodies against C/EBPβ ( 1∶500 , Santa Cruz , sc-150 ) , PGR ( 1∶500 , Santa Cruz , sc-7208 ) , p-SMAD1 , 5 , 8 ( 1∶500 , Cell Signaling , 9511 ) or β-actin . After 1 hour-incubation with horseradish peroxidase-linked secondary antibodies , the membrane was treated with enhanced chemiluminescence reagents ( SuperSignal West Pico Chemiluminescent Substrate , Thermo Scientific ) to visualize the immunoreactivity . ChIP analysis was performed using the Imprint Chromatin Immunoprecipitation kit ( Sigma ) according to the manufacturer's protocol . Briefly , human endometrial stromal cells were treated with decidual medium for 3 days as described above . On the third day of treatment , cells were cross-linked in 1% formaldehyde in DMEM/F12 and after 10 minutes the reaction was stopped by adding 1 . 25 M glycine . Cells were washed with cold PBS , pelleted and treated with Nuclei Preparation buffer and Shearing Buffer containing protease inhibitor cocktail ( PIC ) to prepare the DNA for sonication . The DNA was sheared by sonication: always keeping the cells in ice , we alternated 7 cycles of 20 second sonication with one-minute periods of rest . A small aliquot of the sheared DNA was analyzed by agarose electrophoresis . Debris was precipitated by centrifugation and the supernatant was incubated with 1 µg of Normal Mouse IgG ( negative control ) , Anti-RNA Polymerase II ( positive control ) , anti-SMAD1/5 ( Santa Cruz , sc-7965 ) , or anti-CEBPB ( Santa Cruz , sc-150 ) antibody overnight . The next day , crosslinking was reversed by treatment with proteinase K in Release buffer and Reversing Solution . Finally , the DNA was purified using GenElute Binding Columns . To quantify the amount of template pulled down from the ChIP reactions we performed qPCR using SYBR green detection system . We used various primer sets to amplify specific regions in the proximity of the CEBPB and PGR genes ( Table S1 ) . The putative binding sites of SMAD1/5 were found by searching the conserved consensus sequence GGCGCC ( Bmp Responsive Element , BRE ) in the CEBPB gene ±1 kb; the putative binding sites of C/EBPβ on PGR were identified by using http://cistrome . org/finder/ . | A couple is defined as infertile when failing to become pregnant after one year of regular , unprotected intercourse . Infertility affects more than 10% of couples . The implantation of the embryo in the uterus is one of the most critical steps of pregnancy , and it has been estimated that 75% of pregnancy fails because of peri-implantation defects . An intricate network of molecular pathways regulates the peri-implantation process . It is known that the bone morphogenetic protein ( BMP ) pathways are part of this network , and herein we investigated how one of the BMP signaling receptors interacts with other factors in the uterus . Our results show an essential and conserved role of this BMP receptor during the implantation of the embryo in mice and humans . Furthermore , we discovered that BMPs act in a linear pathway upstream of two other key regulators of implantation , CEBPB and PGR . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2013 | Activin-Like Kinase 2 Functions in Peri-implantation Uterine Signaling in Mice and Humans |
Nematode-trapping fungi ( NTF ) are a large and diverse group of fungi , which may switch from a saprotrophic to a predatory lifestyle if nematodes are present . Different fungi have developed different trapping devices , ranging from adhesive cells to constricting rings . After trapping , fungal hyphae penetrate the worm , secrete lytic enzymes and form a hyphal network inside the body . We sequenced the genome of Duddingtonia flagrans , a biotechnologically important NTF used to control nematode populations in fields . The 36 . 64 Mb genome encodes 9 , 927 putative proteins , among which are more than 638 predicted secreted proteins . Most secreted proteins are lytic enzymes , but more than 200 were classified as small secreted proteins ( < 300 amino acids ) . 117 putative effector proteins were predicted , suggesting interkingdom communication during the colonization . As a first step to analyze the function of such proteins or other phenomena at the molecular level , we developed a transformation system , established the fluorescent proteins GFP and mCherry , adapted an assay to monitor protein secretion , and established gene-deletion protocols using homologous recombination or CRISPR/Cas9 . One putative virulence effector protein , PefB , was transcriptionally induced during the interaction . We show that the mature protein is able to be imported into nuclei in Caenorhabditis elegans cells . In addition , we studied trap formation and show that cell-to-cell communication is required for ring closure . The availability of the genome sequence and the establishment of many molecular tools will open new avenues to studying this biotechnologically relevant nematode-trapping fungus .
A predatory lifestyle is typically associated with animals like lions , tigers or snakes . Dramatic hunts or attacks out of a hiding place are typical scenes in reportage movies . No less dramatic is the situation when some nematode-trapping fungi ( NTF ) catch their prey using sophisticated trapping systems . NTF are an important group of soil microorganisms , which are able to live also saprotrophically . Some other nematode-attacking fungi are obligate pathogens [1–3] . NTF have evolved a variety of trapping structures , including constricting rings and five types of adhesive traps ( sessile adhesive knobs , stalked adhesive knobs , adhesive nets , adhesive columns , and non-constricting rings ) with which they capture nematodes to supplement their diet . There is evidence that NTF originated in the Permian and Triassic [4] . One hypothesis is that dead creatures caused by mass extinctions at that time were rich in carbon but poor in nitrogen and hence direct capture of nitrogen-rich living animals would give predatory fungi a competitive advantage over strictly saprotrophic fungi . Adhesive , non-constricting rings or hyphal loops are considered as ancient structures , after which constricting rings evolved [4] . The organismic interaction between NTF and nematodes reflects more than 400 million years of co-evolution and is highly sophisticated [4] . NTF “smell” their prey and only form traps in the presence of nematodes [5] . On the other hand , fungi secrete volatile substances to attract the worms [6] . It was hypothesized that chemical warfare and sophisticated communication between the two partners occurs also after penetration of the hyphae into the nematode [7] . It is conceivable that secreted proteins may fulfill such functions as described in plant-fungal interactions . For instance , the oomycete Phytophthora infestans attenuates the plant defense system with secreted effector proteins before killing the host cells [8] . In U . maydis the effector Rsp3 is exposed at the hyphal surface during biotrophic growth and inactivates two antifungal peptides [9] . Likewise , the arbuscular mycorrhizal fungus Glomus intraradices secretes an effector that suppresses the host defense and thereby contributes to the establishment of a long-term symbiotic relationship [10] . Hence , effectors play key roles in modulating host cell physiology to promote virulence , biotrophic growth or symbiosis [11–14] . In NTF , certain secreted proteins may be used for modulating the innate immune system of the nematode or target other intracellular processes . Such fungal proteins would share a secretion signal and may contain additional signals for subcellular localization within the host cells . These motifs could be organellar targeting signals , such as the NLS found in the U . maydis effector , See1 ( seedling efficient effector 1 ) , which reactivates host DNA synthesis and cell division [15] . In addition , to the fascinating predatory lifestyle of NTF , NTF were applied as biocontrol agents against nematodes , which colonize the gastrointestinal tract of sheep . Heavy loads of nematodes may cause dramatic reductions in productivity [16] . As nematode eggs are secreted with the feces , and sheep are grazing around their feces , re-infections are very common and thus antihelmintic treatments are only of temporary help . Excitingly , NTF can be used to reduce the nematode number in a field . Duddingtonia flagrans and Monacrosporium thaumasium have the greatest potential as biocontrol agent because they form resistant chlamydospores [17 , 18] . It has been shown that chlamydospores are able to pass the gastrointestinal tract of animals and germinate in the feces where they reduced the number of nematodes [19 , 20] . Multinutritional pellets containing D . flagrans chlamydospores were already used in sheep and horses [21–23] . Nematodes are also a threat for many crop plants . For instance , the nematode Heterodera schachtii attacks sugar beet plants and may cause complete loss of the harvest [24] . There are no effective treatments of such pests , besides cultivation plans with different crops . At least six years are required until the reservoir of infective nematodes is cured . Another example is Xiphinema index , a very large pathogenic nematode , which attacks the roots of grapevine . Besides direct damage of the roots , X . index transmits the Grapevine fanleaf virus ( GFLV ) , which causes large losses in the wine industry [25] . There are estimates that almost 10% of all Chardonnay Vineyards in France are infected with the virus . The last example is the attack of coffee plants by nematodes of the genera Helicotylenchus and Xiphinema [26] . There are estimates that nematodes cause annual losses of more than $ 80 billion worldwide [27] . In this paper , we studied D . flagrans , sequenced its DNA and annotated its genome , developed a transformation system with three different selection markers , established the use of the CRISPR/Cas9 technology and fluorescent proteins for subcellular localization . With these tools , we studied a putative fungal virulence factor , PefB . We show that it is a secreted protein that is upregulated during infection . The mature protein was expressed in C . elegans where it is localized to nuclei . Finally , we generated hypotheses to analyze chlamydospore morphogenesis and proved the hypothesis that intercellular communication is required for ring closure during trap formation .
It is very favorable for the application of microorganisms as biocontrol agents if ( 1 ) the organisms are easy to cultivate and if ( 2 ) robust cellular structures are produced , which can be used to disseminate the agent; both apply to D . flagrans . In comparison to other NTF , D . flagrans grows relatively fast under laboratory conditions on typical media on agar plates . Two days after point inoculation on a potato-dextrose-agar ( PDA ) plate a colony of 1 cm diameter was observed . D . flagrans is able to capture nematodes with three-dimensional trapping networks . In order to induce trap formation , we inoculated D . flagrans spores on low nutrient agar ( LNA ) and added around 100 individuals of C . elegans to the plate . After 6 h incubation at room temperature , the first forming traps were observed . After 12 h , the first trap networks were complete , and nematodes trapped . Completely digested nematodes were usually found after 24 h ( Fig 1A–1C ) . D . flagrans produces ovoid-shaped asexual spores ( conidia ) at the tip of conidiophores after 24 h ( Fig 1D ) , and , under starvation conditions , vegetative hyphae are able to differentiate into asexual spores resistant for survival ( chlamydospores ) ( Fig 1E ) [18] . Under laboratory conditions , the formation of chlamydospores was observed after 3 days of growth on 6 cm LNA Petri dishes and incubation at 28°C . Usually chlamydospores develop from vegetative compartments or at hyphal tips . In addition , we observed that after at least 48 h of co-incubation with C . elegans single compartments of empty traps transformed into chlamydospores ( Fig 1F ) . In order to show that the round spores and the transformed trap compartments were indeed chlamydospores , we tested if they contain glycogen as storage material . When incubated with Lugol’s iodide , chlamydospores showed a brown color , characteristic for glycogen staining , whereas hyphae and conidia were much less or not stained ( Fig 1G and 1H ) . To analyze the predation process in detail , fluorescent microscopy was used . In order to visualize fungal hyphae inside the nematode , fungal cell walls were stained with calcofluor white ( Fig 1I ) , and C . elegans nuclei were visualized by GFP expression . During early stages of the interaction , we observed a ring-like accumulation of chitin at the contact zone of D . flagrans trap cells with the nematode cuticle ( Fig 1I and 1J ) . These accumulations likely represent penetration tubes emerging at the anchorage zone 2–4 h after adhesion [3 , 28] . To test if D . flagrans is able to trap and inactivate plant-pathogenic nematodes , the grapevine-pathogenic nematode X . index was tested as prey . X . index is an obligate pathogen and can be cultivated on grapevine plants or fig trees [29] . X . index is much larger than C . elegans . Adults , which are 3 mm in length , were hand-picked under a stereo microscope and in total 10 individuals were added to 5x104 D . flagrans spores on LNA plates . After 12 h co-incubation first traps were observed and after 24 h first immobilized nematodes . After 48 h the first nematodes were digested ( Fig 1K–1M ) . These results show that D . flagrans is able to attack different nematodes and has thus the potential to be developed as a biocontrol agent against plant-pathogenic nematodes . D . flagrans genomic DNA was sequenced using Pacific Biosciences SMRT to generate 411 , 097 reads with an average read length of 8 , 516 nt . The genome was assembled de novo and 88% of reads were mapped against the genome of the related NTF Arthrobotrys oligospora . The average depth coverage was 102 . 18 ( S1 Table ) . The high-quality sequence was assembled into 9 contigs , reaching from 8 , 347 , 096 bp to 28 , 584 bp in size ( Table 1 , S1 Table ) , summing up to 36 . 64 Mb and with an overall G+C content of 45 . 5% ( S1 Fig , S2 Table ) . Six contigs had sizes between 4 . 4 Mb and 8 Mb , whereas the other eight contigs were only between 0 . 28 Mb and 0 . 028 Mb in length . The six large contigs may represent six chromosomes . In order to improve the annotation , we performed an RNAseq analysis . RNA was extracted from D . flagrans saprotrophic and zootrophic stages and pooled to one sample . The RNA of the different developmental stages should increase the possibility that a maximum number of genes would be expressed and detected in the RNAseq analysis . The sample was sequenced using the Ilumina platform ( BGISEQ-500 ) at the Beijing Genomics Institute ( Beijing , China ) . 23 , 941 , 723 reads were generated and 91 . 79% were mapped to the D . flagrans genome . 88 . 84% had a unique match . Structural gene annotation was then done with BRAKER2 . 9 , 927 genes were predicted , which is close to the gene number in A . oligospora and similar to the coding capacity of other ascomycetes ( S3 Table ) . Among the 9 , 927 protein-coding genes only 5978 showed homology to proteins deposited in the Swiss-Prot database ( S4 Table ) . These results indicate the presence of a high number of novel proteins in D . flagrans . Using A . oligospora as database , we found 9 , 336 similar proteins in D . flagrans ( S5 Table ) . Subcellular localization predictions revealed 41% nucleic , 23% cytosolic , 19% mitochondrial and 12% extracellular proteins . Mitochondrial genes were also annotated ( S6 Table ) . To annotate orthologous proteins , the proteome of D . flagrans was compared ( BLASTP ) to the proteomes of other NTF using a threshold of 80% predicted amino-acid similarity with at least 50% coverage . We found 7 , 488 , 4 , 373 and 2 , 352 orthologous proteins in A . oligospora , Dactylellina haptotyla ( abbreviated as Da . haptotyla ) , and Drechslerella stenobrocha ( abbreviated as Dr . stenobrocha ) , respectively . The four species were used to generate a Venn diagram of orthologous clusters ( Fig 2 ) . Orthologous proteins gathered together in the same cluster , and among the total number of 9 , 800 clusters , 9 , 598 were orthologues ( at least present in two species ) and 4 , 475 were single-copy gene clusters ( Table 2 ) . 4 , 646 cluster gene families were orthologous in all four fungi . Moreover , D . flagrans shares 950 , 75 and 15 gene families with A . oligospora , Da . haptotyla and Dr . stenobrocha , respectively . The highest number of clusters shared between D . flagrans and A . oligospora reflects their close relationship ( Fig 2 and Table 2 ) . 855 proteins of D . flagrans had no orthologous proteins with any of the other three species ( Table 2 ) . We also analyzed the proteome for potential genes involved in the worm attack . We assume that such proteins may be related to proteins involved in pathogenic interactions . Therefore , we analyzed the D . flagrans proteome using the InterProScan and the Pathogen Host Interactions ( PHI ) databases ( S1 Text , S7–S10 Tables ) . Additionally , we analyzed the phylogenetic relationship between D . flagrans and 13 other filamentous fungi by using orthologous proteins ( S3 Fig ) ( S1 Text ) . Secondary metabolites are defined as metabolites which are not essential for the survival of the producing organism in the laboratory , many of which are only produced under specific growth or physiological conditions . A large proportion of secondary metabolites are polyketide or peptide derivatives , which are synthesized by polyketide synthases ( PKS ) and non-ribosomal peptide synthases ( NRPS ) , respectively . The genome of D . flagrans harbors only three PKS and three NRPS type gene clusters [30] . In the genome of A . oligospora , five putative PKS and seven putative NRPS genes were found . Likewise , Drechslerella contains two putative PKS/PKS-NRPSs and three putative NRPSs [26] . These are rather low numbers as compared to other filamentous fungi . In A . alternata ten different PKS/PKS-NRPS and five NRPS clusters are found [31 , 32] , in F . graminearum 12 PKSs and 10 NRPSs [33] , and in A . flavus even 29 PKSs and 27 NRPSs [34] . In D . flagrans one of the PKS displays similarity to 6-MSAS type polyketide synthases with strong homology to the PKS encoded by AOL_s00215g283 . This PKS is described to be negatively involved in trap formation in A . oligospora [35] . It is likely that it fulfills a similar function in D . flagrans . The other 2 PKS genes ( both type I PKS ) could not yet be assigned a function , but they show high homology to the type I PKSs , AOL_s00215g926 , AOL_s00079g496 and AOL_s00043g828 , which were predicted to be involved in lovastatin biosynthesis [36] . The fungal attack of nematodes requires a repertoire of lytic enzymes and perhaps other virulence factors or effectors to modulate the innate immune response or overcome the host defense . In order to identify candidates for both classes of proteins , we analyzed the putative secreted proteins in D . flagrans . Using SignalP , WolfPsort and TMHMM , we identified 638 proteins ( ~6 . 4% of the whole genome ) . Proteins with hydrolase activity , oxidoreductases and peptidase activity are the most abundant . Carbohydrate metabolic and other catabolic processes are also over-represented compared to other functions ( S11 Table ) . In order to identify putative effectors or virulence factors we applied the EffectorP 2 . 0 machine learning method for fungal effector prediction in secretomes . This tool is based on machine learning and is trained with fungal effectors as positive training set and secreted proteins not classified as effectors as the negative set [37] . 117 such proteins were found ( S12 Table ) ( Fig 3 ) . We further used WoLF PSORT to predict the subcellular localization of the mature proteins ( without secretion signal ) to specific host cell organelles . Most proteins ( 60 ) were predicted to reside in the extracellular space . Here , a possible target could be the C . elegans cuticle collagen which plays an important role in barrier integrity and the immune response [38] . Nine putative effectors were predicted to enter nuclei . Only 37 of the 117 proteins could be annotated . Three had predicted pectin lyase domains , four were annotated as glycoside hydrolase , 10 as other enzymes , two with domains of unknown function ( DUF1524 ) . A similar DUF1524 motif is found in protein families in the His-Me finger endonuclease superfamily , which suggests a function as endonuclease . In another approach , the genome was analyzed for small secreted proteins ( SSP ) ( < 300 amino acid ) . SSPs are involved in pathogenesis in numerous fungi [39] and the expression of many SSPs was shown to be highly upregulated in A . oligospora and Da . haptotyla during infection of the host . The D . flagrans secretome comprises 249 SSPs , compared to 312 in the facultative nematode endoparasite Dr . coniospora and 695 in the nematode-trapping fungus Da . haptotyla [40 , 41] . 43 of the proteins identified with EffectorP were also identified as SSPs . Only 24 SSPs of D . flagrans had Pfam annotations in InterProScan . The lack of known homologs and Pfam domains of SSPs is consistent with observations in other fungi as Da . haptotyla , where rapid divergence of the SSPs was shown [42 , 43] . 14 SSPs contained at least five cysteines probably stabilizing them . Using the PHI database 110 of the 249 SSPs where predicted effectors or virulence factors that are likely to act outside the fungal cell or inside the host cell . In the same way as described for D . flagrans , the secretomes of 21 other fungi were predicted ( S4A Fig , S12 Table , S1 Text ) . One major difference between D . flagrans and the closely related A . oligospora is the ability of D . flagrans to form resistant chlamydospores for survival . Hence , the genome of D . flagrans should contain a number of genes not present in A . oligospora . To generate a list of candidate proteins we compared the proteomes of D . flagrans and A . oligospora . 591 proteins were exclusive to D . flagrans . Only 189 proteins resulted in database hits ( RefSeq and Swiss-Prot ) . A protein library of the 591 proteins was generated and compared to the proteomes of the four chlamydospore-forming fungi Botrytis cinerea , Cryptococcus neoformans , Fusarium oxysporum and P . chlamydosporia by protein database search ( BLASTP ) with 58 , 29 , 109 and 119 hits respectively . These proteins were then clustered using OrthoVenn ( S5 Fig ) . The 591 proteins from D . flagrans formed 81 clusters , including 8 orthologous clusters present in all five chlamydospore-forming fungi and correspond to 8 proteins in D . flagrans . One of the clusters did not have any Swiss-Prot hit or GO annotation , the others were annotated as uncharacterized protein YPR170W-B , lipase 2 , cytochrome b-c1 complex subunit 9 , protein transport protein GOT1 , protein GET1 , acyl-coenzyme A oxidase 2 and Dol-P-Man:Man ( 5 ) GlcNAc ( 2 ) -PP-Dol alpha-1 , 3-mannosyltransferase . Chlamydospores are enriched in glycogen ( see above ) , and it was shown that in Candida albicans dolichyl-phosphate beta-D-mannosyltransferase 1 ( Dpm1 ) and Dpm3 mutants are impaired in chlamydospore formation [44] . This could indicate a role of the D . flagrans Dol-P-Man:Man ( 5 ) GlcNAc ( 2 ) -PP-Dol alpha-1 , 3-mannosyltransferase DFL_006194 in chlamydospore formation . In order to further analyze the fungal-nematode interaction , genetic manipulations are required . As a first step , we adapted the protoplast preparation and transformation protocols from A . oligospora and Aspergillus oryzae for D . flagrans [45] . For this purpose , 24 h old D . flagrans liquid culture was used and prepared for protoplast formation and subsequent polyethylene glycol ( PEG ) / calcium chloride ( CaCl2 ) –mediated DNA uptake . For selection of positive transformants the three dominant markers hygromycin-B , nourseothricin and geneticin ( G418 ) were established . The drugs were efficient against protoplast regeneration and growth at 100 μg/ml hygromycin-B , 60 μg/ml nourseothricin or 120 μg/ml G418 . To establish a transformation method , protoplasts were transformed with plasmids ( pVW19 , pRF62 , pUMa1057 ) containing the hygromycin-B phosphotransferase gene hph , the nourseothricin acetyl transferase gene nat or the neomycin phosphotransferase gene neo [46] . Both the hph and nat resistance genes were under the control of the A . nidulans trpC promoter ( ptrpC::hph and ptrpC::nat ) . The neo resistance gene was under the control of the U . maydis o2tef promoter ( po2tef::neo ) . Mutants containing either of the resistance genes were able to grow on medium containing the corresponding drugs at concentrations where wild-type was inhibited ( Fig 4A–4F ) . The integration of the corresponding resistance genes into the genome was confirmed by PCR . Next , we tested the expression of green ( GFP ) and red fluorescent proteins ( mCherry ) . We tried corresponding genes , which were functional in N . crassa , in A . nidulans , in U . maydis , in C . albicans or in Botrytis cinerea . They were expressed using the A . nidulans glyceraldehyde-3-phosphate dehydrogenase ( gpdA ) promoter ( gpdA ( p ) ::GFP or gpdA ( p ) ::mCherry ) . Only the version adapted for B . cinerea worked ( Fig 4G–4H ) [47] . Non-transformed wild-type hyphae did not show any green or red fluorescence ( Fig 4I–4J ) . This is the first report for the function of fluorescent proteins in NTF that form adhesive trapping networks . To get more insights into the cell biology of D . flagrans , mCherry was C-terminally fused to histone H2B to stain nuclei in living cells . The expression of the fusion protein was under the control of the D . flagrans H2B native promoter . Five of seven transformants showed red fluorescent nuclei , which was confirmed using the nucleic acid dye Hoechst 33342 ( Fig 5A–5C ) . Genomic integration of the construct was verified using PCR . The expression of H2B-mCherry revealed multinucleated conidia and chlamydospores ( Fig 5D–5G ) . Interestingly , trap cells and hyphal compartments contained several nuclei ( Fig 5H–5J ) . In addition , the expression of H2B-mCherry enables to monitor the infection inside the nematode , in which nuclei were visualized by GFP ( Fig 5K ) . We used quantitative real time PCR analysis ( qRT PCR ) to identify genes that are up-regulated during infection of C . elegans ( 24 h co-incubation of the two organisms to induce trap formation ) and thus could play a role in virulence . One candidate , gene DFL_009915 ( putative effector protein B , pefB ) was chosen for further analyses ( Fig 3B ) . Biological replicates for the expression analysis varied significantly , with 16 . 1 ± 2 . 1 ( mean ± SD , n = 3 ) , 31 . 7 ( ± 4 . 1 ) and 81 . 1 ( ± 10 . 4 ) fold changes , respectively . The relative expression with gamma actin as reference gene showed no expression of pefB in the uninduced state ( Fig 3B ) . The large variance between replicates depends on differences in the number of traps and the number of worms trapped at certain times . The 274 amino acid long protein has a putative N-terminal signal peptide ( SP ) and a putative nuclear localization signal ( NLS ) ( Fig 3C ) . The corresponding gene does not contain introns . The protein is rich in serine ( 19 . 7% ) and arginine ( 15 . 3% ) and is highly positively charged ( isoelectric point = 12 . 02 ) . PefB does not contain any conserved domains . Analysis of the protein with the “Rapid Automatic Detection and Alignment of Repeats” ( RADAR ) [48] tool revealed three repeat structures within the protein ( Fig 3C ) . A BLAST search resulted in only two hits , revealing 47% identity to the hypothetical A . oligospora gene AOL_s00006g518 ( E-value = E-5−21 ) and 48% identity to a 168 amino acid long region of the 481 amino acid long hypothetical Paramormyrops kingsleyae serine/arginine-rich splicing factor 6-like-protein ( E-value = 0 . 22 ) ( XP_023660799 . 1 ) . Since secretion is one main criterium for effector proteins , or protein-based virulence factors , a laccase-assay was established for D . flagrans to prove the functionality of the SP . PefB was fused to the A . nidulans laccase C ( LccC ) lacking its own SP [49] . Laccases catalyze the oxidation of artificial substrates such as ABTS ( 2 , 2’-azino-bis- ( 3-ethylthiazoline-6-sulfonate ) in the medium to the more stable state of the cation radical . The enzyme activity therefore correlates with an intense blue-green color in the presence of ABTS . Hygromycin-resistant D . flagrans transformants indeed showed a positive reaction ( Fig 6A ) . Genomic integration of the construct was verified using PCR . These results show that the described secretion assay is a suitable method for the verification of predicted protein secretion . The laccase may also be used for other purposes in D . flagrans . In order to test the functionality of the NLS , the protein ( without the signal peptide ) was N-terminally fused with GFP and expressed in D . flagrans . The GFP-PefB fusion protein was visible in the cytoplasm and accumulated in nuclei ( Fig 6B ) . Nuclear localization in D . flagrans should however , not be the natural residence of PefB , because the protein should be secreted during growth in C . elegans . Therefore , we tested whether the NLS was functional in C . elegans host cells ( Fig 6C ) . A C-terminal PefB-GFP fusion protein without signal peptide was expressed in C . elegans . Similar to the localization of PefB in D . flagrans , PefB-GFP localized in spots most likely corresponding to nuclei around the terminal bulb of the pharynx . In the case of the two large nuclei below the pharynx we were able to co-localize the protein with DAPI . The other green spots could not be co-localized , probably because the GFP signal was too weak after fixation for DAPI staining . The smaller spots are probably nuclei of neurons or glia cells , whereas the large ones are most likely from intestinal cells ( http://wormatlas . org ) . Future experiments have to reveal if these are the target cells under natural conditions when D . flagrans colonizes the worm . Nevertheless , the experiment clearly shows that the NLS is functional in C . elegans . One fascinating aspect of D . flagrans and A . oligospora is the ability to form three-dimensional adhesive networks consisting of hyphal rings to trap nematodes . From the cell biological point of view this is not an easy task , since a network is the result of multiple cell fusions ( anastomoses ) of hyphal loops [50] . During trap formation a first loop develops as a branch on a vegetative hypha . The loop mostly consists of three hyphal compartments . Prior to ring closure and the fusion of the hyphal tip with the basal hypha , a small hyphal peg is growing towards the trap-forming hypha , indicating intercellular communication . Such a mechanism of cell-to-cell signalling is observed in many filamentous fungi during germling fusion or hyphal anastomosis . A mitogen-activated protein kinase ( MAPK ) cascade , homologous to the pheromone response pathway of Saccharomyces cerevisiae , plays a central role in this communication [51] . In order to investigate the role of hyphal anastomoses during trap formation , similarity searches against various proteins of the N . crassa signalling cascade allowed the reconstruction of the pathway in D . flagrans ( S7 Fig , S13 Table ) [52] . To test the hypothesis that hyphal anastomosis and cell-to-cell communication is a prerequisite for trap formation , an orthologue of the N . crassa hyphal anastomosis gene soft ( so ) was investigated . The open reading frame of the corresponding D . flagrans gene ( DFL007867 ) consists of 3867 base pairs ( bp ) with four introns ( accession no . MK034754 ) . The introns with lengths of 53 bp , 55 bp , 57 bp , and 60 bp respectively , were confirmed by cDNA cloning and sequencing . They were also in accordance with RNAseq data . The gene was named sofT . We are using the three-letter gene name code in D . flagrans . The SofT protein consists of 1213 amino acids with a predicted molecular mass of 133 kDa . SofT displayed 56% identity with the SO protein of N . crassa , between 44 and 60% identity with Aspergillus oryzae , Epichloe festucae , Dre . coniospora , P . anserina , A . brassicicola , F . verticillioides , F . oxysporum , and 94% with the putative A . oligospora orthologue ( AOL_s00076g148 ) . All SofT proteins harbour a conserved WW domain , which is characterized by two conserved tryptophan ( W ) residues spaced by 20 to 22 amino acids . To investigate the biological function of SofT during trap formation , we established a gene knock-out system for D . flagrans based on homologous recombination ( Fig 7A and 7B ) . 20 transformants were obtained growing on PDA with 100 μg/ml hygromycin-B . They were screened via PCR and Southern blot analysis ( Fig 7C ) . One transformant showed the expected integration . In order to confirm the phenotype of the ΔsofT-deletion strain , and to establish the CRISPR/Cas9 system in D . flagrans , we aimed at achieving gene inactivation with this system . First , we tried the plasmid-based approach established for A . nidulans , A . alternata and other fungi [53 , 54] . However , no stable transformants were obtained . Next , we tested simultaneous transformation of Cas9 ribonucleoproteins ( RNPs ) and the linear resistance gene hph based on recent reports [55–57] . The sgRNA cleavage site was located in the first exon of the sofT gene ( Fig 7D ) . To avoid off-target cleavage , the uniqueness of the protospacer region and the protospacer adjacent motif ( PAM ) sequence were analysed in the D . flagrans genome using BLASTN . In total four transformants were obtained and the region surrounding the protospacer region was sequenced to analyse the mutations . In three of the four transformants , the linear resistance gene hph was inserted into the cleavage site . In addition , small deletions of one or two nucleotides were detected in the ΔsofT mutants . Interestingly the 5’ and 3’ ends of the inserted hph gene were slightly modified by nucleotide deletions . Transformant T3 showed single integration of the hph gene into the genome as shown by Southern blot analysis and was used for further experiments ( Fig 7E and 7F ) . All ΔsofT mutant strains grew with less aerial hyphae and trap formation was disturbed at the step of ring closure , resulting in spiral hyphae ( Fig 8A–8D ) . Anastomoses in normal vegetative hyphae were completely inhibited ( Fig 8E ) . Despite the observed defects , the ΔsofT-deletion strain formed complex three-dimensional trapping networks and was still able to trap C . elegans ( Fig 8F ) . Next , the mutant generated by the CRISPR/Cas9 system was transformed with a plasmid containing the full-length sofT gene including the regulatory regions ( promoter and terminator ) . In two out of six transformants wild-type phenotype was restored . The integration of the construct was verified by PCR . A virulence assay was performed to further investigate the lack of SofT on the virulence of D . flagrans . The nematicidal activity of the fungi was divided into captured ( alive ) or already digested ( dead ) nematodes after 24 and 36 h ( Fig 8G and 8H ) . After 24 h the wild-type captured 10 ± 4 ( mean ± SD ) % while 19 ± 5% were already digested . In the ΔsofT mutant 5 ± 2% ( P = 0 . 1658 unpaired t test ) were captured while 7 ± 4% ( P = 0 . 0235 unpaired t test ) were already digested after 24 h . After 36 h in the wild-type 5 ± 2% were captured while 65 ± 14% were already digested . In the ΔsofT mutant 1 ±1% ( P = 0 . 0514 unpaired t-test ) were captured while 35 ± 12% ( P = 0 . 0476 unpaired t-test ) were already digested . To exclude that the attenuated virulence was due to vegetative growth or fitness defects , germination time and fungal growth were analysed of the mutant ( Fig 8I ) . Wild-type conidia germinated on LNA after 2 . 5 ± 0 . 4 h ( mean ± SD , n = 6 ) while ΔsofT conidia germinated after 2 . 9 ± 0 . 6 h ( P = 0 . 1658 unpaired t test , n = 6 ) . Growth of wild-type hyphae on LNA was 0 . 71 ± 0 . 09 μm/s ( n = 6 ) while the growth of the ΔsofT hyphae was 0 . 68 ± 0 . 08 μm/s ( P = 0 . 5209 unpaired t test , n = 6 ) . Hence , germination and hyphal growth were not very different in wild type and mutant . The reduced virulence of the ΔsofT mutant may be due to decreased stress tolerance . It was shown that the A . oryzae SO orthologue AoSO forms aggregates at the septal pore in response to physical stress , low and high temperature and extreme acidic or alkaline pH [58] . In addition , the AoSO mutant strain has defects in stress granule formation during exposure to heat stress [59] . Thus , the ΔsofT D . flagrans mutant may be more sensitive towards stress . The innate immune response of C . elegans is accompanied by an increase of reactive oxygen species ( ROS ) , to cope with bacterial and fungal pathogens attacking the worm from the intestine or the cuticle [60] . In the insect pathogenic fungus Metarhizium acridum , it has recently been shown that a catalase peroxidase , MakatG1 , is expressed specifically on the host cuticle [61] . D . flagrans SofT could play a role in the signalling pathway of the ROS-induced stress response .
D . flagrans ( CBS 349 . 94 ) was obtained from the CBS-KNAW culture collection ( The Netherlands ) and was cultured at 28°C on PDA . Cultivation and synchronization of C . elegans was performed according to the worm book ( doi/10 . 1895/wormbook . 1 . 101 . 1 ) . X . index was isolated from fig tree roots using a modified Baermann funnel [29] . To induce trap formation around 1x104 D . flagrans spores were inoculated on thin LNA ( pH 6 . 5 ) and around 100 individuals of C . elegans were added . Co-incubation was carried out at room temperature in darkness for at least 6 hours . Experiments with X . index were performed on LNA plates where 10 individuals were co-incubated with 5x104 D . flagrans spores at room temperature . To induce trap formation for RNA extraction 106 D . flagrans spores were incubated on LNA with 1% glucose covered with a cellophane membrane for 24 h at 28°C . 104 individuals of a mixed C . elegans population were added to the membrane and co-incubated for 24 h to induce trap formation . The uninduced group was treated equally without the addition of nematodes . Glycogen of D . flagrans hyphae , conidia and chlamydospores was stained with Lugols’s iodine ( Roth ) . Nuclei were stained using Hoechst 33342 ( 5 μg/ml working concentration , incubation time 5 min , Invitrogen , Karlsruhe , Germany ) or Vectashield antifade mounting medium with DAPI ( Vector Laboratories , Peterborough , UK ) . C . elegans nuclei were stained using Vectashield antifade mounting medium with DAPI after fixation with 95% ethanol [62] . The cell wall was stained with Calcofluor white M2R ( fluorescent brightener 28 , Sigma-Aldrich ) . A 1% ( wt/vol ) stock solution was prepared by dissolving CFW in 0 . 5% ( wt/vol ) KOH and 83% ( vol/vol ) glycerol . The stock solution was diluted 1:1000 ( dH2O ) and incubated for 5 min to stain the cell wall . For extraction of genomic DNA , D . flagrans was grown on CM-agar for 7 days . Spores were collected in 100 ml CM liquid medium and incubated at 28°C for 24 h at 180 rpm . Protoplasts were generated as described above and genomic DNA was extracted using Genomic DNA Buffer Set ( Qiagen ) and Genomic-tip 20/G ( Qiagen ) according to the manufacturer’s instruction . In total 12 μg of DNA were sent for sequencing . The D . flagrans genome was sequenced using the PacBio RS technology ( Pacific Biosciences , Menlo Park , CA , USA ) with libraries prepared with the SMRTbell template prep kit 1 . 0 ( Pacific Biosciences ) . The sequencing runs and assembly of the libraries were carried out by GATC biotech ( Konstanz , Germany ) . The Whole Genome Shotgun project has been deposited at DDBJ/ENA/GenBank under the accession SAEB00000000 . The version described in this paper is version SAEB01000000 . Raw reads were deposited at SRA under the accession SRR8400569 . Total RNA was extracted using the E . Z . N . A . Fungal RNA Kit ( OMEGA bio-tek ) according to the manufacturer’s protocol . Trap induction for the zootrophic stage was induced as described above using LNA without glucose . For saprotrophic growth , around 5x106 D . flagrans spores were inoculated in PDM in 3 . 5 cm petri dishes and were cultivated for 48 h at 28°C . The extracted RNA was checked using a NanoDrop 2000 ( Thermo , CA , USA ) , and the RNA concentration and integrity were assessed using the RNA Nano 6000 Assay Kit of the Agilent Bioanalyzer 2100 system ( Agilent , CA , USA ) to ensure that the RNA Integrity Number ( RIN ) values were above 7 . 0 . Oligo ( dT ) beads were used to isolate poly ( A ) + mRNA , which was fragmented to 250 bp . Fragmentation of the RNA and reverse transcription of double-strand cDNA ( ds cDNA ) was done by using N6 random primers . The synthesized cDNA was subjected to end-repair and then 3′ adenylated . Adaptors were ligated to the cDNA ends . After amplification each cDNA library was sequenced in a single lane of the BGISEQ-500 system using paired-end sequencing with lengths of 100 bp according to the manufacturer’s instructions at the Beijing Genomics Institute ( Beijing , China ) . 23 , 941 , 723 reads were generated and 91 . 79% mapped to the D . flagrans genome and 88 . 84% had a unique match . 85 . 28% of the reads were mapped to genes , 85 . 01% had a unique match . Gene models were predicted with the ab initio predictor BRAKER2 [63–65] . BRAKER2 was implemented with GeneMark-EX and subsequently with AUGUSTUS . Additionally , RNAseq was included to improve intron predictions . The D . flagrans proteome was blasted against the Swissprot using an E-value limitation of E-10−5 . A . oligospora was used as database for BLAST searches of D . flagrans using an E-value limitation of E-10−5 . Protein subcellular localization was predicted using WoLF Psort [66] . The proteomes of A . oligospora , Dre . coniospora , Da . haptotyla , and Dr . stenobrocha were downloaded from NCBI and blasted against the D . flagrans proteome using at least 70% amino acids identity and 80% coverage to D . flagrans proteins in order to identify orthologous proteins . On the other hand , orthologous protein families were determined among D . flagrans , A . oligospora , Da . haptotyla and Dr . stenobrocha in OrthoVenn . Orthologous gene pairs were considered on the basis of amino acid sequence similarities sharing up to 50% of the total length of the shorter gene being analysed ( BLASTP , threshold E-value ≤ 1E-10−5 ) . Signal peptides were predicted using SignalP v4 . 1 [67] . Secreted proteins were further predicted using WoLF PSORT , a protein localization prediction program trained with fungal data . Proteins that may be secreted but probably membrane bound were filtered out using TMHMM v2 . 0 [68 , 69] . Proteins combining these criterions were joined to generate the secretome . The tools were used at Galaxy Uni-Freiburg [70 , 71] and filtration was implemented by a custom Perl script . Nuclear localization was predictied using NLStradamus [72] . The secretomes of A . oligospora ( ADOT00000000 . 1 ) , Fusarium graminearium ( AACM00000000 . 2 ) , Aspergillus nidulans ( AACD00000000 . 1 ) , D . coniospora ( LAYC00000000 . 1 ) , Dr . stenobrocha ( ASQI00000000 . 1 ) , Dr . stenobrocha ( AACP00000000 . 2 ) , Neurospora crassa ( AABX00000000 . 3 ) , Dactylellina haptotyla ( AQGS00000000 . 1 ) , Penicillium brasilianum ( CDHK00000000 . 1 ) , Aspergillus fumigatus ( AAHF00000000 . 1 ) , Candida albicans ( GCA_000182965 . 3 ) , Candida galabrata ( GCF_000002545 . 3 ) , Alternaria alternata ( LXPP00000000 . 1 ) , Magnaporthe oryzae ( AACU00000000 . 3 ) , Trichoderma reesei ( AAIL00000000 . 2 ) , Pochonia chlamydosporia ( LSBJ00000000 . 2 ) , Botrytis cinera ( GCA_000143535 . 4 ) , Metarhizium acridum ( ADNI00000000 . 1 ) , Metarhizium album ( Metarhizium album ) , Hirsutella minnesotensis ( JPUM00000000 . 1 ) , Cordyceps militaris ( AEVU00000000 . 1 ) proteomes were generated on the same way as D . flagrans . Venn diagrams were generated using VennDiagram package in R . The transformation system was adapted from [45] . Briefly 1x108 spores were inoculated in 150 ml PDM and incubated for 24 h at 28°C and 180 rpm . The mycelium was harvested and washed with MN solution ( 0 . 3 mol/l MgSO4 , 0 . 3 mol/l CaCl2 ) and ca . 0 . 5 g of mycelia ( wet weight ) was collected and suspended in 5 ml MN buffer containing 4 mg/ml kitalase ( Fujifilm Wako Chemicals ) and 20 mg/ml VinoTaste Pro ( Novozymes ) followed by incubation at 30°C , 70 rpm for 3 h . Quality of protoplasts was checked using light microscopy . Subsequently undigested mycelium was removed by filtering the protoplast through 3-layer of miracloth tissue . The protoplasts were precipitated at 3350 g for 15 min . After carefully removing the supernatant the protoplasts were washed with 50 ml KTC buffer ( 1 . 2 mol/l KCl , 10 mmol/l Tris-HCl pH 7 . 5 ) and the resulting pellet was resuspended in 500 μl KTC . For transformation 100 μl protoplasts ( 5x106 ) were mixed with 5–8 μg of DNA and incubated for 2 min on ice . Then 1 ml of PTC ( 10 mmol/l Tris-HCl pH 7 . 5 , 50 mmol/l CaCl2 , 60% W/V PEG6000 ) was added and incubated at room temperature for 20 min . 10 ml PDSSA ( 24 g/l potato dextrose broth , 0 . 6 mol/l sucrose , 0 . 3 g/l peptone , 0 . 3 g/l tryptone , 0 . 3 g/l yeast extract , 14 g/l agar ) ( lukewarm ) was added to the transformation mix and was poured onto PDA plates supplemented with 100 μg/ml hygromycin-B , 60 μg/ml nourseothricin , or 120 μg/ml geneticin ( G418 ) , respectively and incubated at 28°C for 4–7 days . The sgRNA for the sofT gene mutation was synthesized using the EnGen sgRNA Synthesis Kit ( New England Biolabs , Frankfurt ) . Ribonucleoprotein ( RNP ) formation was carried out at 37°C for 20 min with 4 μl unpurified sgRNA , 10 μl Cas9 ( 200nM ) , 5 μl 10x Cas9 Nuclease Reaction Buffer and DEPC-treated water in 50 μl total volume . For protoplast transformation , the linear resistance gene hph was amplified by PCR , subsequently purified using FastGene Gel/PCR Extraction kit ( Nippon Genetics ) and lastly about 2 μg purified PCR-product were added to the transformation-mix . The transformation was carried out as described . To create a C-terminal mCherry fusion of the histone H2B under its natural promoter , a 1 . 6 kb fragment including the ORF of DFL_000203 ( 0 . 6 kb without stop codon ) + 1 kb of the 5’ region and a 1 kb fragment 3’ downstream of the locus were amplified by PCR , using D . flagrans genomic DNA as template . The mCherry gene from plasmid 12A_pNDH-OCT [59] and the hph from plasmid pFC332 [53] were amplified by PCR . All fragments were assembled into the pJET1 . 2 vector ( Thermo Fisher , digested with EcoRV ) using the NEBuilder HiFi DNA Assembly Cloning Kit ( New England Biolabs , Frankfurt ) resulting in plasmid pVW23 . For the N-terminal GFP fusion protein of DFL_009915 ( PefB ) without the signal peptide the gene was amplified with a forward primer starting at bp 60 and containing an AscI restriction site and start codon and a reverse primer containing a PacI restriction site . The amplicon was cut with the respective enzymes and ligated into a plasmid that contains the hygromycin-B resistance cassette hph [53] and the GFP gene under control of the constitutive A . nidulans oliC promotor and the glucanase terminator of B . cinerea ( oliC ( p ) ::GFP::pefB::gluC ( T ) ) with a pJET1 . 2 backbone resulting in the plasmid pIH02 . For the ABTS-assay , the modified vector pOF018 was used for the expression of the A . nidulans lccC gene ( AspGD identification AN5397 ) under the constitutive A . nidulans glyceraldehyde-3-phosphate dehydrogenase ( gpdA ) promoter and as positive control for laccase activity [71] . For the selection of D . flagrans the hygromycin B resistance cassette was amplified using pFC332 [53] as template and inserted into the vector after linearization with XbaI , using the NEBuilder HiFi DNA Assembly Cloning Kit [73] . The gene sequence of pefB was amplified with primers adding overlapping regions to the vector . The vector backbone was amplified via PCR with primers binding after the signal peptide sequence ( 60 bp ) to remove the laccase signal peptide . The pefB gene sequence was fused to the 3’ end of the laccase C using the NEBuilder HiFi DNA Assembly Cloning Kit . For the expression of PefB in C . elegans , the open reading frame was cloned , without the predicted signal peptide , into the pLZ29 backbone [74] using the Gibson assembly method [73] . The resulting assembly made an in-frame C-terminally tagged GFP fusion protein under expression of the ubiquitous eft-3 promoter . The sofT gene was deleted by homologous recombination . Two kb flanks homologous to the 5’ and 3’ regions of the sofT gene were amplified by PCR , using D . flagrans genomic DNA as template . The hygromycin-B resistance cassette hph was amplified using pFC332 [53] as template . The PCR-fragments were assembled into the pJET1 . 2 vector ( Thermo Fisher , digested with EcoRV ) using the NEBuilder HiFi DNA Assembly Cloning Kit ( New England Biolabs , Frankfurt ) . The fragments were amplified with Phusion polymerase ( NEB ) using the manufacturers recommended protocol and contained 25 bp overlapping regions to the neighbouring fragment . Standard protocols were used for E . coli transformation and plasmid isolation [75] . To confirm the sofT introns , ca . 5x106 D . flagrans spores were inoculated in PDM in 3 . 5 cm petri dishes and were cultivated for 48 h at 28°C . The mycelium was harvested and frozen immediately in liquid nitrogen . RNA isolation was performed using E . Z . N . A . Fungal RNA Kit ( OMEGA bio-tek ) according to the manufacturer protocol . To remove DNA the RNA was treated with TURBO DNA-free kit ( Invitrogen , Karlsruhe , Germany ) . cDNA synthesis was performed using the SuperScript Double-Stranded cDNA Synthesis Kit ( Thermo Fisher Scientific , Waltham , MA , USA ) according to the manufacturer protocol . The sofT coding sequence was amplified by PCR using D . flagrans WT cDNA as template and assembled into the pJET1 . 2 vector . Sequencing was carried out by MWG-Eurofins . To reintroduce the functional D . flagrans sofT gene into the ΔsofT deletion strain , the whole sofT gene , including its 1-kb upstream and 0 . 5 downstream regulatory regions , was amplified via PCR , using D . flagrans genomic DNA as template . To select for transformants the geneticin G418 resistance cassette was used . In a cloning step , the G418 resistance cassette was assembled between a new trpC ( p ) promoter and a new trpC ( t ) terminator fragment of pVW23 using the NEBuilder HiFi DNA Assembly Cloning Kit . Lastly , the new G418 resistance cassette and the sofT complement were assembled into the pJET1 . 2 vector ( Thermo Fisher , digested with EcoRV ) using the NEBuilder HiFi DNA Assembly Cloning Kit . All primers used for plasmid constructions and sequencing are listed in S14 Table . The plasmid harboring the pefB::GFP fusion construct was injected at a concentration of 50ng/ul into wildtype worms ( N2 ) with a pharyngeal co-injection marker ( myo-2p::tdTomato ) and co-injection marker positive transformants were selected . For RNA extraction the mycelium was scratched off the cellophane membrane and ground in liquid N2 . Total RNA was extracted with Trizol reagent ( Invitrogen , Karlsruhe , Germany ) . DNase digestion was performed using the Turbo DNA-free Kit ( Invitrogen , Karlsruhe , Germany ) and the RNA was diluted to 50 ng/μl . The SensiFast SYBR and fluorescein One Step Kit ( Bioline , Luckenwalde , Germany ) was used for the qRT-PCR analysis on an CFX Connect Real-Time PCR Detection System ( Bio-Rad , Munich , Germany ) . Each reaction mixture contained 0 . 2 μM primers ( S14 Table ) and 80 ng of RNA in a 25-μl total volume . Melting curve analysis was performed to assess the specific amplification of DNA . Fold changes were calculated using the formula 2− ( ΔΔCt ) , with ΔΔCt being ΔCt ( treatment ) -ΔCt ( control ) , ΔCt is Ct ( target gene ) —Ct ( actin ) , and Ct is the threshold cycle . The gamma actin orthologue DFL_002353 was used as internal reference gene for normalization . The qRT PCR was performed using 3 biological replicates . For the ABTS-assay 1x104 spores of D . flagrans transformands were grown on a modified low-nutrient-agar ( LNA ) modified after [73] ( KCL 1g/L , MgSO4 - 7H2O 0 . 2g , MnSO4 - 4H2O 0 . 4mg , ZnSO4 - 7H2O 0 . 88mg , FeCl3 - 6H2O , 3mg , Agar 10g , pH 5 . 5 ) containing 1 mM ABTS ( 2 , 2‘-azino-bis-[3-ethylthiazoline-6-sulfonate] ) at 28°C . Laccase activity was correlated with the formation of blue-green ABTS• visible after 24–48 h . For microscopy around 4x104 spores were inoculated on thin LNA and incubated for at least 14 h at 28°C . Conventional fluorescence images were captured at room temperature using a Zeiss Plan-Apochromat 63x/1 . 4 Oil DIC , EC Plan-Neofluar 40x/0 . 75 , EC Plan-Neofluar 20x/0 . 50 , or EC Plan-Neofluar 10x/0 . 30 objective attached to a Zeiss AxioImager Z . 1 and AxioCamMR . Color images were acquired using the AxioCam 105 color . Images were collected using ZEN 2012 Blue Edition . Confocal images were captured at room temperature using a Leica HCX PL APO 63x1 . 4 oil objective attached to a Leica TC SP5 and conventional photomultiplier tube detectors ( Leica , Wetzlar , Germany ) . Kryo-SEM was performed as previously described [72] . Stereomicroscopy was performed using a Zeiss Lumar . V12 with AxioCam HRc and NeoLumar S 1 . 5x objective . Images were collected using the AxioVision software . To quantify germination time and hyphal growth of D . flagrans wild-type and the ΔsofT strain an AxioObserver Z1 inverted microscope employing a 10x/0 . 30 N . A . objective ( Zeiss ) was used . To compare the virulence of the D . flagrans wild-type and the ΔsofT mutant strain 1x105 spores were plated on a petridish ( 3 . 5 cm diameter ) containing LNA . As bait 200 synchronized C . elegans L2 larvae ( strain BAN126 ) were used and added to the plate and incubated at room temperature . After 24 and 36 h the number of captured or already digested nematodes was counted using a light microscope ( Zeiss AxioImager Z . 1 ) . The assay was performed with three biological replicates . To determine the germination time and hyphal growth of the D . flagrans wild-type and the ΔsofT mutant 1x104 spores were plated on thin LNA . The experiments were carried out at 28°C . The hyphal growth was calculated by kymograph analysis using the Fiji software ( https://fiji . sc ) . | Nematode-trapping fungi are fascinating microorganisms , because they are able to switch from saprotrophic growth to a predatory lifestyle . Duddingtonia flagrans forms adhesive trap systems and conidia and resistant chlamydospores . Chlamydospores are ideal for dissemination in the environment to control nematode populations in the field . We show that D . flagrans is able to catch C . elegans but also the very large wine-pathogenic nematode Xiphinema index . We sequenced the D . flagrans genome and show that it encodes about 10 , 000 genes with a large proportion of secreted proteins . We hypothesize that virulence effector proteins are involved in the interkingdom organismic interaction and identified more than 100 candidates . In order to investigate the molecular biology of D . flagrans and its interaction with nematodes , we established a transformation system and several molecular tools . We show that cell-to-cell communication and hyphal fusion are required for trap formation . Finally , we show that one putative virulence effector protein targets nuclei when expressed in C . elegans . | [
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Materials",
"and",
"methods"
]
| [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"caenorhabditis",
"fungal",
"genetics",
"rna",
"extraction",
"parasitic",
"diseases",
"animals",
"nematode",
"infections",
"animal",
"models",
"fungi",
"caenorhabditis",
"elegans",
"model",
"organisms",
"experimental",
"organism",
"systems",
"extraction",
"techniques",
"research",
"and",
"analysis",
"methods",
"mycology",
"animal",
"studies",
"proteins",
"biochemistry",
"fungal",
"genomics",
"eukaryota",
"proteomes",
"genetics",
"nematoda",
"biology",
"and",
"life",
"sciences",
"protein",
"domains",
"genomics",
"organisms"
]
| 2019 | Intercellular communication is required for trap formation in the nematode-trapping fungus Duddingtonia flagrans |
In eukaryotic cells , most mRNAs are exported from the nucleus by the transcription export ( TREX ) complex , which is loaded onto mRNAs after their splicing and capping . We have studied in mammalian cells the nuclear export of mRNAs that code for secretory proteins , which are targeted to the endoplasmic reticulum membrane by hydrophobic signal sequences . The mRNAs were injected into the nucleus or synthesized from injected or transfected DNA , and their export was followed by fluorescent in situ hybridization . We made the surprising observation that the signal sequence coding region ( SSCR ) can serve as a nuclear export signal of an mRNA that lacks an intron or functional cap . Even the export of an intron-containing natural mRNA was enhanced by its SSCR . Like conventional export , the SSCR-dependent pathway required the factor TAP , but depletion of the TREX components had only moderate effects . The SSCR export signal appears to be characterized in vertebrates by a low content of adenines , as demonstrated by genome-wide sequence analysis and by the inhibitory effect of silent adenine mutations in SSCRs . The discovery of an SSCR-mediated pathway explains the previously noted amino acid bias in signal sequences and suggests a link between nuclear export and membrane targeting of mRNAs .
In eukaryotes , mRNAs are synthesized and processed in the nucleus before they are transported through the nuclear pores into the cytoplasm , where they are translated into proteins . Nuclear export of most mRNAs is mediated by the conserved transcription export ( TREX ) complex that is comprised of the Tho complex , UAP56 , and Aly . In vertebrates , the TREX components are recruited to the 5′ end of newly synthesized transcripts by the combined action of the 5′ cap binding complex , CBP80/20 , and factors that are loaded during the splicing of the intron closest to the 5′ cap [1–4] . Once assembled , the TREX complex recruits the heterodimer TAP/p15 as an export factor [5 , 6] . TAP interacts with nucleoporins directly [7–9] or through the factor Rae1 [10 , 11] and may thus allow bound transcripts to enter and eventually pass through the nuclear pores . It remains unclear how mRNAs exit the pores on the cytoplasmic side , but RNA helicases , such as Dbp5 , may be involved [12 , 13] . Although many details of the export mechanism remain to be clarified , it seems clear that the efficient export of most mRNAs requires both splicing and a functional cap . Not all mRNAs follow this canonical export pathway . In higher eukaryotes , transcripts coding for cyclin D [14] and other regulators of cell division [15] use elements in the 3′ untranslated region ( UTR ) as well as the cap binding protein eIF4E to engage the exportin protein Crm1 . A Crm1-dependent pathway is also used for the export of the intron-containing RNA genome of the human immunodeficiency virus ( HIV ) [16] . In macrophages , the export of interferon-induced transcripts is sensitive to the levels of Nup96 , a component of the nuclear pore , whereas other transcripts are insensitive [17] . In Saccharomyces cerevisiae , the export of some mRNAs requires either the Aly ortholog Yra1p or the TAP ortholog Mex67p , but not both [18] . Transcripts coding for heat-shock proteins , such as Hsp70p , are also exported by an Aly-independent pathway that is stimulated under stress conditions [19] . Together , these data indicate that there may be distinct export pathways for certain classes of mRNAs , which may be related to the different functions of the final translation products . One specialized class of mRNAs codes for secretory proteins . These mRNAs are often translated by ribosomes that are targeted to the endoplasmic reticulum ( ER ) membrane . As a result , the mRNAs also become membrane-bound . In addition , it is possible that these mRNAs associate with the ER membrane in a ribosome-independent manner by interacting with RNA-binding proteins [20–22] . These proteins may be loaded onto the mRNAs in the nucleus , during their passage through the nuclear pores , or in the cytoplasm . Thus , the membrane localization of these mRNAs may require factors that mediate nuclear export and distribution in the cytoplasm , which are distinct from factors used by mRNAs translated on free ribosomes in the cytoplasm . A major characteristic of a secretory protein is a hydrophobic signal sequence close to its N terminus . The signal sequence is first recognized by the signal-recognition particle ( SRP ) as it emerges from the translating ribosome and is then transferred into the protein-conducting channel formed by the Sec61p complex [23–25] . In most cases , the signal sequence is cleaved during or shortly after the polypeptide is transferred into the ER . Although the major requirement for a signal sequence is a stretch of at least 6–7 hydrophobic amino acids , there appear to be other , poorly defined properties that may distinguish different signal sequences [26 , 27] . For example , some signal sequences require auxiliary translocation components [28 , 29] or are sensitive to the drug cotransin [27 , 30] . In addition , signal sequences have other unexplained characteristics . For example , it has been noted that signal sequences in humans tend to be rich in leucine and poor in isoleucine [31] , despite the fact that these two amino acids have similar hydrophobicities [32] and would thus be expected to function equally well to promote translocation . This leucine/isoleucine bias is not seen in prokaryotes . Another puzzling feature is that signal sequences are often conserved across species to a higher degree than expected [33] . These observations raise the possibility that not only the signal sequence per se , but also the encoding nucleotide sequence ( signal sequence coding region or SSCR ) may have a function . We report here on the nuclear export of mRNAs coding for secretory proteins . We used microinjection of mRNAs and DNAs as well as transfection of DNA to demonstrate that the SSCR can promote the export of mRNAs that cannot use the conventional pathway because they lack an intron or a functional cap . By using large-scale sequence analysis , we found that vertebrate SSCRs have a low content of adenine , a feature that may in part be responsible for the preference of leucine versus isoleucine in signal sequences . Consistent with these observations , the incorporation of silent adenine mutations within the SSCR inhibits its nuclear export activity . Even the export of a natural RNA containing introns is facilitated by its SSCR . The discovery of an SSCR-mediated mRNA export pathway thus explains the previously noted amino acid bias in signal sequences and suggests a link between nuclear export and membrane targeting of mRNAs .
To study the nuclear export of mRNA coding for a secretory protein , we used a model mRNA that is derived from a fragment of the fushi tarazu ( ftz ) gene . The original construct contains an intron and was previously used to monitor mRNA splicing and nuclear export in Xenopus oocytes [1 , 34] . The construct was modified by adding a Kozak consensus sequence to allow efficient expression in mammalian cells . Sequences encoding FLAG and hemagglutinin ( HA ) epitopes were included at the 5′ and 3′ ends of the open reading frame ( ORF ) , respectively , to monitor translation of the mRNA . Because the intron contains in-frame stop codons ( Figure 1A; asterisks ) , the HA epitope will only be synthesized if the mRNA is spliced . To target the translation product to the ER , we attached an SSCR derived from the mouse major histocompatibility complex ( MHC ) class 2 molecule H2-K1 . The final construct is called t-ftz-i ( Figure 1A and Figure S1 ) . We first tested the translation of the t-ftz-i mRNA in vitro . When translated in reticulocyte lysate , a polypeptide of 13 kDa was generated , consistent with the size expected from the location of the in-frame stop codon in the intron ( Figure 1B , lane 3 ) . When the intron was deleted , the resulting mRNA ( t-ftz-Δi ) gave rise to a 17-kDa translation product ( lane 4 ) , again in agreement with the expected size of the ORF . With the original ftz constructs that lacked translation initiation signals , no translational products were detected ( lanes 1 , 2 ) . When t-ftz-i mRNA was microinjected into nuclei of NIH 3T3 fibroblasts , it was efficiently spliced within 15 min , as shown by reverse-transcriptase ( RT ) -PCR ( Figure 1C; lanes 2–4 ) . As expected , no detectable splicing was observed when t-ftz-i was microinjected into the cytoplasm ( lane 5 ) . Next , we tested the translation of t-ftz-i transcripts that were microinjected into the nuclei of NIH 3T3 fibroblasts . Nuclear injection was confirmed by co-injecting fluorescently labeled 70-kDa dextran , which is too large to passively cross the nuclear pores ( Figure 1D and 1E; see insets ) . After 4 h , the FLAG and HA epitopes could be detected by immunofluorescence microscopy in over 90% of the injected cells ( Figure 1D and 1E ) . In addition , both epitopes co-localized with the ER resident protein TRAPα ( Figure 1D and 1E ) , indicating that the translation product was translocated into the ER . The protein is probably not secreted efficiently , because it contains only a fragment of ftz and therefore may not be properly folded . When t-ftz-i transcripts were injected into the cytoplasm , the expression of the FLAG but not of the HA epitope was observed ( unpublished data ) , as expected from the presence of the unspliced intron . Both the FLAG and HA epitopes were expressed when t-ftz-Δi transcripts were injected into nuclei or cytoplasm ( unpublished data ) . To monitor the nuclear export of t-ftz-i mRNA , transcripts were microinjected into nuclei of NIH 3T3 cells . The cells were fixed at various time points , and the localization of the injected RNA was probed by fluorescence in situ hybridization ( FISH ) . We optimized the FISH procedure , omitting harsh acid and ethanol treatments , such that intracellular morphology was largely maintained . We estimate that 20 , 000 to 50 , 000 transcripts were injected per cell , which is a small number compared with the total number of transcripts in a typical mammalian cell ( 400 , 000 to 850 , 000 molecules ) [35] . Again , fluorescent 70-kDa dextran was co-injected to identify nuclear-injected cells ( Figure 2 , insets ) . Immediately after injection , the transcripts were confined to the nucleus . Over time , however , the majority of t-ftz-i transcripts ( ∼80% ) accumulated in the cytoplasm ( Figure 2A ) . Quantitation showed that the half-time of mRNA export was ∼15 min ( Figure 2C ) , similar to previous estimates [36 , 37] . About 50% of the mRNA molecules remained intact after 4 h , as shown by the total level of FISH signal ( Figure 2D ) . Surprisingly , when we injected t-ftz-Δi mRNA into nuclei , we also observed efficient export into the cytoplasm ( Figure 2B; quantitation in Figure 2C ) , despite the fact that the absence of an intron should have prevented the recruitment of TREX components and thus nuclear export [2 , 4] . The kinetics of mRNA export was about the same for transcripts containing or lacking the intron ( Figure 2C ) . In contrast to a previous report [38] , a large fraction of the intron-less transcripts remained stable over the time of analysis ( Figure 2D ) . To exclude the possibility that the nuclear export of the intron-less transcript is caused by the introduction of exogenously synthesized mRNAs , we tested the export of mRNA after its synthesis in the nucleus . To this end , we microinjected plasmids containing the t-ftz-Δi or t-ftz-i genes into NIH 3T3 nuclei . After 30 min , the RNA Polymerase II inhibitor α-amanitin was added to inhibit further transcription , and then the distribution of mRNA over time was monitored using FISH . Immediately after α-amanitin addition , most transcripts were in the nucleus , but ∼20% were already found in the cytoplasm . Over time , the nuclear fraction was efficiently exported to the cytoplasm ( Figure 2E and 2F ) with a rate that was slightly lower than that of microinjected RNA . As before , we observed only minor differences between intron-containing and intron-lacking t-ftz transcripts ( Figure 2F ) . Pretreatment of cells with α-amanitin 5 min prior to DNA microinjection completely inhibited mRNA synthesis , as assayed by FISH ( unpublished data ) . From these results , we conclude that export of an intron-lacking mRNA can occur efficiently , regardless of whether or not export is coupled to transcription . Our results are in apparent contradiction to previous observations showing that ftz constructs lacking introns are not exported from nuclei of Xenopus oocytes [1] . A major difference to the transcripts tested in Xenopus oocytes is the presence of an SSCR in t-ftz-Δi . We therefore tested whether the SSCR was required to promote nuclear export of this mRNA . Indeed , a transcript that lacked the SSCR , and thus encoded a cytoplasmic version of ftz ( c-ftz-Δi ) , remained in the nucleus 30 min after injection ( Figure 2G , quantitation in Figure 2C ) , a time during which most of the control t-ftz-Δi mRNA was exported . At later time points , much of the c-ftz-Δi mRNA was degraded ( Figure 2D ) . Because c-ftz-Δi mRNA was stable when injected directly into the cytoplasm ( unpublished data ) , it is unlikely that nuclear-injected c-ftz-Δi mRNA was exported and then rapidly degraded in the cytoplasm . The addition of an intron into c-ftz-Δi mRNA ( resulting in c-ftz-i mRNA ) restored nuclear export and stability of the mRNA ( Figure 2H , quantitation in Figure 2C and 2D ) . A fraction of the exported c-ftz-i transcripts accumulated in stress granules ( unpublished data ) . Together , these experiments suggest that either an SSCR or an intron can serve as a nuclear export signal . To exclude the possibility that the SSCR-mediated export pathway is a peculiarity of NIH 3T3 cells , we microinjected mRNA precursors into the nuclei of COS-7 cells . As before , t-ftz-i mRNA containing both the splicing and the SSCR signals—as well as t-ftz-Δi and c-ftz-i mRNAs , which each contain only one of the two signals—were exported efficiently into the cytoplasm ( Figure 3A and Figure S2; quantitation in Figure 3B ) . In contrast , c-ftz-Δi mRNA lacking both signals was not exported ( Figure 3B and Figure S2 ) . Again , the c-ftz-Δi mRNA was significantly less stable than the other mRNAs ( Figure 3C ) . These data suggest that the SSCR-mediated nuclear export pathway may be present in many mammalian cell types . In COS-7 cells , we were able to visualize the targeting of the mRNAs to the ER membrane . All mRNAs containing an SSCR gave a typical reticular staining in the cytoplasm and co-localized with TRAPα , a marker of the ER membrane ( Figure 3A and Figure S2 ) . To test whether other SSCRs could promote mRNA export , we replaced the SSCR of the MHC class 2 molecule H2-K1 with that of human insulin in the t-ftz-Δi construct ( ins-ftz-Δi , see Figure S3 ) . When injected into COS-7 cell nuclei , ins-ftz-Δi mRNA was efficiently exported ( Figure 3B and Figure S2 ) . In fact , the export kinetics was faster than with t-ftz-Δi mRNA ( Figure 3B ) . Again , the mRNA was targeted to the ER membrane ( Figure S2 ) . Taken together , these results indicate that the SSCR-mediated pathway may be quite general . Finally , we performed transfection experiments in COS-7 cells with plasmids coding for t-ftz and c-ftz mRNAs containing or lacking an intron . The steady-state distribution of the mRNAs between the nucleus and cytoplasm was determined by FISH ( Figure S4 , quantitation in Figure 3D ) . The SSCR-containing mRNAs ( t-ftz-i and t-ftz-Δi ) were mostly found in the cytoplasm , indicating that they are efficiently exported from the nucleus . A significant fraction of mRNAs lacking an SSCR ( c-ftz-i and c-ftz-Δi ) were found in the nucleus , despite the fact that one contained an intron . In the cytoplasm , the SSCR-containing ftz constructs were targeted to the ER membrane , whereas c-ftz constructs partially co-localized with TIA-1 ( Figure S4 ) , a marker of stress granules [39] . These data confirm the presence of an SSCR-mediated mRNA export pathway and suggest that , at least under certain conditions , it may be more efficient than the splicing-mediated export pathway . Next we characterized SSCR-mediated mRNA export , using t-ftz-Δi mRNA that contains the SSCR signal but lacks an intron . When this mRNA was injected into nuclei of NIH 3T3 cells that had been depleted of ATP by treatment with azide [40] , no export into the cytoplasm was observed ( Figure 4A ) . This treatment did not compromise the viability of the cells as judged by phase microscopy or immunostaining of the microtubule network ( unpublished data ) . Blocking the nuclear pores by pre-injecting wheat germ agglutinin ( WGA ) [41 , 42] also inhibited the nuclear export of microinjected t-ftz-Δi mRNA ( Figure 4B ) . Thus the export of t-ftz-Δi mRNA requires energy and functional nuclear pores . As expected , nuclear mRNA export was unidirectional . When transcripts were injected into one of the nuclei of a binucleated cell , 80% of the mRNA was transported into the cytoplasm , but no fluorescence was detected in the uninjected nucleus ( Figure S5 ) , indicating that there is no re-uptake of exported mRNA . Next we attempted to define the components required for t-ftz-Δi mRNA export . To test whether the export factor TAP is required , we took advantage of the viral constitutive transport element ( CTE ) , which binds to and sequesters TAP [43] . Nuclear export of t-ftz-Δi mRNA was inhibited in cells pre-injected with CTE RNA ( Figure 4C , arrow indicates a cell pre-injected with CTE , quantitation is shown in Figure 4D ) . In contrast , cells that were not pre-injected with CTE ( Figure 4C , arrowhead ) , or cells pre-injected with control buffer ( Figure 4D ) , exported t-ftz-Δi mRNA . Previously , it was demonstrated that short mRNAs use the TAP-independent snRNA export pathway [44] , but our results demonstrate that the t-ftz-Δi transcripts exclusively follow the TAP-dependent pathway used by most mRNAs . We then tested the involvement of the TREX complex . To deplete the TREX complex , we incubated HeLa cells with small interfering RNAs ( siRNAs ) against the TREX component UAP56 and its ortholog URH49 , a treatment that has been shown to inhibit the majority of mRNA export in mammalian cells [45] . After one day of RNA interference ( RNAi ) treatment , UAP56 levels were effectively reduced ( Figure 4E ) . Although siRNA-treated cells were compromised for the export of c-ftz-i transcript , these cells exported both t-ftz-Δi and t-ftz-i mRNA ( Figure 4F and 4G ) , suggesting that the SSCR-mediated export pathway is less sensitive to TREX complex depletion than the splicing-mediated pathway . In some cells depleted of UAP56/URH49 , t-ftz-Δi and t-ftz-i transcripts became enriched in the nuclear rim , perhaps due to a pleiotropic effect that inhibits the movement of mRNAs from the nuclear pores into the cytoplasm . Cells treated with siRNA for two days showed even more nuclear rim localization of t-ftz-Δi and t-ftz-i transcripts , and their export was inhibited . Although c-ftz-i mRNA was also not exported , it only accumulated in the nucleoplasm ( unpublished data ) , supporting the idea that different export pathways are used . As expected , the export of the intron-lacking RNA t-ftz-Δi was not reduced when a component of the exon junction complex ( EJC ) , the RNA helicase eIF4AIII , was depleted by RNAi ( Figure 4G and Figure S6 ) . However , even the export of the intron-containing mRNAs t-ftz-i and c-ftz-i remained unaffected , in agreement with previous results that indicated that the EJC is not a main factor in recruiting TREX [3 , 46] . Finally , we tested whether the SSCR-mediated export requires a 5′ end cap , as in splicing-dependent export [3] . Because uncapped transcripts were rapidly degraded after microinjection into NIH 3T3 nuclei ( unpublished data ) , we used mRNAs that were capped with the cap analogs ApppG or trimethyl-2 , 3 , 7-GpppG ( 3mGpppG ) . These modified caps do not associate with the cap binding complex CPB80/20 [47] , which is essential for the subsequent recruitment of the TREX complex [3] and for efficient splicing [47] . Capping of t-ftz-Δi mRNAs with the cap analogs did not inhibit nuclear export ( Figure 4H ) . In addition , the mRNAs were as stable as transcripts containing the natural cap , methyl-7-GpppG ( mGpppG ) ( Figure 4I ) . In contrast , the c-ftz-i transcripts that use the splicing-dependent pathway were not exported when capped with ApppG , as expected ( Figure 4H ) . Thus , we conclude that unlike the splicing-mediated mRNA export , the SSCR-mediated pathway does not require a natural cap . In principle , the SSCR nuclear export signal could either be an RNA element or its translated amino acid sequence , the signal sequence . To distinguish between these possibilities , we altered five nucleotides within the SSCR such that three encoded hydrophobic residues within the core of the signal sequence were changed to arginines ( 3R-ftz , see Figure S3 ) . The mutated signal sequence should no longer target the translation product to the ER membrane . Upon injection into the nuclei of COS-7 cells , 3R-ftz-Δi mRNA was exported with a kinetics similar to that of t-ftz-Δi ( Figure 5A; quantitation in Figure 5B ) . As expected , the mRNA was no longer targeted to the ER , and instead was distributed diffusely in the cytoplasm ( Figure 5A ) . Its translation product was targeted to mitochondria ( unpublished data ) , possibly because the signal sequence was converted into an amphipatic helix typical of mitochondrial targeting sequences [48] . To further ensure that the translation product of the SSCR did not determine mRNA export from the nucleus , we altered the ORF of the SSCR . We added a nucleotide at the beginning of the sequence and deleted a nucleotide at the end , such that the coding region following the SSCR remained unchanged . The altered SSCR codes for a less hydrophobic sequence . The frame-shifted ftz transcript ( fs-ftz-Δi , Figure S3 ) was again efficiently exported into the cytoplasm ( Figure 5A and 5B ) . As expected , it was not targeted to the ER ( Figure 5A ) , and its translation product remained in the cytoplasm ( unpublished data ) . Transcripts that lacked a translation start codon at the beginning of the SSCR ( UUG-ftz-Δi , see Figure S3 ) were also efficiently exported ( Figure 5B and Figure S2 ) . To further confirm that translation of the t-ftz-Δi transcript was not required for its export , cells were pretreated with pactamycin , an inhibitor of translation initiation [49] , before being microinjected . Although translation was effectively inhibited after 15 min of drug treatment , as assayed by 35S-methionine incorporation ( unpublished data ) , an even longer pretreatment with pactamycin ( 20 min ) did not inhibit t-ftz-Δi mRNA export or affect t-ftz-Δi stability ( Figure 5C–5E ) . Interestingly , after 1 h of pactamycin pretreatment , mRNA export was inhibited and nuclear rim staining was seen , as in cells depleted of the TREX components UAP56 and URH49 , which suggests that in both conditions , the synthesis of a factor required for the movement of the mRNA from the nuclear pores into the cytoplasm is impaired . From these experiments , we conclude that the nuclear export signal of the SSCR does not require translation and is thus likely an RNA element . To identify features in SSCRs that might function as nuclear export signals , we performed a large-scale sequence analysis of various genomes . We confined our analysis to the first 69 base pairs ( bp ) following the initiator methionine codon , a length that covers most SSCRs . We used the annotation in Ensembl and PSORTdb to classify genes into SSCR-containing and SSCR-lacking ORFs in genomes ranging from bacteria to humans . As an additional control , we analyzed 69 bp in the central region of each ORF , which should reflect the overall base composition of the coding region . This analysis showed that the SSCRs in all eukaryotes have a marked deficiency of adenines , in contrast to those in bacteria ( Figure 6A ) . Consistent with this observation , eukaryotic SSCRs have long nucleotide stretches devoid of adenines ( no-A tracts ) ( Figure 6B ) . The no-A tracks were longer in vertebrates than in invertebrates . Several factors account for the adenine deficiency in SSCRs . First , hydrophobic amino acids , which are enriched in the signal sequence , have codons that contain few adenines . Another factor is bias among amino acids with similar biochemical properties , such that amino acids encoded by codons with fewer adenines are used . This is illustrated by leucine , which has few adenine-containing codons , and isoleucine , which has at least one adenine in each of its codons . As previously noted , human signal sequences have significantly more leucine residues than equally hydrophobic isoleucine residues [31] . We confirmed this leucine-versus-isoleucine bias for SSCRs in all vertebrates analyzed ( Figure 6C ) . A similar analysis was performed on positively charged amino acids that are frequently found in the amino acid sequence preceding the hydrophobic core of a signal sequence . Arginine is encoded by codons containing relatively few adenines , whereas lysine is encoded by AAA and AAG , and arginines were significantly more frequent than lysines in vertebrate , but not invertebrate , SSCRs ( Figure 6D ) . In total , about 15% to 25% of the adenine deficiency in SSCRs in vertebrates is caused by selection between similar amino acids for those encoded by codons with lower adenine content ( Figure S7 ) . The third factor that contributes to the adenine deficiency is a bias toward codons lacking adenine for amino acids that are encoded by multiple codons ( e . g . , CUC versus CUA ) . This is illustrated by the usage of codons for leucine and serine . In vertebrates , the adenine-lacking codons for both amino acids are more frequently used than the ones containing adenines ( Figure 6E and 6F ) . The bias against adenine-containing codons in vertebrate SSCRs could also be seen when the analysis was extended to all amino acids that are encoded by both adenine-lacking and -containing codons ( Figure 6G ) . In total , bias between synonymous codons accounts for an additional 15% to 25% of the adenine deficiency in vertebrates ( Figure S7 ) . In humans , the bias between similar amino acids and between synonymous codons together account for almost half of the total adenine deficiency ( Figure S7 ) . Our analysis suggests that over the course of evolution , there was a strong selection against adenines that was likely caused by some requirement of the nucleotide sequence , rather than of the encoded amino acid sequence . It should be noted that the two SSCRs that were tested experimentally , completely follow the rules established in the large-scale sequence analysis . The longest no-A track in the SSCRs derived from H2-K1 and insulin were 35 and 40 nucleotides , respectively . Both SSCRs contain leucines and arginines , but no isoleucines or lysines , and both have a high bias against adenine-containing codons . To test whether the bias against adenines within SSCRs is important for promoting mRNA export , we mutated seven nucleotides of the SSCR that were derived from the MHC class 2 molecule H2-K1 to adenines , without altering the encoded amino acid sequence ( 7A-ftz-Δi , see Figure S3 ) . Upon injection into COS-7 cell nuclei , the export of 7A-ftz-Δi mRNA into the cytoplasm was significantly less efficient than that mediated by the original SSCR ( Figure 7A and 7B ) . The nuclear export of an intron-containing version ( 7A-ftz-i ) was not inhibited by the mutations ( Figure 7A and 7B ) , which is consistent with the expectation that this transcript can use the splicing-dependent pathway . The exported 7A-ftz-Δi and 7A-ftz-i mRNAs were targeted to the ER ( Figure 7A ) and resulted in the expression of ER-bound t-ftz protein ( unpublished data ) , indicating that translation of the mRNAs was largely unaffected by the silent mutations . The mutations also did not grossly affect the stability of the mRNAs ( Figure 7C ) . The nuclear export of 7A-ftz-Δi mRNA was also significantly less efficient than that of t-ftz-Δi mRNA when tested in transfection experiments with plasmids coding for these RNAs ( Figure 3D and Figure S4 ) . In fact , the export was as inefficient as with RNA lacking an SSCR altogether , indicating that the adenine mutations severely affected the SSCR-export signal . In addition , a portion of the cytoplasmic fraction of the 7A-ftz-Δi mRNAs was localized to stress granules , as determined by co-staining with the stress granule marker TIA-1 ( Figure S4 ) . An inhibitory effect of adenines on the nuclear export function of an SSCR could also be demonstrated for human insulin mRNA . We first used intronless transcripts , generated by in vitro transcription of human insulin cDNA . Upon injection into COS-7 cell nuclei , insulin mRNA with a wild-type SSCR was efficiently exported into the cytoplasm , whereas mutant mRNA , in which five silent adenine mutations were introduced into its SSCR , was significantly delayed ( Figure 7D and 7E ) . Again , the mutations did not significantly affect the stability of the mRNAs ( Figure 7F ) . To address whether the SSCR contributed to the export of a physiologically transcribed and spliced mRNA , we microinjected plasmids that contained the insulin gene with its two introns under the control of the CMV promoter ( insulin-2i ) . After 30 min , transcription was blocked with α-amanitin , and nuclear export of the newly synthesized transcripts was followed over the course of 2 h . The rate of mRNA export was significantly decreased when silent adenine mutations were incorporated into the SSCR of intron-containing or -lacking mRNAs ( 5A-insulin-2i and 5A-insulin-Δi ) ( Figure 7G ) . On the other hand , deletion of the introns ( insulin-Δi ) had no effect . All the tested transcripts were stable over the tested time course ( Figure 7H ) . From these experiments we conclude that a functional SSCR can enhance the export of mRNAs , regardless of whether they contain or lack introns . These results provide evidence that the SSCR-mediated pathway operates within the context of a natural gene .
Here we describe the surprising discovery of a nuclear export pathway in mammalian cells that appears to be specific for mRNAs coding for secretory proteins . We found that mRNAs lacking an intron or functional cap , which cannot use the canonical , splicing-dependent pathway , can efficiently be exported into the cytoplasm by means of a signal in the SSCR . The signal is an RNA element that is deficient in adenines . Thus , the SSCR not only codes for the hydrophobic amino acid sequence that targets the translation product to the ER membrane , but also functions at the nucleotide level to promote the export of the mRNA from the nucleus . The SSCR even enhances the export of natural transcripts containing introns . Like the splicing-dependent pathway , the SSCR-mediated export pathway requires the export factor TAP , but it is less dependent , or perhaps even independent , of the TREX complex . Our sequence analysis suggests that the SSCR-mediated pathway may exist in all vertebrates . Why would mRNAs coding for secretory proteins use a separate nuclear export pathway ? One possibility is that nuclear export of these mRNAs might be coupled with some downstream event in the cytoplasm . For example , it is possible that the mRNAs emerging from the nuclear pores are initially associated with proteins that keep them translationally silenced and promote their distribution in the cytoplasm . These factors could be distinct from those used by mRNAs coding for cytoplasmic proteins . For example , factors associated with SSCR containing transcripts could have affinity for ER membrane proteins or molecular motors . The recent observation of an association of the TAP homolog , NXF2 , with kinesin lends support to the idea that the export and cytoplasmic distribution of mRNAs may be coupled [50] . Our own data show that in transfection experiments in which the steady-state distribution of mRNAs was investigated , transcripts with defective or lacking SSCRs were not only inefficiently exported but also often accumulated in cytoplasmic stress granules , indicating that their normal cytoplasmic distribution was disrupted ( Figure S4 ) . How nuclear mRNA export would be coupled with downstream events remains to be elucidated . It is also possible that the export of mRNAs coding for secretory proteins is regulated differently from other mRNAs in certain situations . However , we have not found any changes of SSCR-mediated mRNA export upon accumulation of misfolded proteins in the ER ( unpublished data ) . In yeast , a block in secretion results in the relocalization of certain nuclear pore components to the cytoplasm and leads to the inhibition of protein transport between the nucleus and the cytoplasm [51 , 52] . Conceivably , this mechanism may also provide a feedback signal between the demand of secretory protein synthesis in the cytoplasm and the nuclear export of the corresponding SSCR-containing mRNAs . The mechanism by which SSCRs are recognized and trigger nuclear mRNA export also remains to be investigated . A clear nucleotide motif that is common among all SSCRs is not obvious . One possibility is that a negative export regulator would bind to all adenine-containing , non-SSCR sequences . However , we favor models in which either the paucity of adenines per se is recognized by an export-mediating protein with a similar RNA-binding specificity as the Muscleblind family of proteins [53] , or that the adenine-lacking segment of an SSCR folds into a conformation that would recruit an export factor . Regardless of the precise signal , its position at the 5′ end of the mRNA would allow this part of the mRNA to emerge into the cytoplasm first , in the same direction of mRNA export used in the splicing-dependent pathway [3 , 54] . Although many SSCR-containing mRNAs also contain introns , the 5′ end localization of the SSCR signal might allow it to recruit factors to the newly synthesized transcript before the introns are synthesized and thereby overrule the splicing-dependent signals . Our data indicate that the SSCR recruits TAP , likely without involvement of TREX . One possibility is that the serine/arginine-rich ( SR ) proteins serve as adaptors for TAP binding . SR proteins associate with mRNAs during transcription and/or splicing . They are required for splicing [55] , can associate with TAP [56 , 57] , and are involved in the export of intron-lacking histone H2A mRNA [58] . We have not been able to prevent SSCR-dependent mRNA export by pre-injection of an SR antibody that inhibits splicing [55] ( unpublished results ) , but further work is required to test the possible role of the SR proteins in SSCR-mediated export . Our analysis shows that the SSCRs have a nucleotide bias that cannot be explained solely by the encoded amino acid sequence . Although these results indicate that the nucleotide sequence itself is important , there may be additional variability of signal sequences at the level of amino acids , as suggested by experiments in which different signal sequences responded differently to the accumulation of unfolded proteins in the ER [27] . One would assume that the SSCR-mediated mRNA export pathway operates in all eukaryotic cells , but our large-scale sequence analysis showed a clear bias against adenine-containing codons only in vertebrate SSCRs . A slight bias against adenines was seen in lower eukaryotes , but it is possible that SSCRs have additional properties that are found in all eukaryotes , for example , a common folded structure . The same argument might explain the absence of any obvious adenine bias at the 5′ end of genes coding for membrane proteins that lack a signal sequence ( unpublished data ) , even though the corresponding mRNAs also need to be targeted to the ER and would likely take the same pathway as those coding for secretory proteins . Obviously , the surprising discovery of an SSCR-dependent export pathway for mRNAs raises a large number of interesting questions that need to be addressed in the future .
For RT-PCR , total RNA was extracted from injected cells and analyzed using M-MLV reverse transcriptase ( Invitrogen ) and HiFi Taq polymerase ( Invitrogen ) according to the manufacturer's protocol . PCR reactions were carried out using gene-specific primer pairs , Ftz-127F and Ftz-444R ( see Figure 1A ) , and for controls , we used mouse 18S rRNA-448F and mouse 18S rRNA-926R . The ftz constructs were modified to generate t-ftz-i ( Figure 1A and Figure S1 ) , t-ftz-Δi , and their derivatives ( Figure S3 ) . For in vivo mammalian expression experiments , t-ftz constructs were digested with HindIII and XhoI , and then ligated into the pcDNA3 expression vector . Human insulin cDNA was amplified from a cDNA library and the insulin gene was amplified from HeLa genomic DNA using gene-specific primer pairs digested with HindIII and XhoI , and then ligated into the pcDNA3 expression vector . Insulin was modified with PCR primers to generate 5A-insulin ( see Figure S3 ) . In vitro transcription was carried out using the T7 mMESSAGE mMACHINE transcription kit containing excess cap ( Ambion ) . To synthesize mRNAs with cap analogues , the reaction was carried out with 35 mM ApppG or 3mGpppG ( New England Biolabs ) ; 10 mM ATP , UTP , and CTP; and 1 mM GTP . Transcripts were poly-adenylated using Poly ( A ) tailing kit ( Ambion ) , generating poly ( A ) tails of 200–300 nucleotides . mRNA purification was carried out using MEGAclear kit ( Ambion ) . mRNA was then precipitated with 150 mM potassium acetate ( pH 5 . 5 ) and 2 . 5 volumes 100% ethanol . mRNA was resuspended in injection buffer ( 100 mM KCl , 10 mM HEPES , pH 7 . 4 ) . mRNAs were translated in a TnT reticulocyte lysate system in the presence of 35S-methionine ( Promega ) . NIH 3T3 fibroblasts were maintained in DMEM supplemented with 10% calf serum . HeLa and COS-7 cells were maintained in DMEM supplemented with 10% fetal bovine serum . Cells were plated overnight on 35-mm-diameter dishes with glass coverslip bottoms ( MatTek Corp . ) . For RNAi experiments , HeLa cells were transfected with siRNA directed against human UAP56 and URH49 [45] or eIF4AIII [59] . 24 and 48 h post transfection , cells were either plated on 35-mm dishes ( for microinjections ) or collected to assess protein levels by SDS-PAGE and Western blot with rabbit anti-UAP56 serum [4] , rabbit anti-eIF4AIII [59] , or rabbit anti-CBP80 serum [3] . For DNA transfections , cells were transfected with DNA and lipofectamine ( Invitrogen ) using the manufacture's protocol . Microinjections were performed as previously described [60] . mRNA was microinjected at 200 μg/ml along with fluorescein isothiocyanate ( FITC ) –conjugated 70-kDa dextran ( 1 mg/ml; Invitrogen ) . Insulin mRNA was heated to 70 °C for 10 min prior to injections . DNA was injected at 50 μg/ml along with FITC-conjugated 70-kDa dextran , and translation was inhibited with 50 μg/ml α-amanitin ( Sigma ) . For export inhibition experiments , cells were first microinjected with WGA ( 3 mg/ml; Sigma ) or CTE RNA ( 200 μg/ml ) along with cascade-blue–conjugated 10-kDa dextran ( Invitrogen ) , and then incubated for 30 min at 37 °C prior to mRNA microinjection . For azide treatments , microinjected cells were washed three times with Dulbecco's modified PBS ( D-PBS; 10 mM phosphate , pH 7 . 4 , 140 mM NaCl , 3 mM KCl , 0 . 8 mM CaCl2 , 0 . 7 mM MgCl2 ) and incubated in D-PBS supplemented with 10 mM azide ( Sigma ) or 10 mM D-glucose ( Sigma ) . For pactamycin treatments , cells were incubated in DMEM ( 10% fetal bovine serum ) with 200 nM pactamycin 20 min before microinjections . Microinjected cells were washed with D-PBS , fixed in 4% paraformaldehyde ( Electron Microscope Sciences ) in D-PBS , and permeabilized with 0 . 1% Triton X100 ( Peirce ) in PBS . For immunostaining , fixed samples were first incubated with primary antibodies ( rabbit polyclonal against TRAPα [61] , goat polyclonal antibody against TIA-1 [Santa Cruz Biotechnology] , 12CA5 monoclonal antibody against HA [Roche Applied Sciences] , and M2 monoclonal antibody against FLAG [Sigma] ) diluted 1:200 in immunostain solution ( PBS , 0 . 1% Triton X100 , 2 mg/ml RNAse free BSA; Ambion ) for 30 min , washed three times with PBS , and incubated with various Alexa-conjugated secondary antibodies ( Invitrogen ) , diluted 1:200 in immunostain solution . For FISH , fixed cells were washed with 50% formamide in 1X SSC ( 150 mM NaCl , 15 mM NaCitrate , pH 7 . 10 ) and then incubated overnight at 37 °C in 200-ml hybridization buffer ( 50% formamide , 100 mg/ml dextran sulphate , 0 . 02 mg/ml RNAse free BSA , 1 mg/ml Escherichia coli tRNA , 5 mM VRC , 1X SSC ) containing 30–50 ng oligonucleotide probe ( GTCGAGCCTGCCTTTGTCATCGTCGTCCTTGTAGTCACAACAGCCGGGACAACACCCCAT for ftz , GGTCCTCTGCCTCCCGGCGGGTCTTGGGTGTGTAGAAGAAGCCTCGTTCCCCGCACACTA for insulin ) labeled at their 5′ end with Alexa-546 ( Integrated DNA Technologies ) . Cells were washed in five times with 50% formamide in 1X SSC and images were captured using an EM-CCD Camera , Model C9100–12 ( Hamamatsu ) on an inverted microscope ( 200M , Carl Zeiss ) using Metamorph software ( Molecular Devices Corporation ) . Unaltered 14-bit images were quantified in Metamorph and analyzed in Excel ( Microsoft ) . For each image the area ( A ) and the average intensity ( I ) of each injected nuclei ( n ) and cell body ( b ) were recorded . For the background intensity , the average intensity of an un-injected cell ( u ) was used . The cytoplasmic fluorescence was equal to ( Ab ) ( Ib – Iu ) – ( An ) ( In – Iu ) . The ratio of cytoplasmic/total fluorescence equals [ ( Ab ) ( Ib – Iu ) – ( An ) ( In – Iu ) ]/[ ( Ab ) ( Ib – Iu ) ] . For figure production , the contrast and brightness of the aquired 14-bit micrographs were adjusted to optimize the ability to view the fluorescence . The resulting images were converted to 8-bit files using Metamorph . Nucleotide sequences were downloaded from Ensembl Biomart ( http://www . biomart . org/index . html ) and the National Center for Biotechnology Information ( NCBI ) ( E . coli . and Bacillus subtilis ) . Signal sequence containing proteins were determined by annontation in PSORTdb ( E . coli and B . subtilis ) and Ensembl Biomart . A Perl script ( Protocol S1 ) was written to count the nucleotide content , amino acid content , and codon content , of the first 69 nucleotides , middle 69 nucleotides ( offset toward the start to maintain frame as needed ) , and last 69 nucleotides . The script also determined the longest no-adenine and one-adenine tracks completely contained in the first 69 nucleotides . Tabulated results were examined in Excel to determine nucleotide content , codon bias , and amino acid bias . Percent of adenine bias explained by codon bias ( e . g . , CUC versus CUA ) and similar amino acid bias ( e . g . , isoleucine versus leucine ) was calculated by comparing the number of adenines in non–signal sequence containing proteins to the number of adenines found in signal sequence containing proteins normalized for the number of proteins .
The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank ) accession numbers for genes investigated in the paper are as follows: H2-K1 ( GI: 14972 ) , TAP ( GI: 10482 ) , UAP56 ( GI: 7919 ) , URH49 ( GI: 10212 ) , and eIF4AIII ( GI: 9775 ) . The GenBank accession number for human insulin cDNA is GI: 3630 . | In eukaryotic cells , precursors of messenger RNAs ( mRNAs ) are synthesized and processed in the nucleus . During processing , noncoding introns are spliced out , and a cap and poly-adenosine sequence are added to the beginning and end of the transcript , respectively . The resulting mature mRNA is exported from the nucleus to the cytoplasm by crossing the nuclear pore . Both the introns and the cap help to recruit factors that are necessary for nuclear export of an mRNA . Here we provide evidence for a novel mRNA export pathway that is specific for transcripts coding for secretory proteins . These proteins contain signal sequences that target them for translocation across the endoplasmic reticulum membrane . We made the surprising observation that the signal sequence coding region ( SSCR ) can serve as a nuclear export signal of an mRNA that lacks an intron or functional cap . Even the export of an intron-containing natural mRNA was enhanced by its SSCR . The SSCR export signal appears to be characterized in vertebrates by a low content of adenines . Our discovery of an SSCR-mediated pathway explains the previously noted amino acid bias in signal sequences , and suggests a link between nuclear export and membrane targeting of mRNAs . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
]
| [
"cell",
"biology",
"computational",
"biology",
"molecular",
"biology"
]
| 2007 | The Signal Sequence Coding Region Promotes Nuclear Export of mRNA |
A systematic assessment of global neural network connectivity through direct electrophysiological assays has remained technically infeasible , even in simpler systems like dissociated neuronal cultures . We introduce an improved algorithmic approach based on Transfer Entropy to reconstruct structural connectivity from network activity monitored through calcium imaging . We focus in this study on the inference of excitatory synaptic links . Based on information theory , our method requires no prior assumptions on the statistics of neuronal firing and neuronal connections . The performance of our algorithm is benchmarked on surrogate time series of calcium fluorescence generated by the simulated dynamics of a network with known ground-truth topology . We find that the functional network topology revealed by Transfer Entropy depends qualitatively on the time-dependent dynamic state of the network ( bursting or non-bursting ) . Thus by conditioning with respect to the global mean activity , we improve the performance of our method . This allows us to focus the analysis to specific dynamical regimes of the network in which the inferred functional connectivity is shaped by monosynaptic excitatory connections , rather than by collective synchrony . Our method can discriminate between actual causal influences between neurons and spurious non-causal correlations due to light scattering artifacts , which inherently affect the quality of fluorescence imaging . Compared to other reconstruction strategies such as cross-correlation or Granger Causality methods , our method based on improved Transfer Entropy is remarkably more accurate . In particular , it provides a good estimation of the excitatory network clustering coefficient , allowing for discrimination between weakly and strongly clustered topologies . Finally , we demonstrate the applicability of our method to analyses of real recordings of in vitro disinhibited cortical cultures where we suggest that excitatory connections are characterized by an elevated level of clustering compared to a random graph ( although not extreme ) and can be markedly non-local .
The identification of the topological features of neuronal circuits is an essential step towards understanding neuronal computation and function . Despite considerable progress in neuroanatomy , electrophysiology and imaging [1]–[8] , the detailed mapping of neuronal circuits is already a difficult task for a small population of neurons , and becomes impractical when accessing large neuronal ensembles . Even in the case of cultures of dissociated neurons , in which neuronal connections develop de novo during the formation and maturation of the network , very few details are known about the statistical features of this connectivity , which might reflect signatures of self-organized critical activity [9]–[11] . Neuronal cultures have emerged in recent years as simple , yet versatile model systems [12] , [13] in the quest for uncovering neuronal connectivity [14] , [15] and dynamics [16]–[19] . The fact that relatively simple cultures already exhibit a rich repertoire of spontaneous activity [18] , [20] make them particularly appealing for studying the interplay between activity and connectivity . The activity of hundreds to thousands of cells in in vitro cultured neuronal networks can be simultaneously monitored using calcium fluorescence imaging techniques [14] , [21] , [22] . Calcium imaging can be applied both in vitro and in vivo and has the potential to be combined with stimulation techniques like optogenetics [23] . A major drawback of this technique , however , is that the typical frame rate during acquisition is slower than the cell's firing dynamics by an order of magnitude . Furthermore the poor signal-to-noise ratio makes the detection of elementary firing events difficult . Neuronal cultures are unique platforms to investigate and quantify the accuracy of network reconstruction from activity data , extending analysis tools initially devised for the characterization of macro-scale functional networks [24] , [25] to the micro-scale of a developing local circuit . Here we report a new technique based on information theory to reconstruct the connectivity of a neuronal network from calcium imaging data . We use an extension of Transfer Entropy ( TE ) [26]–[28] to extract a directed functional connectivity network in which the presence of a directed edge between two nodes reflects a direct causal influence by the source to the target node [29]–[31] . Note that “causal influence” is defined operationally as “improved predictability” [32] , [33] reflecting the fact that knowledge of the activity of one node ( putatively pre-synaptic ) is helpful in predicting the future behavior of another node ( putatively post-synaptic ) . TE has previously been used to study gene regulatory networks [34] , the flow of information between auditory neurons [35] , to infer directed interactions between brain areas based on EEG recordings [36] or between different LFP frequency bands [37] , as well as for the reconstruction of the connectivity based on spike times [38] , [39] . Importantly , our data-driven TE approach is model-independent . This is in contrast with previous approaches to network reconstruction , which were most often based on the knowledge of precise spike times [40]–[45] , or explicitly assumed a specific model of neuronal activity [43] , [44] . A problem inherent to the indirect algorithmic inference of network connectivity from real data is that the true target topology of the network is not known and that , therefore , it is difficult to assess the quality of the reconstruction . In order to characterize the behavior of our algorithm and to benchmark its potential performance , we resort therefore to synthetic calcium fluorescence time series generated by a simulated cultured neural network that exhibits realistic dynamics . Since the “ground truth” topology of cultures in silico is known and arbitrarily selectable , the quality of our reconstruction can be evaluated by systematically comparing the inferred with the real network connectivities . We use a simplified network simulation to generate surrogate imaging data , improving their realism with the reproduction of light scattering artifacts [6] which ordinarily affect the quality of the recording . Our surrogate data also reproduce another general feature of the activity of neuronal cultures , namely the occurrence of temporally irregular switching between states of asynchronous activity , with relatively weak average firing rates , and states of highly synchronous activity , commonly denoted as “network bursts” [20] , [46] , [47] . This switching dynamics poses potentially a major obstacle to reconstruction , since directed functional connectivity can be very different during bursting and inter-burst phases and can bear a resemblance to the underlying structural ( i . e . synaptic ) connectivity only in selected dynamical regimes in which causal influences reflect dominantly mono-synaptic interactions . To restrict our analysis to such “good” regimes , we resort to conditioning with respect to the averaged fluorescence level , as an indirect but reliable indicator of the network collective dynamics . Appropriate conditioning —combined with a simple correction coping with the poor time-resolution of imaging data— allows the method to achieve a good topology reconstruction performance ( assessed from synthetic data ) , out-performing other standard approaches , without the need to infer exact spike times through sophisticated techniques ( as is required , on the contrary , in [43] , [45] ) . Finally , we apply our algorithm —optimized through model-based validation— to the analysis of real calcium imaging recordings . For this purpose , we study spontaneously developing networks of dissociated cortical neurons in vitro and we address , as a first step toward a full topology reconstruction , the simpler problem of extracting only their excitatory connectivity . Early mature cultures display a bursting dynamics very similar to our simulated networks , with which they also share an analogous state-dependency of directed functional connectivity . Our generalized TE approach thus identifies network topologies with characteristic and non-trivial features , like the existence of non-local connections , a broadened and strongly right-skewed distribution of degrees ( although not “scale free” ) and a moderate but significant level of clustering .
In this study , we consider recordings from in vitro cultures of dissociated cortical neurons ( see Materials and Methods ) . To illustrate the quality of our recordings , in Figure 1A we provide a bright field image of a region of a culture together with its associated calcium fluorescence . As previously anticipated , to simplify the network reconstruction problem , experiments are carried out with blocked inhibitory GABA-ergic transmission , so that the network activity is driven solely by excitatory connections . We record activity of early mature cultures at day in vitro ( DIV ) 9–12 . Such young but sufficiently mature cultures display rich bursting events , combined with sparse irregular firing activity during inter-burst periods ( cfr . Discussion ) . In Figure 1B ( left panel ) we show actual recordings of the fluorescence traces associated to five different neurons . The corresponding population average for the same time window is shown in Figure 1C ( left panel ) . In these recordings , a stable baseline is broken by intermittent activity peaks that correspond to synchronized network bursts recruiting many neurons . The bursts display a fast rise of fluorescence at their onset followed by a slow decay . In addition , during inter-burst periods , smaller modulations above the baseline are sometimes visible despite the poor time-resolution of a frame every few tens of milliseconds . In order to benchmark and optimize different reconstruction methods , we also generate surrogate calcium fluorescence data ( shown in the right panels of Figures 1B and 1C ) , based on the activity of simulated networks whose ground truth topology is known . We simulate the spontaneous spiking dynamics of networks formed by excitatory integrate-and-fire neurons , along a duration of 60 minutes of real time , matching typical lengths of actual recordings . Calcium fluorescence time series are then produced based on this spiking dynamics , resorting to a model introduced in [43] and described in the Materials and Methods section . Although over 1000 cells are accessible in our experiments , we observed in the simulations that neurons suffice to reproduce the same dynamical behavior observed for larger network sizes , while still allowing for an exhaustive exploration of the entire algorithmic parameter space . Furthermore , despite their reduced density , we maintain in our simulated cultures the same average probability of connection as in actual cultures , where this probability ( , see Materials and Methods ) is an estimate based on independent studies [15] , [48] . The fluorescence signal of a particular simulation run or experiment can be conveniently studied in terms of the distribution of fluorescence amplitudes . As shown in Figure 1D for both simulations and experiments , the amplitude distributions display a characteristic right-skewed shape that emerge from the switching between two distinct dynamical regimes , namely the presence or absence of bursts . The distribution in the low fluorescence region assumes a Gaussian-like shape , corresponding to noise-dominated baseline activity , while the high fluorescence region displays a long tail with a cut-off at the level of calcium fluorescence of the highest network spikes . As we will show later , qualitative similarity between the shapes of the simulated and experimental fluorescence distributions will play an important guiding role for an appropriate network reconstruction . Neurons grown in vitro develop on a two-dimensional substrate and , hence , both connectivity and clustering may be strongly sensitive to the physical distance between neurons . At the same time , due to long axonal projections [13] , [49] , excitatory synaptic connections might be formed at any distance within the whole culture and both activity and signaling-dependent mechanisms might shape non-trivially long-range connectivity [50] , [51] . To test the reconstruction performance of our algorithm , we consider two general families of network topologies that cover a wide range of clustering coefficients . In a first one , clustering occurs between randomly positioned nodes ( non-local clustering ) . In a second one , the connection probability between two nodes decays with their Euclidean distance according to a Gaussian distribution and , therefore , connected nodes are also likely to be spatially close . In particular , in this latter case , the overall level of clustering is determined by how fast the connection probability decays with distance ( local clustering ) . Cortical slice studies revealed the existence of both local [52] , [53] and non-local [7] , [54] types of clustering . We will later benchmark reconstruction performance for both kinds of topologies and for a wide range of clustering levels , because very similar patterns of neuronal activity can be generated by very different networks , as we now show . Figure 2 illustrates the dynamic behavior of three networks ( in this case from the non-local clustering ensemble ) . The networks are designed to have different clustering coefficients but the same total number of links ( see the insets of Figure 2B for an illustration ) . The synaptic coupling between neurons was adjusted in each network using an automated procedure to obtain bursting activities with comparable bursting rates ( see Materials and Methods for details and Table 1 for the actual values of the synaptic weight ) . As a net effect of this procedure , the synaptic coupling between neurons is slightly reduced for larger clustering coefficients . The simulated spiking dynamics is shown in the raster plots of Figure 2A . These three networks display indeed very similar bursting dynamics , not only in terms of the mean bursting rate , but also in terms of the entire inter-burst interval ( IBIs ) distribution , shown in Figure 2B . In the same manner , we constructed and simulated local networks —with a small length scale corresponding to high clustering coefficients and vice versa— and obtained the same result , i . e . very similar dynamics for very different decay lengths ( not shown ) . We stress that our procedure for the automatic generation of networks with similar bursting dynamics was not guaranteed to converge for such a wide range of clustering coefficients . Thus , the illustrative simulations of Figure 2 provide genuine evidence that the relation between network dynamics and network structural clustering is not trivially “one-to-one” , despite the fact that more clustered networks have been shown to have different cascading dynamics at the onset of a burst [42] . We focus , first , on the reconstruction of simulated networks , taken from the local and non-local ensembles described above . We compute their directed functional connectivity based on simulated calcium signals . Synthetic fluorescence time series are pre-processed only by simple discrete differentiation , such as to extract baseline modulations associated to potential firing . These differentiated signals are then used as input to any further analyses . Immediately prior to the onset of a burst the network is very excitable . In such a situation it is intuitive to consider that the directed functional connectivity can depart radically from the structural excitatory connectivity , because local events can potentially induce changes at very long ranges due to collective synchronization rather than to direct synaptic coupling . Conversely , in the relatively quiet inter-burst phases , a post-synaptic spike is likely to be influenced solely by the presynaptic firing history . Hence , the directed functional connectivity between neurons is intrinsically state dependent ( cfr . also [55] ) , a property that must be taken into account when reconstructing the connectivity . We illustrate here the state dependency of directed functional connectivity by generating a random network from the local clustering ensemble and by simulating its dynamics , including light scattering artifacts to obtain more realistic fluorescence signals . The resulting distribution of fluorescence amplitudes is divided into seven non-overlapping ranges of equal width , each of them identified with a Roman numeral ( Figure 3A ) . Finally , TE is computed separately for each of these ranges , based on different corresponding subsets of data from the simulated recordings . For simulated data , the inferred connectivity can be directly compared to the ground truth , and a standard Receiver-Operator Characteristic ( ROC ) analysis can be used to quantify the quality of reconstruction . ROC curves are generated by gradually moving a threshold level from the lowest to the highest TE value , and by plotting at each point the fraction of true positives as a function of the fraction of false positives . The quality of reconstruction is then summarized in a single number by the performance level , which , following an arbitrary convention , is measured as the fraction of true positives at 10% of false positives read out of a complete ROC curve . We plot the performance level as a function of the average fluorescence amplitude in each interval , as shown by the blue line of Figure 3B . The highest accuracy is achieved in the lowest fluorescence range , denoted by I , and reaches a remarkably elevated value of approximately 70% of true positives . The performance in the higher ranges II to IV decreases to a value around 45% , to abruptly drop at range V and above to a final plateau that corresponds to the 10% performance of a random reconstruction ( ranges VI and VII ) . Note that fluorescence values are not distributed homogeneously across ranges I–VII , as evidenced by the overall shape of the fluorescence distribution in Figure 3A . For example , the lowest and highest ranges ( I and VII ) differ by two orders of magnitude in the number of data points . To discriminate unequal-sampling effects from actual state-dependent phenomena , we studied the performance level using an equal number of data points in all ranges . Effectively , we restrict the number of data points available in each range to be equal to the number of samples in the highest range , VII . The quality of such a reconstruction is shown as the red curve in Figure 3B . The performance level is now generally lower , reflecting the reduced number of time points which are included in the analysis . Interestingly , the “true” peak of reconstruction quality is shifted to range II , corresponding to fluorescence levels just above the Gaussian in the histogram of Figure 3A . This range is therefore the most effective in terms of reconstruction performance for a given data sampling . For the ranges higher than II , the reconstruction quality gradually decreases again to the 10% performance of purely random choices in ranges VI and VII . The effect of adopting a ( shorter ) equal sample size is particularly striking for range I , which drops from the best performance level almost down to the baseline for random reconstruction . As a matter of fact , range I is the one for which the shrinkage of sample length due to the constraint for uniform data sampling is most extreme ( see later section on dependence of performance from sample size ) . The above analysis leads to a different functional network for each dynamical range studied . For the analysis with an equal number of data point per interval , the seven effective networks are drawn in Figure 3C ( for clarity only the top 10% of links are shown ) . Each functional network is accompanied with the corresponding ROC curve . The lowest range I corresponds to a regime in which spiking-related signals are buried in noise . Correspondingly , the associated functional connectivity is practically random , as indicated by a ROC curve close to the diagonal . Nevertheless , information about structural topology is still conveyed in the activity associated to this regime and can be extracted through extensive sampling . At the other extreme , corresponding to the upper ranges V to VII —associated to fully developed synchronous bursts— the functional connectivity has also a poor overlap with the underlying structural network . As addressed later in the Discussion section , functional connectivity in regimes associated to bursting is characterized by the existence of hub nodes with an elevated degree of connection . The spatio-temporal organization of bursting can be described in terms of these functional connectivity hubs , since nodes within the neighborhood of the same functional hub provide the strongest mutual synchronization experienced by an arbitrary pair of nodes across the network ( see Discussion and also Figure S2 ) . The best agreement between functional and excitatory structural connectivity is clearly obtained for the inter-bursts regime associated with the middle range II , and to a lesser degree in ranges III and IV , corresponding to the early building-up of synchronous bursts . Overall , this study of state-dependent functional connectivity provides arguments to define the optimal dynamical regime for network reconstruction: The regime should include all data points whose average fluorescence across the population is below a “conditioning level” , located just on the right side of the Gaussian part of the histogram of the average fluorescence ( see Materials and Methods ) . This selection excludes the regimes of highly synchronized activity ( ranges III to VII ) and keeps most of the data points for the analysis in order to achieve a good signal-to-noise ratio . Thus , the inclusion of both ranges I and II combines the positive effects of correct state selection and of extensive sampling . The state-dependency of functional connectivity is not limited to synthetic data . Very similar patterns of state-dependency are observed also in real data from neuronal cultures . In particular , in both simulated and real cultures , the functional connectivity associated to the development of bursts displays a stronger clustering level than in the inter-burst periods . An analysis of the topological properties of functional networks obtained from real data in different states ( compared with synthetic data ) is provided in Figure S3 . In this same figure , sections of fluorescence time-series associated to different dynamical states are represented in different colors , for a better visualization of the correspondence between states and fluorescence values ( for simplicity , only four fluorescence ranges are distinguished ) . Our generalized TE , conditioned to the proper dynamic range , enables the reconstruction of network topologies even in the presence of light scattering artifacts . For non-locally clustered topologies we obtain a remarkably high accuracy of up to 75% of true positives at a cost of 10% of false positives . An example of the reconstruction for such a network , with , is shown in Figure 4A . For locally-clustered topologies , accuracy typically reaches 60% of true positives at a cost of 10% of false positives , and an example for is shown in Figure 5A . In both topologies , we observe that for a low fraction of false positives detection ( i . e . at high thresholds ) the ROC curve displays a sharp rise , indicating a very reliable detection of the causally most efficient excitatory connections . A decrease in the slope , and therefore a rise in the detection of false positives and a larger confidence interval , is observed only at higher fractions of false positives . The confidence intervals are broader in the case of locally-clustered topologies because of the additional network-to-network variability that results from the placement of neurons ( which is irrelevant for the generation of the non-locally clustered ensembles , see Materials and Methods ) . The performance level ( fraction of true positives for 10% of false positives , denoted by ) provides a measure of the quality of the reconstruction , and allows the comparison of different methods for different network topologies , conditioning levels , and external artifacts ( i . e . presence or absence of simulated light scattering ) . We test linear methods , XC and GC ( of order 2; the performance of GC of order 1 is very similar and not shown ) , and non-linear methods , namely MI and TE ( of Markov orders 1 and 2; see Materials and Methods for details ) . XC and MI are correlation measures , while GC and TE are causality measures . Note that , for each of these methods , we account for state dependency of functional connectivity , performing state separation as described in the Materials and Methods section . In the case of the non-local clustering ensemble and without light scattering ( Figure 6 , top row ) , even a linear method such as XC achieves a good reconstruction . This success indicates an overlap between communities of higher synchrony in the calcium fluorescence , associated to stronger activity correlations , and the underlying structural connectivity , especially for higher full clustering indices . GC-based reconstructions have an overall worse quality , due to the inadequacy of a linear model for the prediction of our highly nonlinear network dynamics , but they show similarly improved performance for higher . In a band centered around a shared optimal conditioning level , both MI and generalized TE show a robust performance across all clustering indices . This value is similar to the upper bound of the range II depicted in Figure 3A , i . e . it lies at the interface between the bursting and silent dynamical regimes . In particular for TE and in the case of low clustering indices ( which leads to networks closer to random graphs ) , conditioning greatly improves reconstruction quality . At higher clustering indices the decay in performance is only moderate for conditioning levels above the optimal value , indicating an overlap between the functional connectivities in the bursting and silent regimes . Note , on the contrary , that the performance of MI rapidly decreases if a non-optimal conditioning level is assumed . The introduction of light scattering causes a dramatic drop in performance of the two linear methods ( XC and GC ) , and even of MI and TE with Markov order . The performance of TE at Markov order 2 also deteriorates , but is still significantly above the random reconstruction baseline in a broad region of parameters . Interestingly , for the optimal conditioning level the performance of the TE for does not fall below for any clustering level or value . It is precisely in this optimal conditioning range that we obtain the linear relations between reconstructed and structural clustering coefficients , for both the non-local and the local clustering ensembles . A similar trend is obtained when varying the length scale in the local ensembles ( see Supplementary Figure S5 ) . For very local clustering and without light scattering , both XC and TE achieve performance levels up to 80% . The introduction of light scattering , however , reduces the performance of all measures except for MI ( but only in the narrow optimal conditioning range ) and for TE of higher Markov orders ( robust against non-optimal selection of conditioning level ) . Overall , the performance of the reconstruction for the local clustering ensembles is lower than for the non-locally clustered ensembles . This is also true , incidentally , in absence of light scattering since networks sampled from this ensemble tend to be very similar to purely random topologies ( of the Erdös-Rényi type , see e . g . [62] ) as soon as the length scale is sufficiently long , and for which performance is generally poorer ( cfr . top row of Figure 6 , for weak clustering levels ) . Our new TE method significantly improves the reconstruction performance compared to the original TE formulation [Eq . ( 9 ) ] . As shown in Figure 7A for both the local and the non-local clustered networks , reconstruction with the original TE formulation ( Figure 7A , blue line ) yields worse results than a random reconstruction , as indicated by the corresponding ROC curves falling below the diagonal . Such a poor performance is due in large part to “misinterpreted” delayed interactions . Indeed , by taking into account same bin interactions , a boost in performance is observed ( red line ) . Figure 7A also shows that an additional leap in performance is obtained when the analysis is conditioned ( i . e . restricted ) to a particular dynamical state of the network , increasing reconstruction quality by 20% ( yellow line in Figure 7A ) . The determination of the optimal conditioning level is discussed later and takes into account the considerations introduced above ( cfr . Figure 3 ) . Note that the introduction of same bin interactions alone ( red color curves ) or conditioning on the dynamical state of network alone ( yellow color curves ) already brings the performance to a level well superior to random performance . However , at least for our simulated calcium-fluorescence time series , a remarkable boost in performance is obtained only when the inclusion of same-bin interactions and optimal conditioning are combined together ( green color curves ) . Although , in principle , conditioning is enough to indirectly select a proper dynamical regime , the poor time-resolution of the analyzed signals ( constrained not only by the frame-rate of acquisition but also intrinsically by the kinetics of the dissociation reaction of the calcium-sensitive dye [63] ) also requires the potential consideration of causally-linked events occurring in the same time-bin . A different way to represent reconstruction performance are “Positive Precision Curves” ( as introduced in [56] and described in the Materials and Methods section ) , obtained by plotting , at a given number of reconstructed links , the “true-false ratio” ( TPR ) , which emphasizes the probability that a reconstructed link is present in the ground truth topology ( true positive ) . For the same networks and reconstruction as above , we plot the PPCs in Figure S6 ( for the ROC curves see Figure 7A ) . Over a wide range of the number of reconstructed links ( TFS ) , the PPC displays positive values of the TFR , indicative of a majority of true positives over false positives . For both the locally and non-locally clustered networks , the PPC reaches a maximum value of the TPR about 0 . 5 and remains positive up to about 18% of included links for the clustered topology , or up to 12% in the case of the local topology . In Figure 7B , we analyze the performance of our algorithm against changes of the sample size . Starting from simulated recordings lasting 1 h of real time ( corresponding to about 360 bursting events ) and with a full sample number of , we trimmed these recordings producing shorter fluorescence time series with samples , with being a divisor of the sample size . For both network topology ensembles , we found that a reduction in the number of samples by a factor of two ( corresponding to 30 minutes or about 180 bursts ) still yields a performance level of . By further reducing the sample size , we reach a plateau with a quality of for about 40 bursts ( corresponding to 6 minutes ) . All the experiments analyzed in this work are carried out with a duration between 30 and 60 minutes . Since conditioning , needed to achieve high performance , requires one to ignore a conspicuous fraction of the recorded data , we expect long recordings to be necessary for a good reconstruction , albeit the fact that it is possible to increase the signal-to-noise ratio by increasing the intensity of the fluorescent light . However , the latter manipulation has negative implications for the health of neurons due to photo-damage , limiting our experimental recordings to a maximum of 2 hours . We apply our analysis to actual recordings from in vitro networks derived from cortical neurons ( see Materials and Methods ) . To simplify the network reconstruction problem , experiments are carried out with blocked inhibitory GABA-ergic transmission , so that the network activity is driven solely by excitatory connections . This is consistent with previously discussed simulations , in which only excitatory neurons were included . We consider in Figure 8 a network reconstruction based on a 60 minutes recording of the activity of a mature culture , at day in vitro ( DIV ) 12 , in which active neurons were simultaneously imaged . A fully analogous network reconstruction for a second , younger dataset at DIV 9 is presented in Supplementary Figure S7 . In general , fluorescence data neither affected by photo-bleaching nor by photo-damage during this time , as proved by the stability of the average fluorescence signal shown in the Supplementary Figure S8A . The probability distribution of the average fluorescence signal is computed in the same way as for the simulated data . Neuronal dynamics and the calcium fluorescence display the same bursting dynamics that are well captured by the simulations , leading to a similar fluorescence distribution ( Figure 1D ) . Thanks to this similarity we can make use of the intuition developed for synthetic data to estimate an adequate conditioning level . We select therefore a conditioning level such as to exclude the right-tail of high fluorescence associated to to fully-developed bursting transient regimes . We have verified , however , that the main qualitative topological features of the reconstructed network are left unchanged when varying the conditioning level in a range centered on our “optimal” selection . More details on conditioning level selection are given in the Materials and Methods section . Reconstruction analysis is carried out for the entire population of imaged neurons . We analyze a network defined by the top 5% of TE-ranked links , as discussed in a previous section . Such choice leads to an average in–degree of about 100 , compatible with average degrees reported previously for neuronal cultures of corresponding age ( DIV ) and density [14] , [64] .
We have introduced a novel extension of Transfer Entropy , an information theoretical measure , and applied it to infer excitatory connectivity of neuronal cultures in vitro . Other studies have previously applied TE ( or a generalization of TE ) to the reconstruction of the topology of cultured networks [39] , [56] . However , our study introduces and discusses important novel aspects , relevant for applications . As shown in Figure 3 for simulated data and in Supplementary Figure S3 for actual culture data , epochs of synchronous bursting are associated to functional connectivity with a stronger degree of clustering and a weaker overlap with the underlying structural topology . This feature of functional connectivity is tightly related to the spatio-temporal organization of the synchronous bursts . In Figure S2A , we highlight the position of selected nodes ( of the simulated network considered in Figure 3 ) , characterized by an above-average in-degree of functional connectivity ( see Materials and Methods for a detailed definition in terms of TE scores ) . We denote these nodes —different in general for each of the dynamic regimes numbered from I to VII— as ( state-dependent ) functional connectivity hubs . Given a specific hub , we can then define the community of its first neighbors in the corresponding functional network . Consistently across all dynamic ranges ( but the noise-dominated range I ) we find that synchronization within each of these functional communities is significantly stronger than between the communities centered on different hubs ( ( , Mann-Whitney test , see Materials and Methods ) . The results of this comparison are reported in Figure S2B , showing particularly marked excess synchronization during burst build-up ( ranges II , III and IV ) or just prior to burst waning in the largest-size bursts ( VII ) . Therefore , functional connectivity hubs reflect foci of enhanced local bursting synchrony . Other studies ( in brain slices ) reported evidence of functional connectivity hubs , whose direct stimulation elicited a strong synchronous activation [69] , [70] . In [69] , the functional hubs were also structural hubs . In the case of our networks , however only functional hubs associated to ranges II and III have an in-degree ( and out-degree ) larger than average as in [69] . In the other dynamic ranges , this tight correspondence between structural and functional hubs does not hold anymore . Nevertheless , in all dynamic ranges ( but range I ) , we find that pairs of functional hubs have an approximately three-times larger chance of being structurally connected than pairs of arbitrarily selected nodes ( not shown ) . The timing of firing of these strong-synchrony communities is analyzed in Figure S2C . There we show that the average time of bursting of functional communities for different dynamic ranges is shifted relatively to the average bursting time over the entire network ( details of the estimation are provided in Materials and Methods ) . This temporal shift is negative for the ranges II and III ( indicating that functional hubs and related communities fire on average earlier than the rest of the culture ) and positive for the ranges V to VII ( indicating that firing of these communities occurs on average later than the rest of the culture ) . The highest negative time delay is detected in range III , such that the communities organized around its associated functional hubs can be described as local burst initiation cores [71] , [72] . In this study we did not consider inhibitory interactions , neither in simulations nor in experiments ( GABAergic transmission was blocked ) , but we attempted uniquely the reconstruction of excitatory connectivity . We would like to point out that this is not a general limitation of TE , since the applicability of TE does not rely on assumptions as to the specific nature of a given causal relationship – for instance about whether a synapse is excitatory or inhibitory . In this sense , TE can be seen as a measure for the absolute strength of a causal interaction , and is able in principle to capture the effects on dynamics of both inhibitory and excitatory connections . Note indeed that previous studies [39] , [56] have used TE to infer as well the presence of inhibitory connections . However , TE alone could not discriminate the sign of the interaction and additional post-hoc considerations had to be made in order to separate the retrieved connections into separate excitatory and inhibitory subgroups ( e . g . , in [39] , based on the supposedly known existence of a difference in relative strength between excitatory and inhibitory conductances ) . In summary , we have developed a new generalization of Transfer Entropy for inferring connectivity in neuronal networks based on fluorescence calcium imaging data . Our new formalism goes beyond previous approaches by introducing two key ingredients , namely the inclusion of same bin interactions and the separation of dynamical states through conditioning of the fluorescence signal . We have thoroughly tested our formalism in a number of simulated neuronal architectures , and later applied it to extract topological features of real , cultured cortical neurons . We expect that , in the future , algorithmic approaches to network reconstruction , and in particular our own method , will play a pivotal role in unravelling not only topological features of neuronal circuits , but also in providing a better understanding of the circuitry underlying neuronal function . These theoretical and numerical tools may well work side by side with new state-of-the-art techniques ( such as optogenetics or high-speed two-photon imaging [96]–[100] ) that will enable direct large-scale reconstructions of living neuronal networks . Our Transfer Entropy formalism is highly versatile and could be applied to the analysis of in vivo voltage-sensitive dye recordings with virtually no modifications . On a shorter time-scale , it would be important to extend our analysis to the reconstruction of both excitatory and inhibitory connectivity in in vitro cultures , which is technically feasible , and to compare diverse network characteristics , such as neuronal density or aggregation . Our algorithm could be used to systematically reconstruct the connectivity of cultures at different development stages in the quest for understanding the switch from local to global neuronal dynamics . Another crucial open issue is to design suitable experimental protocols allowing to confirm the existence of at least some of the inferred synaptic links , in order to validate statistically the reconstructed connectivity . For instance , the actual presence of directed links to which our algorithm assigns the largest TE scores might be systematically probed through targeted paired electrophysiological stimulation and recording . Furthermore , GFP transfection or inmunostaining might be used to obtain actual , precise anatomical data on network architecture to be compared with the reconstructed one . Finally , it might be interesting to reconstruct connectivity of cultured networks before and after physical disconnection of different areas of the culture ( e . g . by mechanical etching of the substrate or by chemical silencing ) . These manipulations would provide a scenario to verify whether TE-based reconstructions correctly capture the absence of direct connections between areas of the neuronal network which are known to be artificially segregated .
We generated synthetic networks with neurons , distributed randomly over a squared area of 0 . 5 mm lateral size . We chose as the connection probability between neurons [48] , leading to sparse connectivities similar to those observed in local cortical circuits [53] . We used non-periodic boundary conditions to reproduce eventual “edge” effects that arise from the anisotropic cell density at the boundaries of the culture . We considered two general types of networks: ( i ) a locally-clustered ensemble , where the probability of connection depended on the spatial distance between two neurons; and ( ii ) a non-locally clustered ensemble , with the connections engineered to display a certain degree of clustering . For the case of a non-local clustering ensemble , we first created a sparse connectivity matrix , randomly generating links with a homogeneous probability of connection across pairs of neurons . We next selected a random pairs of links and “crossed” them ( links and became and ) . We accepted only those changes that updated the clustering index in the direction of a desired target value , thereby maintaining the number of incoming as well as outgoing connections of each neuron . The crossing process was iterated until a clustering index higher or equal to the target value was reached . The overall procedure led to a full clustering index of the reference random network of ( mean and standard deviation , respectively , across 6 networks ) . After the rewiring iterations , we then achieved standard deviations from the desired target clustering value smaller than 0 . 1% for all higher clustering indices . We measured the full clustering index of our directed networks according to a common definition introduced by [65]: ( 1 ) The binary adjacency matrix is denoted by , with for a link , and zero otherwise . The adjacent matrix provides a complete description of the network topological properties . For instance , the in-degree of a node can be computed as , and the out-degree as . The total number of links of a node is given by the sum of its in-degree and its out-degree ( ) . The number of bidirectional links of a given node ( i . e . links between and so that and are reciprocally connected by directed connections ) is given by . The adjacency matrix did not contain diagonal entries . Such entries would correspond “toautaptic” links that connect a neuron with itself . Note that our directed functional connectivity analysis is based on bivariate time series , and therefore it would be structurally unfit to detect this type of links . For the case of the local clustering ensemble , two neurons separated a Euclidean distance were randomly connected with a distance dependent probability described by a Gaussian distribution , of the form , with a characteristic length scale . To guarantee that a constant average number of links was present in the network , this Gaussian distribution was rescaled by a constant pre-factor , obtained as follows . We first generated a network based on the unscaled kernel and computed the resulting number of links . With this value we then generated a final network based on the rescaled kernel . The dynamics of the generated neuronal networks was studied using the NEST simulator [101] , [102] . We modeled the neurons as leaky integrate-and-fire neurons , with the membrane potential of a neuron described by [103] , [104]: ( 2 ) where is the leak conductance and is the membrane time-constant . The term accounts for a time-dependent input current that arises from recurrent synaptic connections . In the absence of synaptic inputs , the membrane potential relaxes exponentially to a resting level set arbitrarily to zero . Stimulation in the form of inputs from other neurons increase the membrane potential , and above the threshold an action potential is elicited ( neuronal firing ) . The membrane voltage is then reset to zero for a refractory period of . The generated action potential excites post-synaptic target neurons . The total synaptic currents are then described by ( 3 ) where is the adjacency matrix , and is a synaptic time constant . The resulting excitatory post-synaptic potentials ( EPSPs ) have a standard difference-of-exponentials time-course [105] . Neurons in culture show a rich spontaneous activity that originates from both fluctuations in the membrane potential and small currents in the pre-synaptic terminals ( minis ) . The latter is the most important source of noise and plays a pivotal role in the generation and maintenance of spontaneous activity [106] . To introduce the spontaneous firing of neurons in Eq . ( 3 ) , each neuron was driven , through a static coupling conductance with strength , by independent Poisson spike trains ( with a stationary firing rate of , spikes fired at stochastic times ) . Neurons were connected via synapses with short-term depression , due to the finite amount of synaptic resources [104] . We considered only purely excitatory networks to mimic the experimental conditions in which inhibitory transmission is fully blocked . Concerning the recurrent input to neuron , the set represents times of spikes emitted by a presynaptic neuron , is a conduction delay of , while sets a homogeneous scale for the synaptic weights of recurrent connections , whose time-dependent strength depends on network firing history through the equations ( 4 ) ( 5 ) In these equations , represents the fraction of neurotransmitters in the “effective state” , in the “recovered state” and in the “inactive state” [103] , [104] . Once a pre-synaptic action potential is elicited , a fraction of the neurotransmitters in the recovered state enters the effective state , which is proportional to the synaptic current . This fraction decays exponentially towards the inactive state with a time scale , from which it recovers with a time scale . Hence , repeated firing of the presynaptic cell in an interval shorter than gradually reduces the amplitude of the evoked EPSPs as the synapse is experiencing fatigue effects ( depression ) . Random networks of integrate-and-fire neurons coupled by depressing synapses are well-known to naturally generate synchronous events [104] , comparable to the all-or-none behavior that is observed in cultured neurons [46] , [47] . To obtain in our model a realistic bursting rate [47] , the synaptic weight of internal connections was set to result into a network bursting of for all the network realizations that we studied , and in particular for any considered ( local or non-local ) clustering level . Therefore , after having generated each network topology , we assigned the arbitrary initial value of to internal synaptic weights and simulated 200 seconds of network dynamics , evaluating the resulting average bursting rate . If it was larger ( smaller ) than the target bursting rate , then the synaptic weight was reduced ( increased ) by 10% . We then iteratively adjusted by ( linearly ) extrapolating the last two simulation results towards the target bursting rate , until the result was closer than 0 . 01 Hz to the target value . The resulting used values of are provided in Table 1 . Note that we defined a network burst to occur when more than 40% of the neurons in the network were active within a time window of 50 ms . Such a criterion does not play any role in the reconstruction algorithm itself , where state selection is achieved through conditioning , but is only used for the automated generation of random networks with a prescribed bursting rate . Typically , for a fully developed burst , more than 90% of the neurons fire within a 50 ms bin , while , during inter-burst intervals , less than 10% do . Due to the clear separation between these two regimes , our burst detection procedure does not depend significantly on the precise choice of threshold within a broad interval . To reproduce the fluorescence signal measured experimentally , we treated the simulated spiking dynamics to generate surrogate calcium fluorescence signals . We used a common model introduced in [43] that gives rise to an initial fast increase of fluorescence after activation , followed by a slow decay ( ) . Such a model describes the intra-cellular concentration of calcium that is bound to the fluorescent probe . The concentration changes rapidly by a step amount of for each action potential that the cell is eliciting in a time step , of the form ( 6 ) where is the total number of action potentials . The net fluorescence level associated to the activity of a neuron is finally obtained by further feeding the Calcium concentration into a saturating static non-linearity , and by adding a Gaussian distributed noise with zero mean: ( 7 ) For the simulations , we used a saturation concentration of and noise with a standard deviation of 0 . 03 . We considered the light scattered in a simulated region of interest ( ROI ) from surrounding cells . Denoting as the distance between two neurons and and by the scattering length scale ( determined by the typical light deflection in the medium and the optical apparatus ) , the resulting fluorescence amplitude of a given neuron is given by ( 8 ) A sketch illustrating the radius of influence of the light scattering phenomenon is given in Figure S9 . The scaling factor sets the overall strength of the simulated scattering artifact . Note that light scattered , according to the equation shown above , could be completely corrected using a standard deconvolution algorithm , at least for very large fields of view and a scattering length known with sufficient accuracy . In a real setup however , the relatively small fields of view ( on the order of ) , the inaccuracies in inferring the scattering radius , as well as the inhomogeneities in the medium and on the optical system , make perfect deconvolution not possible . Therefore , artifacts due to light scattering cannot be completely eliminated [107] , [108] . The scaling factor , that we arbitrarily assumed to be small and with value , can be seen as a measure of this residual artifact component . In its original formulation [26] , for two discrete Markov processes and ( here shown for equal Markov order ) , the Transfer Entropy ( TE ) from to was defined as: ( 9 ) where is a discrete time index and is a vector of length whose entries are the samples of at the time steps , , … , . The sum goes over all possible values of , and . TE can be seen as the distance in probability space ( known as the Kullback-Leibler divergence [109] ) between the “single node” transition matrix and the “two nodes” transition matrix . As expected from a distance measure , TE is zero if and only if the two transition matrices are identical , i . e . if transitions of do not depend statistically on past values of , and is greater than zero otherwise , signaling dependence of the transition dynamics of on . We use TE to evaluate the directed functional connectivity between different network nodes . In a pre-processing step , we apply a basic discrete differentiation operator to calcium fluorescence time series , as a rather crude way to isolate potential spike events . Thus , given a network node , we define . This pre-processing step also improves the signal-to-noise ratio , thus allowing for a better sampling of probability distributions with a limited number of data points . To adapt TE to our particular problem we need to take into account the general characteristics of the system . We therefore modified TE in two crucial aspects: ( 10 ) We then include all data points at time instants in which this average fluorescence is below a predefined threshold parameter , i . e . we consider only the time points that fulfill . We only make an exception that corresponds to the simulations of Figure 3 and Figure S2 , where we considered time points that fall within an interval bounded by a higher and a lower thresholds , i . e . . Using these two novel aspects , we have extended the original description of Transfer Entropy [Eq . ( 9 ) ] to the following form ( 11 ) Probability distributions have to be evaluated as discrete histograms . Hence , the continuous range of fluorescence values ( see e . g . the bottom panels of Figure 1 ) is quantized into a finite number of discrete levels . We typically used a small , a value that we justify based on the observation that the resulting bin width is close to twice the standard deviation of the signal . The presence of large fluctuations , most likely associated to spiking events , is then still captured by such a coarse , almost non-parametric description of fluorescence levels . Generalized TE values are obtained for every possible directed pair of network nodes , and using a fixed threshold level . The set of TE scores are then ranked in ascending order and scaled to fall in the unit range . A threshold is then applied to the rescaled data , so that only those links with scores above are retained in the reconstructed network . A standard Receiver-Operator Characteristic ( ROC ) analysis is used to assess the quality of the reconstruction by evaluating the number of true positives ( reconstructed links that are present in the actual network ) or false positives ( not present ) , and for different threshold values [110] . The highest threshold value leads to zero reconstructed links and therefore zero true positives and false positives . At the other extreme , the lowest threshold provides both 100% of true positives and false positives . Intermediate thresholds give rise to a smooth curve of true/false positives as a function of the threshold . The performance of the reconstruction is then measured as the degree of deviation of this curve from the diagonal , and that corresponds to a random choice of connections between neurons . To provide a simple method to compare different reconstructions , we arbitrarily use the quantity , defined as the fraction of true positives for a 10% of false positives , as indicator for the quality of the reconstruction . An alternative to the ROC is the Positive Prediction Curve [56] , plotting the “true-false ratio” ( TFR ) against the number of reconstructed links , called “true-false sum” ( TFS ) . The TFR represents the fraction of true positives relative to the false positives . Denoting by #TP the absolute number of true positives and by #FP the number of false positives , TFR is therefore defined in [56] in the following way: ( 12 ) The case corresponds to the case that , for any given reconstructed link , it is on average equally likely that it is in fact a true positive rather than a false positive . To gain further insight into the quality of our reconstruction method , we compare reconstructions based on TE with three other reconstruction strategies , namely cross-correlation , mutual information , and Granger causality . Cross-correlation ( XC ) reconstructions are based on standard Pearson cross-correlation . The score assigned to each potential link is given by the largest cross-correlogram peak for lags between and , of the form ( 13 ) In a similar way , the scores for Mutual Information ( MI ) reconstructions are evaluated as ( 14 ) Analogously to TE , the sum goes over all entries of the joint probability matrix . For the reconstruction based on Granger causality ( GC ) [33] we first model the signal by least-squares fitting of a univariate autoregressive model , obtaining the coefficients and the residual , ( 15 ) In a second step , we fit a second bivariate autoregressive model that includes the potential source signal , and determine the residual , ( 16 ) Note that in the latter bivariate regression scheme we take into account “same bin” interactions as for Transfer Entropy ( index of the second sum starts at ) . Given , the covariance matrix of the univariate fit in Eq . ( 15 ) , and , the covariance matrix of the bivariate fit in Eq . ( 16 ) , GC is then given by the logarithm of the ratio between their traces: ( 17 ) GC analyses were performed at an order . Analyses at yielded however fully analogous performance ( not shown ) . We note that the same pre-processing used for TE is also adopted for all the other analyses . The same holds for conditioning on the value of the average fluorescence , which can be applied simply by only including the subset of samples in which . Connectivity in reconstructed networks is often inhomogeneous , and groups of nodes with tighter internal connectivity are sometimes visually apparent ( see e . g . reconstructed topologies in Figure 3C ) . We do not attempt a systematic reconstruction of network communities [111] , but we limit ourselves to the detection of “causal sink” nodes [112] , which have a larger than average in-degree . We define this property in terms of the sum of TE from all other nodes to one particular node ( ) , choosing the top 20 nodes for each particular network as selected “hub nodes” . We then analyze the dynamics of these selected hub nodes and of their neighbors . Specifically we define as the subgraph spanned by a given hub node and by its first neighbors . We analyze then the cross-correlogram of the average fluorescence of a given group with the average fluorescence of the whole culture: ( 18 ) The -notation indicates that we correlate discretely differentiated average fluorescence time series , rather than the average time series themselves . Indeed , cross-correlograms for these differentiated time series are well modeled by a Gaussian functional form , due to the slow change of the averaged fluorescence compared to the sampling rate . Therefore , we fit a Gaussian to the cross-correlogram : ( 19 ) determining thus a cross-correlation amplitude , a cross-correlation peak lag and the standard deviation . The cross-correlation peak lag indicates therefore whether nodes in a given local hub neighborhood fire on average earlier or later than other neurons in the network . Relative strength of synchrony within a local hub neighborhood can be analogously evaluated by computing XCs , as defined in Eq . ( 13 ) , for all the links within and comparing it with peak XCs over the entire network . Primary cultures of cortical neurons were prepared following standard procedures [14] , [113] . Cortices were dissected from Sprague-Dawley embryonic rat brains at 19 days of development , and neurons dissociated by mechanical trituration . Neurons were plated onto 13 mm glass cover slips ( Marienfeld , Germany ) previously coated overnight with 0 . 01% Poly-l-lysine ( Sigma ) to facilitate cell adhesion . Neuronal cultures were incubated at , 95% humidity , and for 5 days in plating medium , consisting of 90% Eagle's MEM —supplemented with 0 . 6% glucose , 1% 100× glutamax ( Gibco ) , and gentamicin ( Sigma ) — with 5% heat-inactivated horse serum ( Invitrogen ) , 5% heat-inactivated fetal calf serum ( Invitrogen ) , and B27 ( Invitrogen ) . The medium was next switched to changing medium , consisting of of 90% supplemented MEM , 9 . 5% heat-inactivated horse serum , and 0 . 5% FUDR ( 5-fluoro-deoxy-uridine ) for 3 days to limit glia growth , and thereafter to final medium , consisting of 90% supplemented MEM and 10% heat-inactivated horse serum . The final medium was refreshed every 3 days by replacing the entire culture well volume . Typical neuronal densities ( measured at the end of the experiments ) ranged between 500 and . Cultures prepared in these conditions develop connections within 24 hours and show spontaneous activity by day in vitro ( DIV ) 3–4 [14] , [18] , [20] . GABA switch , the change of GABAergic response from excitatory to inhibitory , occurs at DIV 6–7 [14] , [20] . Neuronal activity was studied at day in vitro ( DIV ) 9–12 . Prior to imaging , cultures were incubated for 60 min in pH-stable recording medium in the presence of 0 . 4% of the cell-permeant calcium sensitive dye Fluo-4-AM ( Invitrogen ) . Recording solution includes ( in mM ) : , , , , 10 glucose , and 10 HEPES; pH titrated to 7 . 4 and osmolarity to with sucrose . The culture was washed off Fluo-4 after incubation and finally placed in a chamber filled with fresh recording medium . The chamber was mounted on a Zeiss inverted microscope equipped with a 5× objective and a 0 . 4× optical zoom . Neuronal activity was monitored through high-speed fluorescence imaging using a Hamamatsu Orca Flash 2 . 8 CMOS camera attached to the microscope . Images were acquired at a speed of 100 frames/s ( i . e . 10 ms between two consecutive frames ) , which were later converted to a 20 ms resolution using a sliding window average to match the typical temporal resolution of such recordings [13] , [43] , [45] , [90] , [114] . The recorded images had a size of pixels with 256 grey-scale levels , and a final spatial resolution of . This settings provided a final field of view of that contained on the order of 2000 neurons . Activity was finally recorded as a long image sequence of 60 minutes in duration . We verified that the fluorescence signal remained stable during the recording , as shown in Supplementary Figure S8A , and we did not observe neither photo-bleaching of the calcium probe nor photo-damage of the neurons . Before the beginning of the experiment , inhibitory synapses were fully blocked with bicuculline , a antagonist , so that activity was solely driven by excitatory neurons . Since cultures were studied after GABA switch , the blockade of inhibition resulted in an increase of the fluorescence amplitude , which facilitated the detection of neuronal firing , as illustrated in Figure S8B . The image sequence was analyzed at the end of the experiment to identify all active neurons , which were marked as regions of interest ( ROIs ) on the images . The average grey-level on each ROI along the complete sequence finally provided , for each neuron , the fluorescence intensity as a function of time . Each sequence typically contained on the order of a hundred bursts . Examples of recorded fluorescence signal for individual neurons are shown in Figure 1B . The fluorescence data obtained from recordings of neuronal cultures was analyzed following exactly the same procedures used for simulated data ( e . g . processed in a pipeline including discrete differentiation , TE or other metrics evaluation , ranking , and final thresholding to maintain the top 10% of connections ) . Due to the lack of knowledge of the ground-truth topology , optimal conditioning level cannot be known . However , based on the similarity between experimental and simulated distributions of calcium fluorescence , we select a conditioning level such to exclude the high fluorescence transients associated to fully-developed bursting transients while keeping as many data points as possible . Concretely , this is achieved by taking a conditioning level equal to approximately two standard deviations above the mean of a Gaussian fit to the left peak of the fluorescence histogram . Such a level coincides with the point where , when gradually increasing the conditioning level , the reconstructed clustering index reaches a plateau , i . e . matches indicatively the upper limit of range II in Figure 3 . To check for robustness of our reconstruction , we generated alternative reconstructions based on different conditioning levels . For the selected conditioning value , and for both the experimental datasets analyzed ( Figures 8 and S7 ) , we verified that the inferred topological features , including notably the average clustering coefficient and connection distance , were stable in a range centered on the selected conditioning value and wide as much as approximately two standard deviations of the fluorescence distribution . To identify statistically significant non-random features of the real cultured networks in exam , we compared the reconstructed topology to two randomizations . A first one consisted in a complete randomization that preserved only the total number of connections in the network , but scrambled completely the source and target nodes . The resulting random ensemble of graphs was an Erdös-Rényi ensemble ( see , e . g . [62] ) in which each possible link exists with a uniform probability of connection , where is the total number of connections in the reference reconstructed network . A second partial randomization preserved the in-degree distributions only , and was implemented by shuffling the entries of each row of the reconstructed adjacency matrix , internally row-by-row . In this way , the out-degrees of each node were preserved . In both randomization processes , we disallowed diagonal entries . For both randomizations we calculated the in-degree , the distance of connections and the full clustering index for each node , leading to distributions of network topology features that could be compared between the reconstructed network and the randomized ensembles , to identify significant deviations from random expectancy . | Unraveling the general organizing principles of connectivity in neural circuits is a crucial step towards understanding brain function . However , even the simpler task of assessing the global excitatory connectivity of a culture in vitro , where neurons form self-organized networks in absence of external stimuli , remains challenging . Neuronal cultures undergo spontaneous switching between episodes of synchronous bursting and quieter inter-burst periods . We introduce here a novel algorithm which aims at inferring the connectivity of neuronal cultures from calcium fluorescence recordings of their network dynamics . To achieve this goal , we develop a suitable generalization of Transfer Entropy , an information-theoretic measure of causal influences between time series . Unlike previous algorithmic approaches to reconstruction , Transfer Entropy is data-driven and does not rely on specific assumptions about neuronal firing statistics or network topology . We generate simulated calcium signals from networks with controlled ground-truth topology and purely excitatory interactions and show that , by restricting the analysis to inter-bursts periods , Transfer Entropy robustly achieves a good reconstruction performance for disparate network connectivities . Finally , we apply our method to real data and find evidence of non-random features in cultured networks , such as the existence of highly connected hub excitatory neurons and of an elevated ( but not extreme ) level of clustering . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"circuit",
"models",
"connectomics",
"calcium",
"imaging",
"developmental",
"neuroscience",
"computer",
"science",
"neuroanatomy",
"neural",
"networks",
"computational",
"neuroscience",
"biology",
"computerized",
"simulations",
"neuroscience",
"neural",
"circuit",
"formation",
"neuroimaging"
]
| 2012 | Model-Free Reconstruction of Excitatory Neuronal Connectivity from Calcium Imaging Signals |
In Africa , relapsing fevers are neglected arthropod-borne infections caused by closely related Borrelia species . They cause mild to deadly undifferentiated fever particularly severe in pregnant women . Lack of a tool to genotype these Borrelia organisms limits knowledge regarding their reservoirs and their epidemiology . Genome sequence analysis of Borrelia crocidurae , Borrelia duttonii and Borrelia recurrentis yielded 5 intergenic spacers scattered between 10 chromosomal genes that were incorporated into a multispacer sequence typing ( MST ) approach . Sequencing these spacers directly from human blood specimens previously found to be infected by B . recurrentis ( 30 specimens ) , B . duttonii ( 17 specimens ) and B . crocidurae ( 13 specimens ) resolved these 60 strains and the 3 type strains into 13 species-specific spacer types in the presence of negative controls . B . crocidurae comprised of 8 spacer types , B . duttonii of 3 spacer types and B . recurrentis of 2 spacer types . Phylogenetic analyses of MST data suggested that B . duttonii , B . crocidurae and B . recurrentis are variants of a unique ancestral Borrelia species . MST proved to be a suitable approach for identifying and genotyping relapsing fever borreliae in Africa . It could be applied to both vectors and clinical specimens .
In Africa , relapsing fevers ( RF ) are arthropod-borne diseases caused by four cultured species Borrelia crocidurae , Borrelia duttonii , Borrelia hispanica and Borrelia recurrentis [1] . Transmission is by the bite of Ornithodoros soft ticks for the first three species whereas Pediculus humanus louse feces transmit B . recurrentis [2] , [3] . In Tanzania , molecular investigations of human and tick specimens further provided evidences for two additional , yet uncultured Borrelia species [1] , [4] . Each one of the four cultured Borrelia species is more prevalent in one geographical area of Africa with B . hispanica being reported in Morocco [5] , B . crocidurae in Senegal [6] , B . duttonii in Tanzania [7] and B . recurrentis in Ethiopia [8] . However , the precise area of distribution of each Borrelia is unknown and may overlap as both B . duttonii and B . crocidurae have been reported in Togo and Tanzania [1] , [9] . In these regions of Africa , RF was reported to be the most prevalent bacterial disease , accounting for 8 . 8% of febrile patients in Togo [9] . In Senegal , average incidence is 11 per 100 person-years [10] . The main clinical symptom of infection is recurrent undifferentiated fever associated with high bacteremia; RF are therefore often diagnosed as malaria and cases of malaria co-infection with have been reported [9] , [11] , [12] . RF are treatable by antibiotics . Severity ranges from asymptomatic to fatal , particularly if left untreated and can be associated with significant pregnancy loss or peri-natal mortality [13] , [14] , [15] . The African RF Borrelia are very closely related species as illustrated by 16S rRNA gene sequence variability ≤1% [2] . Accordingly , a previous comparison of B . duttonii and B . recurrentis genomes indicated that the two organisms formed a unique bacterial species [16] . Such a close genetic and genomic proximity challenged the development of laboratory tools for the accurate discrimination between the African RF Borrelia and genotyping [16] . Sequencing the 16S rRNA and the flagellin genes is unsatisfactory since African RF Borrelia differ by only one base in the flagllin gene sequence and have 16S rRNA gene sequence similarity above 99% [17] . Analysis of the intrergenic spacer ( IGS ) located between the 16S and 23S rRNA genes only explored the variability between B . duttonii and B . recurrentis [1] . Moreover , IGS sequence overlapped between one B . duttonii phylogenetic group and one B . recurrentis group [1] with a second overlap disclosed with subsequent analyses of further material [7] . We previously observed that multispacer sequence typing ( MST ) , a PCR-sequencing-based method for bacteria genotyping , was efficient in typing otherwise homogenous bacterial species such as the plague agent Yersinia pestis [18] and the typhus agent Rickettsia prowazekii [19] . Ongoing study of the B . crocidurae genome in our laboratory gave us the opportunity to develop MST for African RF Borrelia and to deliver the proof-of-concept that MST is a suitable method for both the species identification and genotyping of RF Borrelia in Africa .
B . crocidurae Achema strain , B . recurrentis A1 strain and B . duttonii Ly strain were grown in BSK-H medium ( Sigma , Saint Quentin Fallavier , France ) supplemented with heat-inactivated 10% rabbit serum ( Eurobio , Courtaboeuf , France ) . B . recurrentis DNA was extracted from 21 blood specimens collected in 1994 in Addis Ababa , Ethiopia Dr . S . J . Cutler ( School of Health , Sports and Bioscience , University of East London , London UK ) . Likewise , B . recurrentis DNA extracted from 9 blood specimens collected in 2011 in Bahir Dah , Highlands of Ethiopia were provided by SC Barker ( Parasitology section , School of Chemistry and Molecular Bioscience , University of Queensland , Brisbane , Australia ) and KD Bilcha and J Ali ( University of Gondar , Ethiopia ) . In addition , B . duttonii DNA extracted from 17 blood specimens collected in Mvumi , Tanzania were also provided by Dr . S . J . Cutler . B . crocidurae DNA was extracted from 13 blood specimens collected in 2010 in Senegal by C . Sokhna ( URMITE , Dakar , Senegal ) including 11 specimens from Dielmo and 2 specimens from Ndiop . DNA was extracted from these specimens using QIAamp DNA Blood mini kits ( QIAGEN , Hilden , Germany ) according to the manufacturer's instructions . The B . crocidurae genome ( Genbank accession number CP003426–CP003465 ) has been sequenced and annotated in our laboratory using pyrosequencing technology on a Roche 454 GS FLX sequencer . The draft genome is comprising of one closed chromosome and scaffolds representing the plasmids . Spacer sequences extracted from B . crocidurae strain Achema , B . recurrentis strain A1 ( Genbank accession number CP000993 ) and B . duttonii strain Ly ( Genbank accession number CP000976 ) genomes using perl script software were compared using ssaha2 software [20] . Spacers were pre-selected for a 300 to 800-bp length . Pre-selected spacers were further analyzed for sequence similarity in order to exclude spacers with <0 . 1% interspecies sequence similarity . PCR primers were then designed using primer3 software ( http://fokker . wi . mit . edu ) in order to amplify the entire sequence of each of the selected spacers . Five microliters of Borrelia DNA and 10 pmol of each primer ( Eurogentec , Seraing , Belgium ) were added to the PCR mixture , containing 0 . 4 U Phusion DNA Polymerase ( Finnzymes , Espoo , Finland ) , 4 µl of 5× Phusion HF Buffer ( Finnzymes ) and 0 . 4 µl of 10 mM dNTPs . The volume was adjusted to 24 µL by adding distilled water . Thermal cycling was performed on a 2720 DNA thermal cycler ( Applied Biosystems , Courtaboeuf , France ) with an initial 30-sec cycle at 98°C followed by 35 cycles consisting of 10 seconds at 98°C , 30 seconds at 58°C and 1 minute at 72°C , followed by a 10-min final extension step at 72°C . To rule out amplicon carry-over , nucleotide-free water negative control was used throughout the steps of the protocol . PCR products were purified prior to sequencing by using the Nucleo-Fast 96 PCR Kit ( Macherey-Nagel , Hoerdt , France ) . Three microliters of the resulting DNA were added to each primer mixture comprised of 10 pmol of each primer , 4 µL water and 3 µL BigDye Terminator reaction mix ( Applied Biosystems ) . Sequencing thermal cycling was performed on a Applied Biosystems DNA thermal cycler with an initial 5-min cycle at 96°C followed by 25 cycles consisting of 30 seconds at 96°C , 20 seconds at 55°C , and 4 minutes at 60°C , followed by a 7-min final extension step at 15°C . Sequencing products were purified using sephadex plates ( Sigma-Aldrich , Saint Quentin Fallavier , France ) and sequencing electrophoresis was performed on a 3130 Genetic Analyzer ( Applied Biosystems ) . The nucleotide sequences were edited using ChromasPro software ( www . technelysium . com . au/chromas . html ) . Similarities between spacers were determined after multiple alignments using the MULTALIN software [21] . MST discrimination power was calculated using the Hunter-Gaston Index [22]:where D is the numerical index of discrimination , N is the total number of isolates in the sample population , s is the total number of different types , and nj is the number of isolates belonging to the jth type . The five spacer sequences analyzed herein were concatenated and neighbor-joining phylogenetic tree was reconstructed using the maximum likelihood method in PhyML 3 . 0 [23] . Each particular sequence of a given spacer was assigned to a spacer type ( ST ) number . This study was approved by the IFR48 Ethic Committee . All patients provided informed written consent .
Chromosome sequence alignment of the three Borrelia reference genomes studied herein revealed that 23 intergenic spacers that were common to all three species . Of these , five spacers fulfilled our selection criteria and were named MST2 , MST3 , MST5 , MST6 and MST7 . Use of the PCR primers listed in Table 1 to amplify each of the five spacers produced amplicons ranged from 333-bp to 738-bp and sequence reads ranging from 246-bp to 543-bp ( Table 1 ) . Pairwise comparison of the five spacers ( Table 2 and figure 1 ) revealed they had species-specific sequence with interspecies sequence differences relying on single nucleotide polymorphism in 36 ( 90% ) cases , deletion in 3 ( 7 . 5% ) cases and insertion in 1 ( 2 . 5% ) case . Comparing B . duttonii MST7 with B . crocidurae and B . recurrentis MST7 yielded 93% and 97% similarity , respectively , whilst comparing B . crocidurae MST7 with B . recurrentis MST7 showed 93% similarity . The other four spacers yielded pairwise sequence similarity of 97–99% ( Table 2 ) . Sequences for each allele of each spacer have been deposited in GenBank under accession number ( JQ398815: JQ398841 ) as well as in our local data base ( http://www . ifr48 . com ) . While the concatenation of the five spacers yielded a discrimination index of 0 . 825 , 1 , this index was of 0 . 7814 for MST2 , 0 . 6896 for MST6 , 0 . 6749 for MST5 , 0 . 6623 for MST7 , and 0 . 6579 for MST3 . Concatenation of the five spacers yielded 8 STs named ST6–ST13 for the 13 B . crocidurae samples and the B . crocidurae Achema type strain ( Table 3; Figure 2 ) . 3 STs named ST1–ST3 for the 18 B . duttonii samples and the B . duttonii Ly type strain and 2 STs named ST4–ST5 for the 30 B . recurrentis samples and the B . recurrentis A1 type strain . MST2 sequencing classified latter samples into ST-4 ( 11 samples ) and ST-5 ( 19 samples ) due to the insertion of a G at position 190 . The genotype ST-5 represented 47 . 6% ( 10 out 21 samples ) detected in 1994 and all the nine samples detected in 2011 . The phylogenetic tree constructed after concatenation of the five intergenic spacer sequences separated the RF Borrelia into three clades , each clade containing only one Borrelia species ( figure 2 ) . A first clade comprised of all the 30 B . recurrentis isolates; a second clade comprised of three groups representing the three B . duttonii spacer types and a last clade comprised of 7 B . crocidurae spacer types .
PCR-derived data reported herein were interpreted as authentic as the negative controls used in every PCR-based experiment remained negative , all the PCR products were sequenced and experiments yielded reproducible sequences . We therefore established the proof-of-concept that MST could be used for species identification and genotyping of 3 out of 4 cultured RF borreliae ( B . hispanica was not available for this study ) in Africa . MST combines the sensitivity of PCR with unambiguous , portable data yielded by sequencing . Indeed , all the sequences determined are freely available in GenBank and in our local database website at ifr48 . com . Therefore , any laboratory with a capacity in PCR-sequencing could easily confirm and compare their data with that reported herein to further increase the knowledge of RF Borrelia species and genotypes circulating in African countries . In the present study , five intergenic spacers were selected from the alignment of B . crocidurae , B . duttonii and B . recurrentis reference genomes , representing approximately ∼0 . 2% of the total genome length . The spacers were scattered across the chromosome thus representative of the whole genome . Such a multi-target approach offers distinct advantages over the one single locus methods previously used , such as the 16S–23S IGS for typing that may be less representative of the whole genome . Based on this spacer sequencing , a total of 61 RF strains could be separated into 12 STs . Interestingly , we observed that isolates grouped into three clades corresponding to the three Borrelia organisms under study . Indeed , MST yielded no overlap between B . duttonii and B . recurrentis organisms contrary to that observed when using IGS typing [1] , [7] . We observed that sequencing MST7 spacer alone accurately discriminated between B . duttonii and B . recurrentis with 3% sequence divergence , a result not previously achieved . Therefore , sequencing MST7 spacer alone could be used for the molecular identification of RF Borrelia in Africa at the species level , but not for genotyping which requires sequencing the four other spacers in addition to MST7 . Further analysis indicated that each one of the three Borrelia species under study was comprised of several spacer-types . B . recurrentis was the least diverse Borrelia comprising of only two very closely related groups . This finding supports the previous genomic analysis that concluded that B . recurrentis was a subset of B . duttonii [16] . In our study also , there was an inverse correlation between the RF Borrelia MST diversity and the reported mortality rate for these RF Borrelia [8] , [15] . Despite the fact that we tested a small set of B . crocidurae , nevertheless we found a high diversity index in this species since 13 B . crocidurae samples collected in Senegal yielded 7 MST types and the B . crocidurae Achema type strain collected in Mauritania yielded an additional MST type . This first genotyping method for B . crocidurae is therefore very promising to probe its geographic repartition as well as potential association of B . crocidurae genotypes with vectors . Indeed , four genogroups could be identified in O . sonrai ticks collected in Senegal and Mauritania [24] . In this study , B . crocidurae flagellin sequence was found identical among the four O . sonrai tick groups but the B . crocidurae infection rate significantly differed among the four tick groups; MST may help studying such discrepancy and may reveal previously unknown relationships between B . crocidurae genotypes and O . sonrai genotypes . Moreover , a recent study indicated that B . crocidurae may be transmitted by soft tick Ornithodoros erraticus in Tunisia , challenging O . sonrai as the only B . crocidurae vector in West Africa [25] . MST is new laboratory tool to question whether the unexpected higher diversity in B . crocidurae than in B . duttonii and B . recurrentis is linked to a more complex cycle involving several mammals and ticks species . Present data indicate that MST is offering a new sequencing-based technique for further exploring the identification and genotypes of RF Borrelia in vectors and clinical specimens collected in Africa . | In Africa , relapsing fevers are caused by four cultured species: Borrelia crocidurae , Borrelia duttonii , Borrelia hispanica and Borrelia recurrentis . These borreliae are transmitted by the bite of Ornithodoros soft ticks except for B . recurrentis which is transmitted by louse Pediculus humanus . They cause potentially undifferentiated fever infection and co-infection with malaria could also occur . The exact prevalence of each Borrelia is unknown and overlaps between B . duttonii and B . crocidurae have been reported . The lack of tools for genotyping these borreliae limits knowledge concerning their epidemiology . We developed multispacer sequence typing ( MST ) and applied it to blood specimens infected by B . recurrentis ( 30 specimens ) , B . duttonii ( 18 specimens ) and B . crocidurae ( 13 specimens ) , delineating these 60 strains and the 3 type strains into 13 species-specific spacer types . B . crocidurae strains were classified into 8 spacer types , B . duttonii into 3 spacer types and B . recurrentis into 2 spacer types . These findings provide the proof-of-concept that that MST is a reliable tool for identification and genotyping relapsing fever borreliae in Africa . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"medicine",
"bacterial",
"diseases",
"infectious",
"diseases",
"borrelia",
"infection",
"zoonoses",
"neglected",
"tropical",
"diseases"
]
| 2012 | Multispacer Sequence Typing Relapsing Fever Borreliae in Africa |
The Complementarity Determining Regions ( CDRs ) of antibodies are assumed to account for the antigen recognition and binding and thus to contain also the antigen binding site . CDRs are typically discerned by searching for regions that are most different , in sequence or in structure , between different antibodies . Here , we show that ∼20% of the antibody residues that actually bind the antigen fall outside the CDRs . However , virtually all antigen binding residues lie in regions of structural consensus across antibodies . Furthermore , we show that these regions of structural consensus which cover the antigen binding site are identifiable from the sequence of the antibody . Analyzing the predicted contribution of antigen binding residues to the stability of the antibody-antigen complex , we show that residues that fall outside of the traditionally defined CDRs are at least as important to antigen binding as residues within the CDRs , and in some cases , they are even more important energetically . Furthermore , antigen binding residues that fall outside of the structural consensus regions but within traditionally defined CDRs show a marginal energetic contribution to antigen binding . These findings allow for systematic and comprehensive identification of antigen binding sites , which can improve the understanding of antigenic interactions and may be useful in antibody engineering and B-cell epitope identification .
Antibody-Antigen ( Ab-Ag ) interactions are based on non-covalent binding between the antibody ( Ab ) and the antigen ( Ag ) . Correct identification of the residues that mediate Ag recognition and binding would improve our understanding of antigenic interactions and may permit the modification and manipulation of Abs . For example , introducing mutations into the V-genes has been suggested as a way to improve Ab affinity [1]–[3] . However , mutations in the framework regions ( FRs ) rather than in the Ag binding residues themselves are more likely to evoke an undesired immune response [4] . Knowing which residues bind the Ag can help direct such mutations and be beneficial to Ab engineering [5]–[7] . It has been shown that Ag binding residues are primarily located in the so called complementarity determining regions ( CDRs ) [7]–[9] . Thus , the attempt to identify CDRs , and particularly the attempt to define their boundaries , has become the focus of extensive research over the last few decades [7] , [8] , [10] . Kabat and co-workers [9] , [11] attempted to systematically identify CDRs in newly sequenced Abs . Their approach was based on the assumption that CDRs include the most variable positions in Abs and therefore could be identified by aligning the fairly limited number of Abs available then . Based on this alignment they introduced a numbering scheme for the residues in the hypervariable regions and determined which positions mark the beginning and the end of each CDR . The Kabat numbering scheme was developed when no structural information was available . Chothia et al . [12] , [13] analyzed a small number of Ab structures and determined the relationship between the sequences of the Abs and the structures of their CDRs . The boundaries of the FRs and the CDRs were determined and the latter have been shown to adopt a restricted set of conformations based on the presence of certain residues at key positions in the CDRs and the flanking FRs . This analysis suggested that the sites of insertions and deletions in CDRs L1 and H1 are different than those suggested by Kabat . Thus , the Chothia numbering scheme is almost identical to the Kabat scheme , but based on structural considerations , places the insertions in CDRs L1 and H1 at different positions . As more experimental data became available , the analysis was performed anew , re-defining the boundaries of the CDRs . These definitions of CDRs are mostly based on manual analysis and may require adjustments as the structure of more Abs become available . Abhinandan et al . [14] aligned Ab sequences in the context of structure and found that approximately 10% of the sequences in the manually annotated Kabat database have erroneous numbering . A more recent attempt to define CDRs is that of the IMGT database [15] which curates nucleotide sequence information for immunoglobulins ( IG ) , T-cell receptors ( TcR ) and Major Histocompatibility Complex ( MHC ) molecules . It proposes a uniform numbering system for IG and TcR sequences , based on aligning more than 5000 IG and TcR variable region sequences , taking into account and combining the Kabat definition of FRs and CDRs [16] , structural data [17] and Chothia's characterization of the hypervariable loops [12] . Their numbering scheme does not differentiate between the various immunoglobulins ( i . e . , IG or TcR ) , the chain type ( i . e . , heavy or light ) or the species . A drawback of these numbering schemes is that CDRs length variability is accommodated with either annotation of insertion ( Kabat and Chothia ) or by providing excess numbers ( IMGT ) . Abs with unusually long insertions may be hard to annotate this way , and therefore their CDRs may not be identified correctly . Honegger and Pluckthun [18] suggested a structurally improved version of the IMGT scheme . Instead of introducing unidirectional insertions and deletions as in the IMGT and Chothia schemes , they were placed symmetrically around a key position . MacCallum et al . [8] have proposed focusing on the specific notion of Ag binding residues rather than the more vague concept of CDRs . They suggested that these residues could be identified based on structural analysis of the binding patterns of canonical loops . Other studies have dubbed those Ag binding residues Specificity Determining Regions ( SDRs ) [5] , [7] . Here , we analyze Ag-Ab complexes and show that virtually all Ag binding residues fall within regions of structural consensus . We refer to these regions as Ag Binding Regions ( ABRs ) . We show that these regions can be identified from the Ab sequence as well . We used “Paratome” , an implementation of a structural approach for the identification of structural consensus in Abs [19] . While residues identified by Paratome cover virtually all the Ag binding sites , the CDRs ( as identified by the commonly used CDR identification tools ) miss significant portions of them . We refer to the Ag binding residues which are identified by Paratome but are not identified by any of the common CDR identification methods , as Paratome-unique residues . Similarly , Ag binding residues that are identified by any of the common CDR identification methods but are not identified by Paratome are referred to as CDRs-unique residues . We show that Paratome-unique residues make crucial energetic contribution to Ab-Ag interactions , while CDRs-unique residues have a rather minor contribution . These results allow for better identification of Ag binding sites and thus for better identification of B-cell epitopes . They may also help improve vaccine and Ab design .
The outline of our structure-based ABRs identification method is delineated in Figure 1 . Briefly , the algorithm structurally aligns all known Abs and marks the residues that contact the Ag in each of them . We have shown [19] , [20] that in this multiple structure alignment there is a consensus among Abs that some structurally aligned positions contact the Ag . These positions form six sequence stretches along the Ab sequence that roughly correspond to the six CDRs . Beyond the edges of these stretches there were no structurally aligned positions in which more than 10% of the Abs contact the Ag . Thus , we defined the boundaries of the ABRs based on these stretches and marked the ABRs in all the Abs in our dataset . Figure 2 depicts the automated ABRs identification tool we developed . Given a query sequence ( Figure 2A ) a BLAST search is performed against all Abs in the dataset described above . The best hit ( i . e . , lowest E-value ) is used to infer the positions of the ABRs in the query sequence , based on its alignment to the annotated Ab from the dataset . When the query Ab has a known 3-D structure , it can be used to identify the ABRs as described in Figure 2B ( see Methods ) . Figure 3 summarizes the number of residues identified by each method on the test set . In all regions except L1 and H2 , Paratome identified a slightly larger number of residues than any other method . The largest differences were recorded in L2 and H2 . In L2 , Paratome had 50% more residues identified than Kabat and Chothia and four times the number of residues identified by IMGT . For H2 , Kabat and Paratome identified twice the number of residues suggested by Chothia and IMGT . For each Ab in our test dataset we recorded the average recall of the residues that actually bind the Ag by each method . Given the typical trade-off between recall and precision in which the increase of one is at the cost of decreasing the other , we measured the average precision of each method . The results are presented in Figure 4 . The ABRs identified by Paratome included 94% of Ag binding residues , followed by Kabat ( 85% ) , IMGT ( 81% ) and Chothia ( 79% ) CDRs . Precision rates ranged between 48% ( IMGT ) and 41% ( Kabat ) , with Chothia ( 44% ) and Paratome ( 42% ) in between . Table 1 compares the consensus sets and the method specific sets of residues . The Paratome-Kabat consensus set is the largest ( 3476 residues ) , covering 83 . 54% of the Ag binding sites . Paratome-Chothia consensus set covered 77 . 08% of the Ag binding sites ( 3203 residues ) , and Paratome-IMGT consensus set covered 79 . 47% of the Ag binding sites ( 3077 residues ) . In all consensus sets , approximately 50% of the residues are Ag binding residues . ΔParatome contains a substantially larger percentage of Ag binding residues than ΔKabat , ΔChothia and ΔIMGT ( 20 . 8% , 26 . 23% and 20 . 6% respectively , compared with 5 . 03% , 4 . 88% and 6 . 88% respectively ) . Moreover , ΔParatome residues cover a significantly larger portion of the Ag binding sites . ΔParatome residues covered 10 . 77% of the Ag binding sites while ΔKabat covered merely 1 . 78% of the Ag binding sites . The coverage of ΔParatome ( 14 . 84% ) was 20 times larger than that of ΔIMGT ( 0 . 76% ) . When compared to Chothia , the coverage of ΔParatome ( 17 . 23% ) was , again , more than an order of magnitude greater than that of ΔChothia's ( 0 . 86% ) . In each comparison , Paratome-specific residues covered a significantly larger portion of the Ag binding sites than the alternative method-specific residues . Thus , indicating that structural consensus regions capture more of the Ag binding portion of Abs . Figure 5A shows the average precision for each ABR/CDR on the light and heavy chains as defined by each of the methods . L2 has the lowest precision in all methods . For L3 , all the methods have a similar precision , with a slightly higher rate for Paratome ( 0 . 55 ) . IMGT has the highest precision for L1 ( 0 . 46 ) , followed by Paratome ( 0 . 38 ) and Chothia and Kabat has the lowest precision ( 0 . 27 ) . The largest difference between the methods is in H2 where Chothia has the highest precision ( 0 . 69 ) , followed by IMGT ( 0 . 57 ) , then Paratome ( 0 . 43 ) and Kabat ( 0 . 37 ) . Figure 5B summarizes the average recall of each method for each of the six regions . For all methods , L2 has the lowest recall ( 2–7% ) . This is expected considering L2 has the lowest precision ( see Figure 5A ) . For L1 , all methods show similar recall ( 11–12% ) . The same holds for H3 , which covers the largest fraction of the Ag binding sites ( 24–25% ) . H2 shows the highest diversity; For Paratome and Kabat it covers 21% of the Ag binding sites while for Chothia and IMGT recall ranged between 13–15% , respectively . In all cases , Paratome shows the highest recall . Note that while the overall recall ranges between 0 . 7–1 ( see Figure 4 ) , the recall of each of the six regions ranges between 0–0 . 3 . This is due to the fact that the total recall is the accumulation of the recall obtained by each of the six regions . To gain insight into the extent to which Paratome-unique residues contribute to Ag binding , we searched the non-redundant set of Abs for Ag binding residues residing within structural consensus regions that are not identified by any of the CDR identification methods . We obtained 153 Paratome-unique residues , originating from 104 Abs ( Table S3 ) . Using the FoldX algorithm [21] , [22] , we performed an in-silico alanine scan in which each Paratome-unique residue and each Ag binding residue identified by the CDR identification methods ( 2707 residues ) within the 104 Abs were mutated to Alanine . Additionally , we searched the non-redundant set of Abs for Ag binding residues residing within CDRs that are not identified by Paratome ( i . e . CDRs-unique residues ) . We found 59 CDRs-unique residues , stemming from 41 Abs ( Table S4 ) . To each CDRs-unique residue we performed an in-silico alanine scan in which it was mutated to Alanine . The distribution of the predicted interaction energy ( ΔΔG ) of these mutants is presented in figure 6A . Destabilizing residues in this analysis ( ΔΔG>0 . 25 ) are residues whose mutation to alanine is predicted to destabilize the Ab-Ag complex . These residues , therefore , are likely to be important for Ag binding . Paratome-unique residues have a slightly higher percentage of destabilizing residues ( 49% ) than Ag binding residues that fall within the CDRs according to Kabat , Chothia or IMGT ( 44 . 15% ) . While it is not clear whether the differences between Paratome-unique and Ag binding residues within the CDRs are significant , it is obvious that the former are at least as important to stability as the latter . In contrast , CDRs-unique residues have substantially lower contribution to binding: only 27% of them are destabilizing and the vast majority of them ( 70% ) are neutral . To demonstrate the importance of Paratome-unique residues we show a more detailed analysis of the complex of IL-15 with an anti-IL-15 Ab ( PDB ID 2xqb ) . Two Ag binding residues , LEU46 and TYR49 , which were identified by Paratome to be part of ABR L2 , were not identified by any of the CDR identification methods ( Table S1 ) . Figure 6B shows these residues relative to the surface of the Ag . It can be seen that TYR49 protrudes into the surface of the Ag , while LEU46 is located opposite to the antigenic LEU52 , forming a hydrophobic interaction . As shown is Figure 6C , only seven residues from L2 interact with the Ag , and two of them are Paratome-unique residues . TYR49 forms one of the two hydrogen bonds between the Ag and ABR L2 . The results of the FoldX in-silico single-point mutations analysis indicate that mutating ARG50 , ARG53 and TYR49 to Alanine have the most significant destabilizing effect ( Table S2 ) . Not surprisingly , due to the salt bridge it forms with antigenic GLU46 , mutating ARG50 had the most prominent destabilizing effect . The next most destabilizing mutation to Alanine was of TYR49 which forms a hydrogen bond between ABR L2 and antigenic GLU53 . The third most destabilizing mutation to Alanine was of ARG53 , which forms a cation-п interaction with TYR49 . As expected , mutating LEU46 to Alanine has a weak destabilizing effect on the binding energy . Hence , Ag binding residues within the structural consensus regions that fall outside the CDRs may play a pivotal role in Ag binding and recognition . The amino acid composition of Paratome-unique residues is presented in Table S8 .
Ab-Ag recognition is the basis for the vast usage of Abs for molecular identification in research and in the clinic [23]–[26] . Thus , identifying Ag binding sites facilitates the understanding of the underlying biology as well as Ab design and engineering . In a previous study [19] , we have shown that structural analysis can lead to the identification of residues that roughly correspond to the CDRs . Here we further developed this approach , and tried to determine whether it can be used to identify the Ag binding regions within Abs . To our knowledge , this study is the first to quantitatively compare the residues identified by the most commonly used CDR identification methods . The residues that reside within the structural consensus regions cover most of the observed Ag binding residues ( 94% ) , a significantly higher coverage than with the other methods . The coverage obtained by Kabat , Chothia and IMGT stemmed almost entirely from the residues that were within the structural consensus regions . While CDR residues unique to Kabat , Chothia and IMGT comprised less than 2% of the Ag binding sites , ABRs residues unique to Paratome covered 10–17% of the Ag binding residues . Nevertheless , there are cases in which the structural consensus regions did not contain Ag binding residues while a CDR identification method identified them . For a detailed example , see Figure S1 . Approximately 2% of the Ag contacting residues are located remotely from the ABRs/CDRs and thus should be considered as true negatives . Therefore , the actual recall of Paratome is 96% . Interestingly , all Paratome-unique residues come from either L2 or H2 . However , when we compare each method separately to Paratome there are differences in other CDRs as well . Table S9 shows the distribution of CDRs from which the Ag binding residues that are identified by Paratome but are missed by one of the other methods originated , in a pairwise comparison . MacCallum et al . [8] demonstrated that for some of the CDRs , the residues that contact the Ag correspond better with the Kabat definition of CDRs than with that of Chothia . This finding may , to some extent , explain the fact that the ABRs residues have the highest overlap with the residues identified by Kabat , and that for both H2 rather than H3 , comprises the largest number of residues . Attempts to increase Ab affinity have suggested that CDRs L3 and H3 are prevalently responsible for high energy interactions with the Ag [27] , [28] . This coincides with our observation that ABRs/CDRs L3 and H3 have the largest fraction of Ag binding residues for both Paratome and Kabat . For Chothia and IMGT , however , the CDRs with most Ag contacting residues are H2 and H3 ( Figure 5A ) . Notably , for all methods except for Chothia , H2 and H3 rather than L3 and H3 , cover a significantly larger percentage of the Ag binding residues ( Figure 5B ) . This analysis of Ag binding residues recognition demonstrates that relying on structural consensus rather than sequence differences , enables to identify Ag binding residues significantly better than the commonly used CDR identification methods . Additionally , a detailed in-silico single point mutation analysis of all Ag binding residues demonstrates that Paratome-unique residues contribute to Ag binding at least as much as residues within the CDRs and substantially more than Ag binding residues that are not identified by Paratome and are identified by CDR identification methods . This may prove useful for applications aimed at identifying and manipulating Ag binding residues .
The outline of our structure-based ABRs identification method is delineated in Figure 1 . To identify all Ab-Ag structures in the PDB [29] we performed a BLAST [30] search against the August 2009 version of the PDB using an arbitrarily chosen Fab sequence as a query . The search was performed separately for the light and heavy chains and thus two lists were obtained , a heavy chains list ( 2000 chains from 962 structures ) and a light chains list ( 2500 chains from 1047 structures ) . To obtain an E-value cut-off that will ascertain that the hits for the light chain do not contain any heavy chains and vice versa , we performed a BLAST search using the heavy chain of the query Fab against the hits of light chain and another BLAST search using the light chain of the query Fab against the hits of the heavy chain . Based on these analyses we determined that results with an E-value≤1e-6 should be further analysed ( 1280 heavy chains from 855 structures and 1846 light chains from 961 structures remained ) . To discard all T-cell receptors or MHC molecules complexes from our lists , we searched for a BLAST E-value that will exclude all T-cell receptors and MHC molecules from the dataset . We arbitrarily chose MHC-I , MHC-II , TCR type A and TCR Type B sequences and performed a separate BLAST search against the hits of light and heavy chains . Results with an E-value of 1e-6 , 1e-6 , 1e-12 and 1e-28 , respectively , or smaller , were discarded . Furthermore , we removed files that contained the keywords TcR or MHC , duplicate chains from the same PDB and complexes that did not contain both a heavy Ab chain and a light Ab chain . This resulted in a list containing 1568 Ab chains from 784 structures . We then screened the list so each complex holds one heavy Ab chain , one light Ab chain , and a single Ag chain which is not an Ab and contains at least five amino acids . We did not include non-peptide Ags in the analysis . The final list from which we removed redundancy contained 352 structures . Redundancy removal was performed using Blastclust [31] with sequence identity ≥97% and coverage ≥95% . We ran Blastclust separately for the sequences of the light chains and for the sequences of the heavy chains and obtained 96 clusters , 48 for each . To determine which sequences to remove in each cluster , we chose the Ab-Ag interactions as the distinguishing criterion for redundancy removal . For each PDB complex in a given Blastclust cluster , we identified all residue-residue contacts ( see below for contact definition ) . The similarity between any two complexes ( i . e . , lists of Ab-Ag contacts ) within each cluster was measured as the number of identical contacts ( i . e . , the same amino acid and alignment position within the Ab and the same amino acid and position within the Ag ) divided by the total number of contacts in the shorter of the two lists . Since the similarity score on its own is not sufficient for separating the non-redundant complexes from the redundant ones , we plotted a histogram of the similarity scores to obtain a discriminating cut-off . Most of the complexes in any given Blastclust cluster had a similarity score greater than 0 . 90 while only 25% of all complexes had a similarity score smaller than 0 . 77 . Therefore the latter was chosen as the cut-off , rendering complexes with a similarity score <0 . 77 non-redundant . For each group of complexes with a similarity score above the cut-off , the complex with the highest number of interactions was chosen as the representative complex . This process removed 152 redundant complexes and the resulting non-redundant set included 200 experimentally determined 3-D structures of Ab-Ag complexes from the PDB . Using a structure-based approach [19] that is presented in Figure 1 , we determined the ABRs of the Abs in our dataset of non-redundant known Ab-Ag complexes from the PDB . The algorithm structurally aligns all Abs whose 3-D structure was experimentally determined bound to their protein Ag , using the MUSTANG multiple structure alignment algorithm [32] . Next , it marks the residues in each structure that contact the residues on the Ag . Then , it searches for structurally aligned positions that create such contacts across at least 10% of the Abs . These positions form six sequence stretches along the Ab sequence that correspond to the six CDRs . Beyond the edges of these clusters , there were no structurally aligned positions in which more than 10% of the Abs created contacts with the Ag . Therefore , these were defined as the ABRs edges . Applying this algorithm to the dataset , we automatically identified all ABRs defined by our method without any manual intervention . Kabat , Chothia and IMGT establish Ab sequence numbering schemes that define in a straightforward manner the location of the CDRs within the sequence . Applying the various numbering schemes to the Ab sequences in our dataset we obtained the residues composing the CDRs according to each of the methods . We used the online AbNum tool [8] to number the Abs in our dataset according to Kabat and Chothia . The boundaries of the CDRs were defined as described in AbNum ( see table of CDR definitions [33] ) . To obtain the CDRs according to IMGT , we coupled the Kabat numbering obtained by applying AbNum with a conversion code available at the IMGT web site . Considering that the three dimensional structure of most known Abs is not yet known , the ability to identify the ABRs based merely on its sequence , is highly desirable . We constructed an automated ABRs identification tool capable of identifying the ABRs of an Ab from its amino acid sequence . Given a query sequence , the tool works as follows: First , a BLAST search is performed using the query Ab sequence , against all Abs in our dataset . As described above , all Abs in this non-redundant set were annotated and the ABRs within each of them were identified based on the multiple structure alignment ( MSTA ) . The best hit from this BLAST search ( i . e . , lowest E-value ) is then used to infer the positions of the ABRs in the query sequence from the annotated sequence . Since our aim is to identify the ABRs , aligning the entire input Ab sequence , including its yet unidentified stretches of ABRs , may lead to misalignments . Therefore , standard application of global sequence alignment algorithms is not suitable for this task . To overcome this , we align only the FRs , using the Smith-Waterman local alignment algorithm [34] . This is based on the premise that the FRs originate from a limited set of genes and undergo fewer mutations than the highly variable ABRs [25] . Therefore aligning the FRs is less error-prone than aligning the entire Ab . The FRs of the annotated sequence are identified based on the location of the ABRs in the MSTA , which in turn is used to identify the FRs of the query sequence and thus identifying the ABRs . Figure 2A summarizes the process of sequence based ABRs identification . When the query Ab has a known 3-D structure , it can be used to identify the ABRs . The first stage in this case is , once again , to identify the top BLAST hit for the sequence of the query Ab . Next , the query Ab and the top BLAST hit in the non-redundant Abs set are structurally aligned using the Combinatorial Extension ( CE ) algorithm [35] . Finally , the MSTA and the pairwise structural alignment are used to transfer the locations of the ABRs from the annotated Ab structure to the structure of the query Ab . Figure 2B summarizes the process of structure based ABRs identification . To assess the identification of Ag binding sites , we curated a test set of Ab-Ag complexes that were not used to construct our tool . We started by recording all Ab complexes added to the PDB between September 1st 2009 and February 22nd 2011 ( 8522 entries ) . We identified which of these entries are Abs by performing a BLAST search against the same arbitrarily chosen light and heavy chain Ab Fabs as well as MHC-I , MHC-II , TCR-A and TCR-B sequences described previously at the process of constructing the initial dataset of Ab-Ag complexes ( see Extraction of 3D structures ) . A PDB entry was retained if i ) At least one of its chains was similar to the light chain ( e-value≤2e-32 ) , ii ) At least one of its other chains ( belonging to the same biological unit , as defined by remark 350 within the PDB file ) was similar to the heavy chain ( e-value≤2e-26 ) and iii ) Neither of these chains was similar to any of the MHC or TCR molecules . The resulting list was further analysed to include only interacting Ab-Ag complexes . An Ab-Ag complex was defined as interacting if at least one of the atoms in the Ab molecule was ≤6 Å from any atom of the Ag molecule [36] . The final list contains 69 distinct Ab-Ag complexes . For each Ab in the test set of 69 Ab-Ag complexes , we identified all Ab residues that contact the Ag ( see next paragraph for definition of contact ) . Then we applied to each Ab the implementation of the CDR identification tools of Kabat , Chothia and IMGT . Finally , we used the structure based approach described above to identify the ABRs . Thus , each Ab in the test dataset has five lists , the list of all Ab binding residues ( “gold standard” ) and four lists of identified residues , one for each of the compared methods . Applying the sequence based approach presented above to the test set resulted in identification of a set of Ag binding residues that is 99% identical to the set identified using the structural approach . An Ab amino acid and an Ag amino acid are considered as interacting if at least one of their respective atoms were ≤6 Å of each other [36] . While this is a permissive cut-off that introduces into the dataset many non-interacting residues , it allows for an unbiased dataset [37] . To demonstrate that the results are not sensitive to this choice , we repeated the analysis also for distance cut-offs of 4 Å , 4 . 5 Å , 5 Å and 5 . 5 Å ( Figure S2 ) . The results show that the recall of Paratome remains superior to that of all other methods , and that the inevitable changes in recall and precision that stem from changing the positive set have the same effect on all the methods . For each method and for each Ab within the test dataset we recorded: We then performed a pairwise comparison between Paratome and each of the other methods . We examined three sets of residues: the consensus residues ( i . e . residues identified by both our method and the other method ) , and two sets of method-specific residues ( i ) residues identified by Paratome yet not by the other method ( ΔParatome ) and ( ii ) residues identified by the other method yet not by Paratome ( e . g . , ΔKabat . ) . For each of the three sets , we recorded the number of residues it includes , the number of Ab binding residues and what fraction of the Ag binding sites is covered by these residues . Figure 7 shows an example of the comparison of method-specific residues and consensus residues for all methods , for one Ab light chain . All data are available in tables S6 and S7 . To assess the performance of the extracted sets according to each of the methods , we computed for each Ab , the precision: ( 1 ) where tp is the number of true positive predictions , namely the number of residues predicted to bind the Ag that were observed in the structure to be in contact with the Ag . fp ( false positive ) is the number of residues predicted to bind the Ag that were not observed to be in contact with the Ag . Measures such as precision are typically complemented by the recall measure , defined as: ( 2 ) where fn is the number of false negative predictions ( i . e . residues not predicted to bind the Ag that were observed to be in contact with the Ag ) . The average recall and precision were computed for the Abs in the test set . The ABRs/CDRs were obtained for each of the Abs in the non-redundant train set according to Kabat , Chothia , IMAGT and Paratome as described above . Next , all Ag binding residues were extracted from the PDB structure , and mapped to the ABRs and CDRs . We then searched for Ag binding residues that are identified by Paratome but not by any CDR identification method . Next , we searched for Ag binding residues that are identified by any of the CDR identification methods but not by Paratome ( Table S3 , S4 and S5 ) . To assess the contribution of Paratome-unique Ag binding residues to Ab-Ag binding , we used the FoldX algorithm [21] , [22] , a molecular modelling software that computationally predicts the effect of mutations on the binding energy . It is based on empirical energy terms correlated with experimental ΔΔG measurements [21] . We computed the effect of mutating each Ag binding residue to Alanine on Ab-Ag binding energy ( ΔΔG ) . We applied FoldX ( version 3 . 0b4 ) to all Paratome-unique residues ( Table S3 ) and to all CDRs-unique residues ( Table S4 ) , in the following manner: i ) 3D structures were taken from the PDB ( PDB accession numbers and relevant chains are listed in Table S5 ) and optimized using the FoldX repairPDB function , ii ) Structures corresponding to each of the single-point mutants were generated using the FoldX BuildModel protein mutagenesis function , iii ) The interaction energy of the WT structure and the mutated structure with the Ag were calculated using the ComplexAnalysis energy calculation function of FoldX , iv ) . ΔΔG values were obtained using the following equation: ( 3 ) Mutations for which ΔΔG>0 . 25 were defined as destabilizing mutations , whereas mutations for which ΔΔG<−0 . 25 were defined as stabilizing . Mutations were defined as neutral if 0 . 25≥ΔΔG≥−0 . 25 . | Antibodies are a primary adaptive defence mechanism against infection , and function by recognizing and binding to non-self antigens . While most of the sequence of all antibodies of a given individual is identical , relatively small variations turn each antibody into a specific binder of one antigen . It is widely assumed that antigen binding sites correspond to the so called Complementarity Determining Regions ( CDRs ) of the antibody , which are defined as the elements that are most different between antibodies . We analysed all known antibody-antigen complexes and found that about 20% of the residues that actually bind the antigen fall outside the CDRs . However , we also found that virtually all antigen binding residues fall within regions of structural consensus between antibodies . Moreover , we demonstrate that antigen binding residues that reside within these structural consensus regions but outside of the traditionally-defined CDRs make significant energetic contribution to antigen binding . Furthermore , we show that these regions are organized along the sequence of the antibody chains and are identifiable from the sequence of the antibody . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"immunology",
"biology",
"computational",
"biology"
]
| 2012 | Structural Consensus among Antibodies Defines the Antigen Binding Site |
Bacteria form dense surface-associated communities known as biofilms that are central to their persistence and how they affect us . Biofilm formation is commonly viewed as a cooperative enterprise , where strains and species work together for a common goal . Here we explore an alternative model: biofilm formation is a response to ecological competition . We co-cultured a diverse collection of natural isolates of the opportunistic pathogen Pseudomonas aeruginosa and studied the effect on biofilm formation . We show that strain mixing reliably increases biofilm formation compared to unmixed conditions . Importantly , strain mixing leads to strong competition: one strain dominates and largely excludes the other from the biofilm . Furthermore , we show that pyocins , narrow-spectrum antibiotics made by other P . aeruginosa strains , can stimulate biofilm formation by increasing the attachment of cells . Side-by-side comparisons using microfluidic assays suggest that the increase in biofilm occurs due to a general response to cellular damage: a comparable biofilm response occurs for pyocins that disrupt membranes as for commercial antibiotics that damage DNA , inhibit protein synthesis or transcription . Our data show that bacteria increase biofilm formation in response to ecological competition that is detected by antibiotic stress . This is inconsistent with the idea that sub-lethal concentrations of antibiotics are cooperative signals that coordinate microbial communities , as is often concluded . Instead , our work is consistent with competition sensing where low-levels of antibiotics are used to detect and respond to the competing genotypes that produce them .
While the traditional model for bacterial life was one of cells swimming in liquid , it is now realized that bacteria commonly live in surface-associated communities known as biofilms [1–4] . These groups of bacteria carry significant medical and economic importance that includes a role in chronic diseases , antibiotic tolerance , biofouling , and waste-water treatment [5–7] . The importance of biofilms has led to their intensive study in many species of bacteria . However , this work has revealed that the genetic and biochemical mechanisms underlying biofilm formation are extremely variable across strains and growth conditions [8 , 9] . This variability has made it difficult to identify common principles of biofilm formation across species . One property common to the biofilms of many bacterial species , both gram-negative and gram-positive , is that they are induced by sub-lethal concentrations of antibiotics [10–17] . The diversity of the species displaying this response is striking [13] and suggestive of the general principles that so often elude the study of biofilms . This , and the observation that sub-lethal antibiotics induce a range of other physiological changes [18] , has led to the hypothesis that antibiotics may not function as killing agents in nature . Instead , it is suggested that antibiotics may function as cooperative signals between species that contribute to the homeostasis of bacterial communities [13 , 16 , 19–21] . In parallel , there are a growing number of studies concluding that natural microbial communities are cooperative enterprises in which strains and species work together , both in biofilm formation and metabolism [22–31] . Cooperative phenotypes—such as secreted enzymes or polymers—are important in biofilms for cells of a single genotype [4 , 32–35] . However , evolutionary theory and ecological experiments caution against the idea that this cooperation will always extend to cells of different genotypes [32 , 34 , 36–38] . Competition between strains and species appears to be commonplace and it is not clear how cooperative signaling via antibiotics could widely evolve [39] . An alternative to the idea that sub-lethal antibiotics are cooperative signals between species is that they are cues used by competing species to detect competitors and respond appropriately [39–41] . Under the cue model , antibiotics are not secreted in order to signal to others [42 , 43] , they are secreted to attack but can inadvertently provide information to others at low concentration . The idea that bacteria use antibiotics and other factors as cues of competition is central to the idea of “competition sensing” [41] . This is based on the observation that both nutrient and antibiotic stress , which are associated with ecological competition , often cause bacteria to release their own antibiotics . Bacteria appear to attack back when they are themselves attacked . If sub-lethal concentrations of antibiotics are cues of competition from other strains , and biofilms are induced by such concentrations , this implies that biofilm formation can be a response to ecological competition from other strains . Here we test this hypothesis by co-culturing natural isolates of the opportunistic pathogen Pseudomonas aeruginosa and studying the effects on biofilm formation . We focus on members of a single species here because these are expected to have the strongest ecological overlap and , therefore , the strongest ecological competition; something first noted by Darwin: “The struggle almost invariably will be most severe between the individuals of the same species , for they frequent the same districts , require the same food , and are exposed to the same dangers” [44] . More concretely , data from the cystic fibrosis lung suggest that different isolates can colonize and , at least temporarily , co-exist in young patients [45–47] , which is consistent with different strains meeting . P . aeruginosa also forms robust biofilms that have been intensively studied [6 , 48] . The co-culture of P . aeruginosa strains then provides a test case to evaluate the link between ecological competition and biofilm formation . Specifically , if biofilm formation is a response to ecological competition , we expect to observe ( i ) evidence of competition between strains and ( ii ) evidence of increased biofilm formation in co-cultures relative to monocultures .
We begin by recapitulating the known effects of antibiotic stress on bacterial survival and biofilm formation . Fig 1 ( panels A , B ) shows the effects of antibiotics on Pseudomonas aeruginosa where we use optical density as a qualitative indicator of cell number . As expected , increasing the concentration of three widely used clinical antibiotics—ciprofloxacin , rifampicin and tetracycline—monotonically reduces the optical density of shaking cultures . However , as previously shown , this monotonic decrease is not seen when the same concentrations are applied to standing cultures . In standing cultures , the liquid is not agitated and cells can readily attach to the edge of the well and establish biofilms . Under this condition , biofilm formation actually increases for many concentrations of the antibiotics , until the concentrations become so high that toxicity dominates [13] . Applying sub-lethal concentrations of antibiotics then leads to an increase in biofilm formation . Can we understand these widely reported biofilm responses to antibiotics in terms of ecological competition ? We tested this by mixing natural isolates of P . aeruginosa in standing cultures and studying the impact on biofilm formation . Consistent with the hypothesis that biofilm formation is a response to ecological competition , we find that biofilm formation is increased in mixed cultures compared to pure cultures , both in a broad scale survey that mixed combinations of 22 natural isolates ( Fig 1C ) and in defined pairwise comparisons in which we show that the strength of the biofilm response depends on the relative proportion of each strain at the start of the co-cultures ( Fig 1D ) . The observed biofilm response is consistent with a competitive response . However , an alternative hypothesis is that the strains cooperate with one another to produce more biofilm ( see Introduction ) . In order to better assess whether the strains were cooperating to make more biofilm , we labeled five of our P . aeruginosa natural isolates with constitutive fluorescent proteins ( cyan fluorescent protein , CFP , and yellow fluorescent protein , YFP ) by putting a single copy of the desired fluorescent protein gene at a specific site in the genome [50] . We then set up a round-robin tournament , in which the five strains were mixed in all combinations , and measured biofilm formation , planktonic growth , and the relative frequency of the two strains ( Fig 2 ) . This revealed a strong competitive effect with one strain typically dominating after 24 h ( median proportion of dominant strain in biofilm = 0 . 98 , Wilcoxon-signed rank test versus 0 . 5 , n = 10 , p < 0 . 01 ) . Thus , the biofilm response is a manifestation of competition between strains rather than a cooperative synergy in which both strains benefit from the presence of one another [34] . There are some cases in which biofilm formation in the mixture is less than that expected from the average of the two starting strains ( Fig 2 ) . However , accounting for which strain actually makes the final biofilm reveals that mixing nearly always causes an increase in biofilm formation by the dominant strain . For example , mixtures of 7 and 19 result in less biofilm than monocultures of strain 19 . However , strain 7 eliminates strain 19 in co-cultures , producing more biofilm than it does when alone ( Fig 2 ) . More generally , we can calculate the expected biofilm formation from the pure culture data using a weighted average , which weights by the frequencies of the two strains in the biofilm after competition . This reveals that the mixed culture biofilm normally increases relative to the expectation based upon pure cultures ( 9/10 mixes increase , Wilcoxon signed rank test of expected/observed biofilm against a null of 1 , n = 10 , p < 0 . 01 ) , while the amount of cells in the planktonic phase decreases ( n = 10 , p < 0 . 01 ) . Moreover , the increases in biofilm and decreases in plankton that we observe upon mixing strains are significantly negatively correlated across the different strain mixes ( Spearman rank correlation , n = 10 , r = -0 . 782 , p < 0 . 01 ) . This correlation agrees with recent observations on multispecies interactions during biofilm formation [31] . The biofilm response is not an idiosyncrasy of our particular growth medium . S1 Fig shows the effect of varying levels of nutrients , buffering , and soluble iron for two genotype mixtures . Despite the success of strains in pairwise competition being strongly dependent on the environment , the biofilm response is detected across all conditions in at least one of the two strain combinations , but typically in both ( S1 Fig ) . We next focus on the cue that drives the increase in biofilm formation in co-cultures . We first examined whether the biofilm response requires cell–cell contact by growing pairs of strains on either side of a 0 . 4 μm permeable membrane ( Fig 3A ) . Increased biofilm is observed across the membrane in most cases . In one case , there is a strong inhibitory effect on both biofilm and planktonic growth , which is consistent with the responder strain losing in co-culture ( specifically , strain 4 is inhibited and outcompeted by strain 1 , Fig 2 ) . We also found that the biofilm response is recapitulated by the addition of cell-free culture supernatant of different strains . We focus on strain 4 , which responds robustly to the addition of the supernatant from other strains ( Fig 3B ) . The nature of the response is dose dependent: lower concentrations tend to induce biofilm formation while high concentrations inhibit both planktonic growth and biofilm formation . In order to identify the active component within the supernatant , we first performed anion exchange chromatography on the supernatant of strain 69 as it generates a robust response in strain 4 ( Fig 3B ) . The active fraction that induced biofilm formation in strain 4 was further separated based on size by gel filtration . A single fraction ( C9 ) appeared to contain the bulk of the active component as it strongly inhibited growth in strain 4 , while the two flanking fractions C8 and C10 ( i . e . , lower concentrations of the effector ) induced the biofilm response ( Fig 4A ) . We subjected fraction C9 to protein mass spectrometry ( see Methods ) . Of the proteins identified , three are notable for their known role in microbial interactions: homologs ( PA14_8070 , PA14_8090 , PA14_8210 ) of the R and F pyocins in P . aeruginosa strain PA14 ( S1 Table ) [51] . Given that PA14 is a well-characterized strain , we next sought to explore how the pyocins of PA14 induce the biofilm response in our natural isolates . We first evaluated the effects of cell-free culture supernatant from PA14 on biofilm formation in the five strains from the round robin tournament . Strain 19 , which was isolated from a cystic fibrosis infection , is a very poor competitor ( Fig 2 ) and showed no detectable growth in our supernatant assays . It was therefore excluded from further analysis . Among the remaining strains , increased biofilm was observed in only strain 4 ( Fig 4B ) , and we confirmed that direct mixtures of PA14 and strain 4 also lead to increased biofilm formation ( S2 Fig ) . Why is the biofilm response absent in the other three strains ? To address this question , we grew the four strains in shaking culture in the presence or absence of the supernatant from PA14 and followed their growth over time . The results obtained show that the strains that do not show the biofilm response to the supernatant of PA14 also are not harmed by it ( Fig 4C ) . This correlation between toxicity and increased biofilm helps explain the lack of response in the three strains . Cell damage appears to be required for the biofilm response . We next studied the effect of various mutants of pyocin production in PA14 on their ability to damage and induce the biofilm response in strain 4 . We focused particularly on a strain of PA14 that has the repressor of pyocins , prtR , mutated to a version that is constitutively active ( S3 Table ) [52] . We added cell-free supernatant of the PA14 prtR mutant to strain 4 and compared the effect to the addition of WT PA14 supernatant by propidium iodide staining . Propidium iodide preferentially enters cells with envelope damage ( see Methods ) , and this assay confirmed that the pyocins produced by PA14 cause cell damage to strain 4 ( S3 Fig ) , which agrees with published literature on pyocin-driven lethality [51] . Moreover , additional experiments showed that the pyocins of PA14 are the major effectors of both population growth inhibition ( Fig 4C ) and the biofilm response of strain 4 ( Fig 4D ) . Focusing on the biofilm response , the use of mutants that lack only one class of pyocin suggested that F pyocins have some effect on biofilm formation at high concentrations , but it is particularly the R pyocins that strongly stimulate biofilm formation . This is , furthermore , consistent with the importance of cell damage because R pyocins are known to be efficient killers that can contribute significantly to competition among P . aeruginosa strains [53 , 54] . To further explore the potential of cell damage being required for the biofilm response , we screened 12 natural isolates for inhibition by the supernatant of PA14 in shaking culture and found that three strains were inhibited like strain 4 . We randomly selected two of the inhibited strains for further study in the biofilm assay . This revealed that both exhibit a biofilm response comparable to that of strain 4 and , again , the response occurs in a pyocin-dependent manner ( S3 Fig ) . We also tested the response of the strain PAK under the same conditions . This serves as a good control case since PAK is known to be inhibited by the pyocins produced by PA14 [53] . Again , we find evidence of pyocin-driven biofilm formation ( S4 Fig ) . These data further support the idea that pyocin-driven damage under these conditions is sufficient to drive the biofilm response . We note , however , that this does not exclude the possibility that there are additional mechanisms by which P . aeruginosa strains damage and induce biofilm in other susceptible strains . Indeed , the link between pyocin-based cell damage and the biofilm response is consistent with experiments using clinical antibiotics , in which antibiotic resistance mutants require higher concentrations of antibiotics to elicit the biofilm response [10] . To confirm this effect in our assays , we tested a rifampicin resistant mutant of PAO1 , with defined mutations in rpoB [55] and found that it does not respond to the same concentrations of rifampicin that trigger the biofilm responses in the wild type ( S5 Fig ) . In sum , we see comparable effects on biofilm formation from diverse forms of cell damage . This suggests that cell damage from a diverse range of bacteria interactions , both between strains and between species , has the potential to promote biofilm formation . Our experiments so far have utilized the crystal violet microtiter dish assay of standing cultures [48] . Although this experimental approach is widely used to quantify biofilm responses to antibiotics [13] , it comes with two notable limitations . Firstly , indirect proxies are used to estimate biofilm formation ( crystal violet staining ) and planktonic cell numbers ( optical density ) . Secondly , the presence of a large planktonic population in the standing culture means that the planktonic cells can affect the biofilm , and vice versa , in an undefined manner throughout the experiment . We therefore sought to recapitulate our results in a more defined assay using flow cells and direct imaging of the biofilms ( Fig 5A ) . We begin by studying one of our focal combinations , strain 1 ( PAO1 ) and strain 4 . As we find in the standing culture assay ( Figs 1D and 2B ) , PAO1 dominates strain 4 and largely excludes it from the biofilm after 24 h ( Fig 5B ) . Moreover , we observe an increase in the volume of biofilm formed by PAO1 . Our observations in the less-defined crystal violet assays , therefore , are recapitulated in flow cell biofilms: strain-mixing results in a single strain dominating and an increase in biofilm formation . Do we also observe this biofilm response in strain 4 , which was the focus of detailed analyses in the previous section ? In order to study this , we needed to create a condition where strain 4 is not eliminated from the biofilm . Preliminary tests with our two focal mixtures ( 1 + 4 , 4 + 69 , Fig 1D ) revealed that strain 4 will persist with strain 69 , if it is over-represented at inoculation . We , therefore , tested strain 4 mixed with strain 69 in the flow cells at a 3:1 initial ratio . Under these conditions , strain 4 dominates the biofilms and largely excludes strain 69 and , most importantly , we see a clear increase in biofilm formation in strain 4 here when compared to its single-strain counterpart ( Fig 5C ) . Experiments using flow cells have a separate attachment phase at the beginning of the experiment to seed the surface with bacteria ( see Methods ) . Counting attached cells after the attachment phase revealed that strain 4 attaches more in the presence of strain 69 but strain 1 does not attach more in the presence of strain 4 ( Fig 5 ) . Both initial attachment and subsequent biofilm accumulation , then , contribute to the response to strain mixing but while we always see increased biofilm formation , we do not always see an initial difference in attachment ( also see next section ) . In sum , the flow cell experiments provide validation from a second assay that P . aeruginosa strains increase biofilm formation in response to the presence of competing strains . Our data suggest that bacteriocins produced by natural isolates of P . aeruginosa act as a cue that drives biofilm formation . This raises the possibility that the widely reported biofilm responses to clinical antibiotics [10–17] also have their origins in responses to ecological competition from strains and species . But does the addition of antibiotics , or cell-free culture supernatant from competitors , also result in increased biofilm formation in a microfluidic device ? Experiments with a single concentration of antibiotic revealed that the addition of cell-free culture supernatant or clinical antibiotics commonly results in cell detachment at the initial stage of the experiment . This is consistent with the effects of supernatants and antibiotics in the crystal violet assay being strongly concentration dependent , whereby lower concentrations cause the biofilm response and higher concentrations cause growth inhibition ( Fig 3 ) . We , therefore , developed a more robust method to study the effects of cell-free supernatant and antibiotics on biofilms in microfluidic devices . We moved to use microfluidic platforms with two inlet channels in which only one of the two inlets contains the antibiotics ( Fig 6A ) . The goal of this experimental approach was to create a gradient inside the main chamber with a wide range of antibiotic concentrations , from sub-lethal to lethal ( see Methods ) . Our protocol always allows some parts of the chamber to contain negligible amounts of the antibiotic , which ensures that not all cells detach when antibiotics are flowed through . Importantly , the cells will experience a wide range of concentrations of antibiotic in a single experiment , and we can directly observe how they respond . Under these conditions , we observe a large increase in biofilm formation upon the addition of multiple classes of antibiotics in multiple strain backgrounds ( Figs 6 , S6 and S7 ) . Our data show that antibiotics do indeed promote biofilm formation in flow cells . In addition , they support the value of studying gradients for understanding cellular responses to stress . Such gradients are common in biofilms [56 , 57] , but their effects can be missed by static culture assays . We next asked whether pyocins are also capable of promoting biofilm formation in our dual inlet experiment . In support of this , we find that flowing the cell-free culture supernatant of PA14 over strain 4 results in increased biofilm formation , but not when we use the supernatant of the pyocin null-mutant ( Fig 6 ) . What drives the increase in biofilm ? Increased biofilm accumulation could occur through stronger attachment and better retention of cells throughout the experiment , or through the acceleration of cell division . In order to test the impact of pyocins on attachment , we allowed strain 4 cells to attach in the presence of cell free supernatant of PA14 and compared this to the effect of cell free supernatant of the PA14 pyocin null-mutant ( S8 Fig ) . This shows that pyocins do indeed lead to increased attachment . This phenotype is not simply sedimentation of damaged cells . We flush unattached cells out before imaging and , importantly , we see increased attachment in the presence of pyocins both on the bottom and the top of the channel ( S9 Fig ) . Moreover , propidium iodide staining shows that , concurrently , there is increased cell damage from the supernatant of PA14 than with the supernatant of the pyocin null mutant ( S8 Fig ) , consistent with what we found for shaking culture ( S3 Fig ) . We see , therefore , that pyocins from competitors both cause cell damage ( S3 and S8 Figs ) and an increase in biofilm formation ( Figs 6 , S4 , S8 and S9 ) . The attachment experiments ( S8 and S9 Figs ) suggest that the biofilm response may occur via an increased propensity for cells to attach , and subsequently persist , in the biofilm . An increased rate of cell division could also contribute to the biofilm response in some or all strains . Direct assessments of cell division in three-dimensional biofilms is challenging and was not possible with our current methods . However , there is evidence from Escherichia coli that sub-lethal levels of antibiotics can increase growth rate [58] . More generally , there is evidence for widespread transcriptional and metabolic changes in response to antibiotics [16 , 59–66] , which have recently been the focus of intense debate [62 , 67–70] but still lack evolutionary explanation . If these changes are also involved in the responses to other bacterial strains and species , it would suggest that many of the physiological and metabolic changes seen upon antibiotic use have their natural origins in competition sensing [41] and the ecology of the species concerned .
Biofilms are central to microbiology . Despite this , we still understand relatively little about the drivers of biofilm formation in a natural context . Our data suggest that a major factor in biofilm formation in nature is ecological competition with other strains . We have shown that co-culturing strains of P . aeruginosa often leads to increased biofilm ( Figs 1 , 2 and 5 ) . We have also shown that this increase in P . aeruginosa should not be interpreted as a cooperative interaction between strains . Mixing pairs of strains shows that one of the two strains is largely eliminated during the process of biofilm formation ( Figs 2 and 5 ) . Moreover , our data suggest that biofilm formation is stimulated as a response to cell damage caused by other strains: cell-free supernatant of one strain only induces biofilms in another strain when it can also causes cell damage ( Figs 4 , S3 and S8 ) . In pairwise mixtures of natural isolates , we have shown that one driver of the biofilm response are R and F pyocins , the narrow-spectrum antibiotics produced by P . aeruginosa . We , therefore , expect that pyocin-mediated competition will often be an inducer of biofilm formation in P . aeruginosa . However , our work also emphasizes that the biofilm responses to clinical antibiotics are highly comparable to those seen with pyocins . The different sets of antibiotics studied here have very different modes of action , including inhibition of DNA replication ( ciprofloxacin ) , transcription ( rifampicin ) , or translation ( tetracycline ) and , in the case of R and F pyocins , disruption of membrane potentials . The responses we observe , therefore , are very general and suggest that any toxin made by one P . aeruginosa strain has the potential to induce biofilm formation in another . Is there a common mechanism linking the diverse forms of cell damage to the biofilm response ? Bacteria have a number of stress responses that are triggered by cell damage , which have been proposed as candidates for sensing the presence of competing strains [41] . Perhaps most notably , this includes the responses to oxidative stress , which is associated with the action of clinical antibiotics and diverse damaging agents from competing microbes [59 , 62 , 68 , 71 , 72] . What is the evolutionary basis of P . aeruginosa strains increasing biofilm formation in response to ecological competition ? Mechanistically , our experiments suggest that the biofilm response is caused , at least in part , by an increase in surface-attachment in the presence of other strains . Whether a biofilm response is caused by differential attachment or another mechanism , the key effect is to increase the frequency of the responder relative to competing genotypes . We hypothesize that the evolutionary advantage of this increase in frequency is that it allows strains to better defend against and attack other genotypes . Biofilms are known to promote resistance to a wide range of stressors , including R pyocins [73] . In addition , competitive ability in a biofilm is expected to be linked to the relative density of each strain . At high density , cells can better control the environment with secreted products that favor their genotype over others [4 , 74] . Dominating a biofilm early , then , is likely to enable a genotype to dominate in the long term . In this model , the well-known biofilm responses to antibiotics [10–17] have their roots in natural selection to detect competing genotypes and respond in a manner that maximized the ability to defend and attack [39 , 41] . The fact that antibiotics are known to induce biofilm formation in diverse species suggests that ecological competition may increase biofilm formation in many microbial communities in which many strains and species meet . Consistent with this prediction , recent work on assemblage of multiple species , including known antibiotic producers , found that strain mixing increases biofilm formation [26 , 30 , 31] , although it was interpreted there as cooperation between species . Understanding the effects of ecological competition on biofilm formation also has implications for health and disease . An important new idea in the treatment of disease is that we can harness ecological competition between microbes to our advantage [75 , 76] . Probiotic supplements for the gut are already an important industry [77] , and a major goal of synthetic biology is to arm benign bacteria with the antibiotics they need to combat disease-causing strains [78] . We must be cautious , then , that this escalation of ecological competition does not result in the outcomes that we observe here: a more abundant pathogen .
A total of 22 strains of Pseudomonas aeruginosa were used , including both clinical and environmental isolates from the Kolter strain collection ( S2 Table ) . A subset of these strains was subject to detailed analyses , including the well-studied clinical strain PAO1 ( here referred to as strain 1 ) , and strains 4 , 7 , 19 , and 69 . These five strains were chromosomally tagged with constitutively-expressed fluorescent tags ( cfp or yfp ) using the mini-Tn7 system [50] . The pyocin-null prtRS153A ( constructed as in [52] ) , producing constitutively active repressor , and its parental strain PA14 were kindly provided by Jon Penterman . The R and F pyocin mutants are from the PA14 Non-Redundant Transposon Insertion Mutant Set [79] . We found no differences in our assays between the two sources of PA14 strains and only show the data from the Penterman source in the figures . Other strains are described in S3 Table . These assays were based on standard protocols [48] . Briefly , cells from overnight cultures of P . aeruginosa were sub-cultured and from these , exponential phase cells were inoculated in peg-lidded 96-well plates ( Nunc , conical bottom ) in 100 μl tryptone broth ( 10 g Bacto triptone per 1 l water ) at starting optical density of 0 . 0025 ( A600 ) , and allowed to grow at 22°C . After 24 h , 50 μl of medium was removed and the optical density at A600 was measured to estimate the density of planktonic cells . Pegs were then washed , stained with 0 . 1% crystal violet , washed again , and then the remaining stain was dissolved with 200 μl of 33% acetic acid . Finally , the optical density at A595 was measured and used as a proxy for biofilm formation . For direct strain-mixing experiments ( Figs 1C , 1D and 2 ) , we used plates without pegs because this allows cells to be more easily removed and sampled from the biofilms when required ( see below ) . The only substantive difference is that the biofilm on the edge of the well is assessed using crystal violet rather than the biofilm on a peg . We confirmed that the two assays give very similar results ( S10 Fig ) . While the experiments presented in the main text were done in tryptone broth , the biofilm response was also investigated in other standard bacterial growth media , including Luria broth ( LB ) , Pseudomonas minimal medium ( PMM ) and casamino acid ( CAA ) medium , along with different concentrations of iron or buffer , which did not have qualitative effects in the cellular responses to mixing ( S1 Fig ) . Three antibiotics belonging to different structural families , namely ciprofloxacin , rifampicin , and tetracycline , were chosen to recapitulate the effect of sub-inhibitory concentrations of antibiotics on biofilm formation of P . aeruginosa [13] . Stock solutions of each antibiotic ( obtained from Sigma Chemical Company ) were stored at -20°C and diluted to the desired concentration on the day of each experiment . All antibiotics experiments used tryptone broth as the growth medium . Biofilm formation and growth in shaking culture were assayed as detailed in the sections “Static Biofilm Assays” ( above ) and “Growth Curves” ( below ) , respectively . For the dual-flow microfluidic experiments ( Fig 6 ) we used 10x the concentration that inhibited growth in shaking culture , as shown in Fig 1 . These are 1 μg/mL for ciprofloxacin , 50 μg/mL for rifampicin , and 1 , 000 μg/mL for tetracycline . The initial survey of how strain mixing affects biofilm formation was based on 22 strains of P . aeruginosa . Biofilm formation after 24 h was evaluated in 22 single strains , 22 pair-wise combinations , eight combinations of five strains , and two of ten strains . Strains were evenly mixed based upon the optical density of stationary phase cultures . The strain combinations in the mixtures were chosen to ensure a near-even representation of the strains across the mixtures . Each strain occurred twice in the pairs , twice in the mixtures of fives ( with one strain appearing only once ) , and once in a mixture of ten ( with one strain excluded ) . Sixteen replicates were performed in each experiment for each strain or mixture , to give a total of 432 biofilm assays in each experiment . All treatments were performed on two different days . The second set of experiments mixed five fluorescently-labeled strains in all combinations . We assessed biofilm formation and planktonic cell density in 96 well plates after 24 h . Samples of planktonic cells and biofilm cells were taken from each treatment to record the frequency of the two strains using fluorescence microscopy . The biofilm sample was taken by vigorously shaking the washed plates in phosphate buffer for one hour ( experiments revealed that this treatment gives identical frequency estimates as shaking with 500 μm glass beads and removing the whole biofilm ) . In addition , direct images of the biofilms were taken by growing strains in 2 ml of media containing 25 x 25 mm2 glass coverslips in 6-well plates under conditions otherwise identical to the 96-well plate assay . The bacteria form a biofilm on the coverslip slightly below the air-water interface , which is then visualized by fluorescent microscopy ( Fig 2 ) . All experiments were repeated twice on different days . Comparison of these data with the pair-wise mixtures in the first experiments ( above ) confirmed that the mixing responses were not affected by fluorescent labeling . All methods were the same as before with the addition that we diluted overnight cultures of each strain and grew them for 2 h to obtain exponential phase cells to inoculate the biofilm assays ( Fig 2 ) . This did not affect biofilm formation , but it helped reduce variability in the strain frequency in the mixtures . For the two large-scale strain-mixing experiments ( Figs 1C and 2 ) , we present data that is the combination of the two repeats from different days as variability between days was relatively high . Biofilm formation was also monitored under laminar flow conditions using the BioFlux 200 system ( Fluxion Biosciences , South San Francisco , CA , United States ) . Experiments were performed essentially as described elsewhere [80] . First , microfluidic channels ( 350 μm x 75 μm x 4 . 89 mm for width , height , and length , respectively ) were primed with growth media , here tryptone broth , to prevent air bubbles from entering the channels on subsequent loading steps , and just after , inoculated with cells at the exponential phase . The concentration of these cultures was adjusted so that the optical density at 600 nm in tryptone broth was 0 . 25 . Then , after the attachment period without flow ( 20 min at 22°C ) , fresh medium was pumped from inlet wells through the channels to outlet , firstly with a flow rate of 40 μl/h for 20 min to flush out planktonic cells , and then with a constant flow rate of 4 μl/h until the end of the experiments ( 24 or 48 h ) . To address the effect of pyocins on cellular attachment we allowed cells to attach in the channels for 20 min with tryptone broth made with 10% cell-free supernatant , either from strain PA14 WT or a mutant that has the repressor of pyocins , prtR , mutated to a version that is constitutively active ( S3 Table ) . We then flushed out planktonic cells , as explained above , and imaged to assess attachment . For consistency , we used only one plate type for all microfluidic experiments , the “Invasion” plate of Fluxion Biosciences that has two parallel inlets and one outlet ( Fig 6 ) . When studying co-cultures , the two inlets contained the same medium , and when studying the effect of antibiotic gradients we added the required concentration of the antibiotic ( or cell-free culture supernatant ) to one of the inlets only , thus creating a gradient of the compound being tested in the main chamber . Each experiment was repeated twice , on different days , and the images of biofilms taken come from the middle of the channels , where the channels are the flattest , when accessing the effects of strain mixing . In all treatments , images were taken at the same distance from the respective inlets . When accessing the effect of antibiotics or cell-free supernatant we imaged directly downstream of the inlets , as shown in Fig 6 , where the gradient is the steepest . To access the effects of diffusible substances in the biofilm response , two sets of experiments were performed . Firstly , we used 96-well Transwell plates ( Corning ) with two different inlays separated by a 0 . 4 μm pore size membrane . In these experiments , the strain in the bottom of the well was inoculated at an optical density of 0 . 0025 ( A600 ) in 75 μl of media . This bottom strain was the tester strain in which biofilm formation was analyzed . The top strain was the inducer and was inoculated at a higher density of 0 . 1 ( A600 ) with 50 μl media . These revealed cell-to-cell contact is not strictly necessary for the biofilm response ( Fig 3A ) . To investigate this further , we studied responses of strains to cell-free supernatants from other strains . The supernatant of the strains were obtained as follows . Exponential phase cells of all strains were inoculated in 6-well plates in 4 ml of tryptone broth at starting optical density of 0 . 025 ( A600 ) , and allowed to grow at 22°C for 21 h . Cultures were then spun down ( 3 , 000 g during 10 min ) and filtrated using 0 . 22 μm filters ( Millipore ) . The spent media assay followed the same methods for the biofilm assay , except that exponential phase cells of the responder strain were inoculated in media containing ( 2 . 5/25% ) cell-free culture supernatant from itself or others . Fractionation of the cell free supernatant was performed by GE 1 ml resource Q anion exchange chromatography ( buffer A: 50 mM Tris-HCl [pH 7 . 5] , 2% glycerol , 2 mM β-mercaptoethanol; buffer B: buffer A + 1 M NaCl ) followed by GE Superdex 200 10/300 GL gel filtration in SEC buffer . The active fraction , as determined by the crystal violet assay ( Fig 4 ) , was subject to mass spectrometry by the FAS Center for Systems Biology Core Facilities , Harvard University , US . Briefly , the sample was reduced , carboxyamidomethylated and digested with trypsin . Peptide sequence analysis of each digestion mixture was then performed by microcapillary reversed phase high-performance liquid chromatography coupled with nanoelectrospray tandem mass spectrometry ( μLC-MS/MS ) on an LTQ-Orbitrap mass spectrometer ( ThermoFisher Scientific , San Jose , CA , US ) . Exponential phase cells from overnight cultures were inoculated in flat-bottomed 96-well plates ( Nunc ) in 200 μl of tryptone broth at the starting optical density ( A600 ) of 0 . 0025 , as for the biofilm assays , and allowed to grow at 22°C , shaking , inside a plate reader ( Tecan , Infinite M200 Pro ) . Test treatments were supplemented with 25% of cell-free supernatant from PA14 , the respective pyocin-null , or distilled water ( control ) . The optical density at A600 was measured automatically every 500 s . When required , cell damage was assessed by staining shaking cultures or attached cells in a microfluidic device with propidium iodide ( at the final concentration of 3 μl/ml ) [81] . Propidium iodide-stained cells were visualized using epifluorescence microscopy with DsRed-based filters . Images of BioFlux biofilms were obtained using a Zeiss confocal laser scanning microscope ( LSM 700 ) , a digital camera ( AxioCam MRm ) , and the associated Zen 2011 software ( blue or black edition , for fluorescence or confocal microscopy , respectively ) . Objectives used were 10x , 20x , or 40x , as specified in the figure legends . A MATLAB ( MathWorks ) script was used to determine the number of attached cells on the surface of microfluidics , along with the volume of biofilms after 24 h and 48 h of incubation . The volume calculation uses the area of fluorescence signal of biofilms in each confocal slice , which is then summed across the slices in z-plane . | Bacteria often attach to each other and to surfaces and make biofilms . These dense communities occur everywhere , including on us and inside us , where they are central to both health and disease . Biofilm formation is often viewed as the coordinated action of multiple strains that work together in order to prosper and protect each other . In this study , we provide evidence for a very different view: biofilms are formed when bacterial strains compete with one another . We mixed together different strains of the widespread pathogen Pseudomonas aeruginosa and found that pairs often make bigger biofilms than either one alone . Rather than working together , however , we show that one strain normally kills the other off and that biofilm formation is actually a response to the damage of antibiotic warfare . Our work helps to explain the widespread observation that treating bacteria with clinical antibiotics can stimulate biofilm formation . When we treat bacteria , they respond as if the attack is coming from a foreign strain that must be outnumbered and outcompeted in a biofilm . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Conclusions",
"Methods"
]
| []
| 2015 | Biofilm Formation As a Response to Ecological Competition |
Mitochondrial membrane biogenesis and lipid metabolism require phospholipid transfer from the endoplasmic reticulum ( ER ) to mitochondria . Transfer is thought to occur at regions of close contact of these organelles and to be nonvesicular , but the mechanism is not known . Here we used a novel genetic screen in S . cerevisiae to identify mutants with defects in lipid exchange between the ER and mitochondria . We show that a strain missing multiple components of the conserved ER membrane protein complex ( EMC ) has decreased phosphatidylserine ( PS ) transfer from the ER to mitochondria . Mitochondria from this strain have significantly reduced levels of PS and its derivative phosphatidylethanolamine ( PE ) . Cells lacking EMC proteins and the ER–mitochondria tethering complex called ERMES ( the ER–mitochondria encounter structure ) are inviable , suggesting that the EMC also functions as a tether . These defects are corrected by expression of an engineered ER–mitochondrial tethering protein that artificially tethers the ER to mitochondria . EMC mutants have a significant reduction in the amount of ER tethered to mitochondria even though ERMES remained intact in these mutants , suggesting that the EMC performs an additional tethering function to ERMES . We find that all Emc proteins interact with the mitochondrial translocase of the outer membrane ( TOM ) complex protein Tom5 and this interaction is important for PS transfer and cell growth , suggesting that the EMC forms a tether by associating with the TOM complex . Together , our findings support that the EMC tethers ER to mitochondria , which is required for phospholipid synthesis and cell growth .
Mitochondria are critical cellular components that are needed for energy production , lipid metabolism , calcium regulation , and apoptosis . Most proteins and lipids necessary for mitochondrial biogenesis are not synthesized in mitochondria and must be imported . Although protein import into mitochondria is relatively well understood , much less is known about phospholipid transfer to mitochondria . Phospholipid synthesis occurs largely in the endoplasmic reticulum ( ER ) , and mitochondria acquire phospholipids from the ER at regions of close contact between these organelles [1]–[3] . Zones of close contact between organelles , often called membrane contact sites , are regions where lipids , small molecules , and other signals are transferred between organelles [4]–[6] . Contacts between the ER and mitochondria are not only important for lipid exchange and signaling between these organelles , but have also been proposed to play a role in calcium signaling , apoptosis , Alzheimer's disease pathology , and viral replication [7]–[10] . Protein complexes proposed to mediate ER–mitochondria contacts have been identified in mammalian cells and in Saccharomyces cerevisiae [11]–[14] . The only such complex that has been found in yeast to date is called the ER–mitochondria encounter structure ( ERMES ) , which comprises the integral ER glycoprotein , Mmm1 , the cytosolic protein , Mdm12 , and two proteins of the outer mitochondrial membrane , Mdm10 and Mdm34 [14] . The ERMES complex may play a role in phospholipid exchange between the ER and mitochondria . Notably , three of the four proteins contain a domain that may facilitate lipid exchange between closely apposed membranes [15] . The study that identified the ERMES complex found that phospholipid exchange between the ER and mitochondria decreases 2–5-fold in cells missing this complex [14] . However , two subsequent studies found that the transfer of phosphatidylserine ( PS ) from the ER , where it is produced [16] , to mitochondria , did not significantly slow in cells lacking the ERMES complex [17] , [18] . These findings suggest that protein complexes in addition to ERMES mediate ER–mitochondria tethering , as lipid exchange between these organelles probably occurs only at contacts . In addition , the fact that cells lacking ERMES are viable suggests there may be other tethering complexes , as ER–mitochondria tethering is probably essential [19] . The mechanism of phospholipid exchange between the ER and mitochondria at sites of contact between these organelles is not well understood , but is thought to be nonvesicular [2] , [3] . A small number of mutants that reduce ER to mitochondria lipid transfer have been identified . One is missing Met30 , a subunit of an ubiquitin ligase complex [20] that ubiquitinates the transcription factor Met4 , which in turn regulates PS transfer to mitochondria [21] . Cells missing both the ERMES complex and proteins needed to shape the ER have a similar ∼50% decrease in ER to mitochondria PS transfer [18] . However , why lipid exchange slows in these mutants is not understood . PS transfer to mitochondria is required for the synthesis of phosphatidylethanolamine ( PE ) in mitochondria [2] . PE is critical for mitochondrial function [22] , [23] . Although PE can be made outside mitochondria , for unknown reasons this PE is not efficiently transferred to mitochondria [24] . Failure to import PE may explain why all cells from yeast to humans have an enzyme that converts PS to PE in the mitochondrial matrix . In yeast , this protein is called PS decarboxylase 1 or Psd1 ( Figure 1A ) [25] , [26] . PE produced in mitochondria by Psd1 can be transferred back to the ER and converted to phosphatidylcholine ( PC ) by the methyltransferases , Cho2 and Opi3 ( Figure 1A ) . There is a second PS decarboxylase in yeast , called Psd2 , which may reside in the Golgi complex , endosomal system , or vacuole [27] . PE can also be synthesized from diacylglycerol ( DAG ) and CDP–ethanolamine , a metabolic pathway known as the Kennedy pathway ( Figure 1A ) [28] . The Kennedy pathway can also produce PC from DAG and CDP–choline ( Figure 1A ) . In this study , we employed an array-based genetic interaction screen to identify genes required for phospholipid exchange between the ER and mitochondria . We found that mutants missing multiple proteins of the ER-membrane protein complex ( EMC ) had defects in PS transfer to mitochondria , reduced ER–mitochondrial tethering , and impaired mitochondria function . This complex contains six conserved proteins , called Emc1–6 [29] . The EMC has been suggested to play roles in the cellular response to ER stress , in membrane protein folding , or the unfolded protein response ( UPR ) in the ER [29]–[33] . However , the molecular functions of the EMC are not known . Our findings indicate that the EMC mediates lipid transfer from the ER to mitochondria by facilitating tethering between these organelles .
In a prior synthetic genetic array ( SGA ) screen for the PSD1 gene , we uncovered an aggravating genetic interaction with the CHO2 gene . We found that the growth defect of the Δpsd1Δcho2 mutant was rescued by the addition of ethanolamine or choline to the medium ( Figure 1B ) , which allows cells to make PE and PC via the Kennedy pathway , indicating that this mutant had a defect in PE synthesis from PS ( Figure 1A ) . We reasoned that genes required for transfer of PS from ER to mitochondria and PE from mitochondria to ER would similarly have negative genetic interactions with CHO2 that would be rescued by ethanolamine or choline . Therefore , we performed SGA screens for the CHO2 gene in the absence and presence of choline to identify genes that functioned in lipid transfer between ER and mitochondria . The results of this genome-wide screen are shown in Figure 1C , in which we plotted the growth of double mutants in the absence of choline versus their growth in its presence . We identified 209 aggravating genetic interactions in total , 187 that we deemed were rescued by choline addition ( Figure 1C and Table S1 ) . We found significant enrichment for various functional classifications for this set of genes , and these are shown in Figure 1D ( see also Table S2 ) . Membrane-associated functions were highly represented among these groups and included ER , vacuole , and endosomal compartments . In the Cellular Component category , we noticed an almost 30-fold enrichment for subunits of the EMC , a conserved uncharacterized ER membrane protein complex ( Figure 1D ) [29] . In our screen , genes encoding the six subunits of the EMC showed strong aggravating genetic interactions with CHO2 that were rescued by choline addition ( Figure 2A ) . We verified these genetic interactions and their rescue by choline by spot assay ( Figure S1 ) . Interestingly , these interactions were not rescued by ethanolamine , suggesting that the EMC has functions distinct from PE production by Psd1 . To uncover additional functional information about the EMC , we examined the global genetic interaction network , which is a comprehensive map of pairwise genetic interactions in which genes with similar functions form coherent clusters [34] . We noticed that all EMC genes except for EMC5 , which was not present in the global network , formed a discrete cluster , suggesting EMC genes share similar functions ( Figure 2B ) . Interestingly , also in this cluster were CHO2 and OPI3 , which encode the two methyltransferases that convert PE to PC ( Figure 1A ) , supporting a role for EMC genes in phospholipid metabolism . EMC genes showed primarily aggravating genetic interactions with a cluster of genes with roles in lipid metabolism including ICE2 , INO2 , and INO4 . The latter two encode both subunits of the Ino2/4 transcriptional activator complex required for expression of genes involved in phospholipid synthesis , including OPI3 . Next , we performed an SGA screen for one of the EMC genes , EMC6 , to further define functions for the EMC complex . We identified 37 aggravating and 45 alleviating genetic interactions with EMC6 ( Table S3 ) . Genetic interactions with EMC6 revealed enrichment for functions associated with lipid metabolism , mitochondria , membrane traffic , cell polarity and morphogenesis , cell signaling , and chromatin ( Figure 2C ) . Although EMC genes have been found to have links to the UPR [29]–[31] , we did not find significant enrichment for ER stress response functions and EMC6 did not interact genetically with either HAC1 or IRE1 , two key factors required for induction of the UPR . Our screen did uncover genetic interactions with key regulators of phospholipid metabolism , INO2 , INO4 , and SCS2 , as well as CHO2 ( Figure 2C ) , further supporting a role for EMC proteins in phospholipid synthesis . We also did not observe aggravating genetic interactions between EMC6 and any of the remaining EMC genes ( Table S3 ) . EMC proteins 1–6 were first identified by their ability to interact in an affinity purification experiment [29] . We now examined their individual localizations by tagging the endogenous proteins with GFP and imaging by confocal microscopy . We found that each of Emc1–6 localized throughout the yeast ER and were expressed at similar levels ( Figure 3A ) . Next we verified interactions between each EMC protein by Protein-Fragment Complementation Assay ( PCA ) in which we tagged each endogenous Emc protein with one half of the Venus fluorescent protein . When two proteins interact , the two halves of Venus bind each other ( probably irreversibly ) and form a fluorescent protein . Shown in Figure 3B is the matrix containing all pair-wise interactions between the EMC proteins . We observed PCA interactions in the ER between most EMC proteins , and there was no single EMC protein that failed to interact with any other EMC protein , suggesting that all six EMC proteins did indeed form a complex within the ER . Interestingly , we did not detect interactions between Emc1 and Emc3 or Emc1 and Emc4 , whereas all other EMC proteins interacted , suggesting that Emc1 was organized distinctly within the complex . To determine if the EMC plays a role in phospholipid exchange between the ER and mitochondria , we used an in vivo assay to measure PS import into mitochondria from the ER . After synthesis in the ER , PS can be transferred to mitochondria and converted to PE by Psd1 [28] . Thus , the conversion of newly synthesized PS to PE has been used to estimate the amount of PS transfer from the ER to mitochondria [21] . We metabolically labeled cells with [3H]serine , which is used for PS production in the ER [16] . Previously , we have shown that using the labeling conditions described in Materials and Methods , cells produce PS and PE at linear rates and that little of the radiolabeled PE is converted to PC [35] . Strains were labeled with [3H]serine for 30 min and the ratio of [3H]PE to [3H]PS calculated . In a wild-type strain , this ratio was 2 . 5 . Cells missing either Psd1 or Psd2 had a significant decrease in the [3H]PE to [3H]PS ratio , and this ratio was close to zero in a strain missing both proteins ( Figure 4A ) . Using this assay we determined the amount of PS converted to PE in cells missing any one of the Emc proteins . We found that the amount of PS converted to PE did not decrease significantly in these strains ( Figure 4A ) . However , in cells missing multiple EMC proteins , PS to PE conversion was significantly slowed . A strain lacking Emc2 and Emc6 had a ∼25% decrease in the ratio of [3H]PE to [3H]PS compared to wild-type . Cells missing additional Emc proteins had more substantial transfer defects; a strain missing Emc1 , Emc2 , Emc3 , and Emc6 ( 4x-emc ) or one missing these proteins and Emc5 ( 5x-emc ) had ∼50% reduction in the ratio of [3H]PE to [3H]PS ( Figure 4A ) . These strains contain Psd2 , which is outside mitochondria and converts a significant fraction of newly synthesized PS to PE ( Figure 4A ) . Because the amount of PS to PE conversion in 4x-emc and 5x-emc cells was about the same as that of cells lacking Psd1 , our findings suggest that very little PS to PE conversion occurs in the mitochondria of these strains . To rule out that the decrease in PS to PE conversion in 5x-emc cells was caused by a reduction in Psd1 activity or mislocalization of Psd1 , we determined the amount of Psd activity in mitochondria derived from 5x-emc cells . An in vitro Psd assay was performed using a fluorescent PS analog [7-nitro-2–1 , 3-benzoxadiazol-4-yl]-PS ( NBD-PS ) . We found mitochondrial Psd activity was not reduced in mitochondria from 5x-emc cells compared to those from wild-type cells , but rather was significantly increased , for unknown reasons ( Figure 4C ) . In addition , we ruled out that the decrease in PS to PE conversion in 5x-emc cells was due to differences in total PS production during the labeling; there was no significant difference between total PS produced in wild-type and the EMC mutants ( Figure S2A ) . We also confirmed that the rates of PS and PE production were linear in wild-type [35] and 5x-emc cells over the course of the 30-min labeling ( Figure S2B ) . These findings confirm that 5x-emc cells have a significant decrease in the transfer of PS from the ER to mitochondria . To test whether most of the PS converted to PE in 5x-emc cells was due to Psd2 , we sought to make a 5x-emc psd2Δ strain and hence to measure PS to PE conversion only in the mitochondrial pathway . However , we found that 5x-emc psd2Δ cells were not viable . We transformed 5x-emc cells with a plasmid containing PSD2 and URA3 and then deleted PSD2 on the chromosome . The resulting strain was not able to grow on medium with 5-fluoroorotic acid ( 5-FOA ) , which is toxic to URA3 strains and selects against the plasmid carrying the URA3 and PSD2 genes ( Figure 4B ) , confirming that the 5x-emc psd2Δ mutant was not viable . Adding ethanolamine to the medium , which is used to make PE by the Kennedy pathway ( Figure 1A ) , and which restored viability of psd1Δ psd2Δ cells , did not rescue the 5x-emc psd2Δ mutant ( Figure 4B ) . This suggests that the 5x-emc mutant might have lipid metabolism defects in addition to significantly reduced PS transfer to mitochondria . For example , it is possible that PE synthesis by the Kennedy pathway or the conversion of PE to PC is defective in 5x-emc cells . However , we found that both processes occurred at similar rates in wild-type and 5x-emc cells ( Figure S3 ) . Therefore , it remains unclear why 5x-emc psd2Δ cells are not viable even when grown in media containing ethanolamine . Because 5x-emc cells have reduced PS transfer from ER to mitochondria , we suspected that mitochondria from 5x-emc cells would have reduced amounts of PS and PE . To measure phospholipid levels , wild-type and 5x-emc cells were labeled with [3H]acetate for at least 3–4 generations and the relative abundance of the major phospholipids in purified mitochondria was determined . We found that PS levels in the mitochondria from 5x-emc cells were about 50% lower than those in wild-type mitochondria ( Figure 5A ) . Therefore , reduced ER to mitochondria PS transfer in 5x-emc cells results in decreased steady-state PS levels in mitochondria . Notably , PE levels were also reduced about 50% . This reduction is probably caused by the defect in ER to mitochondria PS transfer in 5x-emc cells , as most PE in mitochondria is generated from PS by Psd1 [24] . Interestingly , the relative abundance of other phospholipids was increased in 5x-emc mitochondria , particularly phosphatidic acid ( PA ) and cardiolipin ( CL ) ( Figure 5A ) . It may be that the transfer of PA and other CL precursors into mitochondria is not as sensitive to the loss of the EMC as is PS transfer . Alternatively , the increase in PA and CL may reflect a mechanism by which cells compensate for low levels of mitochondrial PE , which is thought to be critical for proper mitochondrial function [22] . Because the lipid profile of mitochondria in 5x-emc cells was dramatically altered , we wondered if the mitochondria were functional . Yeast strains with nonfunctional mitochondria cannot grow on media containing nonfermentable carbon sources such as glycerol . We found that 5x-emc cells and other strains missing multiple EMC proteins did not grow on the glycerol-containing medium , YPGly , as has previously been found for cells lacking the ERMES component Mmm1 ( Figure 5B ) [36] . Therefore cells missing multiple EMC proteins do not have functional mitochondria , probably because of the abnormal levels of phospholipids in the mitochondria of these strains . Interestingly , 5x-emc cells also had a substantial growth defect even on glucose-containing media ( Figures 4B and 5B ) . Because we found that 5x-emc cells have reduced ER to mitochondria PS transfer in vivo , we wondered if a similar defect could be detected in vitro . We used a previously established two-step assay to monitor the transfer of newly synthesized PS from the ER to mitochondria [18] , [37] . In the first step , crude mitochondria are incubated for 20 min with [3H]serine and Mn2+ . We have shown that crude mitochondria have tightly associated ER-derived membranes that contain PS synthase [18] . The presence of Mn2+ is required by PS synthase but inhibits the conversion of [3H]PS to [3H]PE by Psd1 [37] . Thus , in the second step of the reaction , Mn2+ is chelated by EDTA and the PS to PE conversion rate was determined . Because Psd2 is not active in this assay [18] , all PS to PE conversions in this assay are mediated by Psd1 and indicate the rate of PS transfer from ER to mitochondria . Using mitochondria derived from wild-type cells , we found that about 1% of [3H]PS synthesized was converted to PE per minute ( Figure 6A ) . When mitochondria from 3x-emc ( missing Emc2 , Emc5 , and Emc6 ) and 5x-emc cells were used , this rate decreased about 2- and 3-fold , respectively ( Figure 6A ) . It should be noted that for all the strains tested , the rate of PS to PE conversion was linear ( R2≥0 . 9 ) . Because mitochondria derived from 5x-emc cells have Psd activity that is not lower than those from wild-type ( Figure 4C ) , these findings indicate that the rate of ER to mitochondria PS transfer is significantly reduced in crude mitochondria derived from 5x-emc cells . Because the ERMES complex is thought to function as a tether between the ER and mitochondria , we wondered whether PS transfer would be slower in crude mitochondria missing Emc proteins and the ERMES complex . As disruption of a single ERMES component causes disassembly of the entire complex [14] , we sought to delete one of the four genes encoding the ERMES proteins in 5x-emc cells . However , we were unable to delete MMM1 in 5x-emc cells , probably because the resulting strain was not viable . Therefore , we introduced the conditional mmm1-1 allele into 5x-emc cells . The 5x-emc mmm1-1 strain was not viable at the nonpermissive temperature of 37°C ( Figure 6B ) , indicating that cells required either ERMES or the EMC for viability . We next asked if ER to mitochondria PS transfer was more reduced in cells missing both ERMES and the EMC than in cells missing only the EMC . We found that when 5x-emc mmm1-1 cells were shifted to the restrictive temperature ( 37°C ) they stopped growing after ∼4 h . Therefore , we isolated crude mitochondria from 5x-emc mmm1-1 cells 3 h after shift to 37°C , to insure that the cells were still viable . In these mitochondria , the rate of PS to PE conversion was reduced ∼5-fold ( Figure 6A ) . Because mitochondria derived from 5x-emc mmm1-1 cells do not have less Psd activity than those from wild-type ( Figure 4C ) , these findings suggested that cells missing both ERMES and the EMC have very little PS transfer from the ER to mitochondria , which was consistent with the growth defect of these cells . The difference between the rates for 5x-emc and 5x-emc mmm1-1 mutants was not statistically significant , likely because of their already low rates of transfer . However , we found that the rate of PS to PE conversion in membranes derived from mmm1-1 single mutant cells was not significantly lower than wild-type , suggesting that the EMC was the critical component mediating ER to mitochondria PS transfer in vitro . The decrease in ER to mitochondria PS transfer in cells missing EMC proteins could be caused by a decreased ability to transfer PS or by inefficient tethering of the ER to mitochondria . The EMC might play a direct role in these processes or it could regulate the proteins that mediate lipid transfer or tethering . To characterize the nature of the PS transfer defect in cells missing EMC proteins , we determined if artificially tethering the ER and mitochondria corrects the PS transfer defect in mitochondria derived from 5x-emc and 5x-emc mmm1-1 cells . For these studies we used a fusion protein called ChiMERA , which has previously been shown to tether the ER and mitochondria [14] . When this fusion protein was expressed in 5x-emc and 5x-emc mmm1-1 cells , it corrected the ER to mitochondria PS transfer defect in mitochondria derived from these strains ( Figure 6B ) . It also restored the ability of 5x-emc mmm1-1 cells to grow at elevated temperature ( Figure 6C ) . These findings suggest that the defect in ER to mitochondria PS transfer in 5x-emc and 5x-emc mmm1-1 cells may be caused by inefficient tethering of the ER and mitochondria . To more directly determine if cells missing the EMC or ERMES complexes have defects in ER–mitochondria tethering , we used a previously described assay to quantitatively measure the association of these organelles [12] . A subdomain of the ER , often called the mitochondrial-associated membrane ( MAM ) , remains tightly associated with mitochondria after their purification , and is thought to be the portion of the ER that forms ER–mitochondria contacts . To quantitatively measure these contacts , we determined the percent of the ER that copurifies with mitochondria by measuring the fraction of the ER proteins , Dpm1 and Kar2 , in purified mitochondria . We found a significant decrease in the percent for ER co-purifying with mitochondria from mmm1Δ or 5x-emc cells ( Figure 7A ) . Because there were no significant differences in the percentages of mitochondria purified from the strains ( Figure 7A , right panel ) , these results indicate that there is a dramatic decrease in ER–mitochondria tethering in mmm1Δ and 5x-emc cells . We confirmed the decrease in ER–mitochondrial tethering in mmm1Δ and 5x-emc by ultrastructural analysis using transmission electron microscopy ( EM ) . We quantified the length of contacts where the ER membrane was within a distance of 30 nm to the mitochondria membrane and the ratio of the length of ER–mitochondria contacts to mitochondrial perimeter . Contacts were significantly reduced in both mmm1Δ and 5x-emc cells ( Figure 7B ) . There was also a significant difference between 5x-emc cells and 5x-emc mmm1-1 cells , suggesting that both EMC and ERMES tether independently of one another . Notably , this is the first demonstration of a reduction in ER–mitochondrial tethering in cells missing the ERMES complex . The defect in ER–mitochondria tethering we found in 5x-emc cells was not because ERMES complex formation was compromised in these cells . In wild-type cells , the ERMES complex forms about 1–10 puncta per cell [14] . These structures are thought to form at sites of ER–mitochondria tethering . In cells missing any one of the four ERMES proteins , the remaining proteins do not form puncta [14] . Therefore , if ERMES puncta are visible in 5x-emc cells , the EMC is not necessary for ERMES assembly . Indeed , we found that the ERMES protein Mmm1-GFP forms puncta in 5x-emc cells ( Figure 8A ) . In contrast , Mmm1-GFP was seen all over the ER in cells missing the ERMES protein Mdm10 ( Figure 8A , right panel ) . We also compared the number and intensity of Mmm1-GFP puncta between wild-type and 5x-emc cells . The ERMES puncta in the mutant were more numerous but slightly less intense than those in the wild-type cells , indicating that ERMES complex formation is not dramatically affected in the mutant . Thus , the ERMES complex can still assemble in 5x-emc cells , indicating that the ER–mitochondria tethering defect in 5x-emc cells is likely independent of ERMES-mediated tethering . This finding suggests that there must be ER–mitochondrial tethering proteins in addition to the ERMES complex . Even though the EMC is not required for ERMES assembly , it might play a role in the assembly of other tethering complexes , as it has been implicated in protein folding or quality control in the ER [29]–[33] . If there was a general defect in protein folding in mutants missing the EMC , the UPR might be induced in these cells . However , we found that the UPR was not induced in 5x-emc cells and that a UPR-responsive promoter is activated by ER stress in these cells ( Figure S4 ) . Because the EMC might directly participate in ER–mitochondria tethering , we sought to identify proteins in the outer mitochondrial membrane with which it could partner . We searched for genes identified in the CHO2 SGA screen that encoded mitochondrial outer membrane proteins . One gene , TOM5 , showed a strong genetic interaction with CHO2 that was rescued by choline ( Table S1 ) . TOM5 encodes one of three small integral membrane proteins in the outer membrane of mitochondria , which are nonessential subunits of the translocase of the outer membrane ( TOM ) complex that imports proteins into mitochondria [38] . Thus , Tom5 was a good candidate for interacting with the EMC , and we used PCA to determine if it interacts with EMC proteins . We found that all the Emc proteins interacted with Tom5 in puncta ( Figures 9A , B and S5A ) that were suggestive of the localization of the ERMES complex [14] . Interestingly , when we deleted the transmembrane domain of Tom5 , Tom5ΔTM now interacted with Emc2 on the ER ( Figure 9C ) , indicating that insertion of Tom5 in the outer mitochondrial membrane was not required for its interaction with the EMC . We next determined if the Emc–Tom5 PCA occurs at the same ER–mitochondria junctions where ERMES is localized . Remarkably , we found that in 100% of the cells we examined , the Emc1–Tom5 PCA puncta colocalized with ERMES foci ( Figure 9D ) . This indicated that the EMC–Tom5 interaction likely formed a tether between ER and mitochondria at the same or nearby contact sites as defined by ERMES . We did not detect an interaction between Tom5 and another integral ER protein , Ale1 , using PCA , even though Ale1–GFP localized throughout the ER similar to EMC proteins ( Figure S5B ) . Ale1 has a cytoplasmically oriented C-terminus , which should be accessible to Tom5 on mitochondria [39] . These findings support a specific role for the EMC at contacts between ER and mitochondria . To obtain more direct evidence for the interaction between the EMC and Tom5 , we used co-immuoprecipitation . Lysates from cells expressing GFP-tagged Emc1 , Emc2 , or Emc5 and Tom5 fused to the tandem affinity purification ( TAP ) tag were immunoprecipitated with agarose beads conjugated to anti-GFP antibodies . TAP–Tom5 co-immunoprecipitated with all three of the Emc proteins we tested ( Figure 9E ) . These results suggest that Emc proteins bind Tom5 and that the EMC may directly participate in ER–mitochondria tethering by interacting with the Tom complex . These data are also consistent with our PCA data showing that all the Emc proteins interact with Tom5 , which may explain why it was necessary to ablate a number of these proteins to obtain significant defects in PS transfer from the ER to mitochondria . Next we examined the functional relevance of the EMC–Tom5 interaction . To begin , we deleted TOM5 and measured PS transfer using the in vivo PS transfer assay; we found no PS transfer defect in cells missing Tom5 . However , we noticed that the PCA interaction on the ER between Emc2 and Tom5ΔTM caused cells to grow very poorly ( Figure 9F ) , whereas the PCA with full-length Tom5 grew similarly to wild-type , suggesting that the interaction of nonmitochondrial Tom5ΔTM with the EMC on the ER interfered with tethering and PS transfer from the ER to mitochondria . Therefore , we measured PS transfer in vivo in cells expressing either Emc2–VF2 or Tom5ΔTM–VF1 or both ( VF1 and VF2 are the two halves of the Venus fluorescent protein used in the PCA ) . ER to mitochondria PS transfer was decreased to ∼40% of that of wild-type in cells expressing both Emc2–VF2 and Tom5ΔTM–VF1 , whereas transfer was only modestly decreased in cells expressing only Tom5ΔTM–VF1 ( Figure 9G ) . Therefore , the PCA between Tom5ΔTM and Emc2 slowed PS transfer from the ER to mitochondria , suggesting the Emc may directly tether the ER and mitochondria by interacting with Tom5 in the outer mitochondrial membrane , and that this interaction is disrupted by the PCA between Emc2 and Tom5ΔTM .
Lipid exchange between the ER and mitochondria is critical for mitochondrial membrane biogenesis and lipid metabolism . We found that cells missing multiple components of a conserved protein complex in the ER , called EMC , had dramatic reductions in the amount of PS transferred from the ER to mitochondria both in vitro and in intact cells . These cells also had significantly lower levels of PS and PE in mitochondria . We found that the reduced ER to mitochondria PS transfer is probably caused by a significant reduction in the extent of tethering between the ER and mitochondria . Finally , we also demonstrate a role for the EMC in phospholipid metabolism and mitochondrial function by our unbiased global genetic analyses . Taken together , these findings strongly support that the EMC mediates lipid trafficking between the ER and mitochondria . We propose the EMC mediates ER–mitochondrial tethering independent of the previously identified ERMES complex . A number of findings support this claim . First , we found a significant decrease in the amount of ER associated with mitochondria both by determining the fraction of ER that copurified with mitochondria and by EM . Second , the PS transfer defect in cells missing multiple EMC proteins was ameliorated by expression of the artificial ER–mitochondria tether , ChiMERA . Third , we found a strong aggravating genetic interaction between the 5x-emc mutant and an ERMES temperature-sensitive allele , mmm1-1 , suggesting that the EMC contributes to an essential overlapping function with the ERMES . And finally , expression of ChiMERA completely rescued the synthetic lethality of the 5x-emc mmm1-1 combined mutant , confirming that this lethality arose as a result of deficient ER–mitochondrial tethering . How might the EMC mediate ER–mitochondrial tethering ? Because the EMC has been implicated in protein folding and quality control in the ER [29]–[33] , the EMC could be necessary for the folding or processing of proteins directly involved in tethering ( or lipid transfer ) , rather than participating in tethering directly . However , we found that ERMES complex assembly was not compromised in 5x-emc cells , indicating that the tethering defect we found in this strain was not due to a failure of ERMES proteins to assemble . We also found that in 5x-emc cells the UPR was not induced and occurred normally in response to ER stress ( Figure S4 ) . Importantly , we provide evidence by two independent methods that multiple Emc proteins interact with the mitochondrial outer membrane protein Tom5 , suggesting that the EMC–Tom5 interaction is important for tethering ER and mitochondria . Consistent with this , loss of the mitochondrial anchor for this tether by deleting the C-terminal transmembrane domain of Tom5 interfered with PS transfer to mitochondria and disrupted mitochondrial function . Therefore , we favor a model in which the EMC directly facilitates ER–mitochondrial tethering . Our findings indicate that the EMC and ERMES perform tethering functions independently of one another , as we found a significant decrease in tethering in mutants missing either complex . Whether the tethering defects due to loss of either complex are additive remains unclear . Why there are multiple tethering complexes is not yet known , but a plausible explanation is that each tether may facilitate a different type of communication between the ER and mitochondria . Given that there is a significant decrease in ER to mitochondria PS transfer in cells missing EMC , but not in cells lacking ERMES [17] , [18] , it may be that ERMES facilitates the exchange of other lipids or perhaps is involved in other forms of signaling/transport between the ER and mitochondria . Another possibility is that the proteins necessary for PS transfer associate with the EMC but not ERMES ( Figure S6 ) . Another notable difference between the EMC and ERMES is that only the EMC is conserved in higher eukaryotes [29] , [40] . Whether the mammalian EMC homologue also facilitates ER–mitochondria tethering and PS transfer remains an important question . Interestingly , our findings also suggest that 5x-emc cells have lipid metabolism defects in addition to reduced PS transfer between the ER and mitochondria . We found that 5x-emc psd2Δ cells were not viable and did not grow even when supplemented with ethanolamine . Because psd1Δ psd2Δ cells grow when ethanolamine is present in the medium , 5x-emc psd2Δ cells must have greater defects in lipid metabolism in addition to altered PS transfer to Psd1 in mitochondria . This conclusion is further supported by the results of our SGA screen with EMC6 , which identifies functional links to global transcriptional control of phospholipid metabolism ( INO2 , INO4 ) and fatty acid biosynthesis ( HTD2 , MCT1 ) . It is also likely that the EMC has functions in addition to its role in ER-mitochondrial membrane tethering , as indicated by other studies and as is implied by its uniform localization throughout the ER , which is not restricted to contacts [29]–[33] . One of the more perplexing aspects of EMC function is the apparent redundancy between individual Emc proteins and Tom5 in tethering and PS transfer . That is , why is deletion of individual Emc proteins or Tom5 not sufficient to reduce PS transfer to mitochondria ? Although there is no shared sequence identity amongst Emc proteins that would suggest obvious functional redundancy , our finding that multiple Emc proteins interact with Tom5 provides a sound alternative explanation for functional overlap in tethering . However , one would then expect that deletion of Tom5 would disrupt tethering and PS transfer , similarly to loss of multiple Emc proteins . That this appears not to be the case suggests that there is additional redundancy in tethering on the mitochondrial membrane , perhaps involving other TOM complex proteins . Although there is no sequence identity between Tom5 and the two other small TOM subunits , Tom6 and Tom7 , all these proteins are similar in size and topology and appear to be functionally redundant within the TOM complex [41] , [42] . Hence , the EMC may interact with Tom6 and Tom7 , and , indeed , preliminary PCA evidence indicates that Tom6 and Tom7 interact with certain Emc proteins ( unpublished data ) . A role for the TOM complex in tethering between the mitochondrial outer and inner membranes through interaction with the translocase of the inner membrane ( TIM ) complex is well established [43] . Thus , our work further implies that tethering between TOM and the EMC , and TOM and TIM may facilitate efficient PS transfer from the ER to the inner mitochondrial membrane , the location of PS decarboxylation by Psd1 [16] .
Strains and plasmids used in this study are listed in Table S13 . Media used were YPD ( 1% yeast extract , 2% peptone , 2% glucose ) , YPGly ( 1% yeast extract , 2% peptone , 3% glycerol ) , and synthetic complete ( SC ) media ( 2% glucose , 0 . 67% yeast nitrogen base without amino acids , and amino acid dropout mix from BIO101 ) . Where indicated , ethanolamine or choline was added to a concentration of 5 mM . 5-FOA was added at a concentration of 1 mg/ml . Single deletion strains were obtained from freezer stocks of the haploid yeast deletion collection ( BY4741 , Mat a , KanMX , a gift from C . Boone ) unless otherwise stated . Other gene deletions were constructed using the PCR method with the heterologous markers S . pombe HIS5 ( pKT128 ) , K . lactis URA3 ( pKT209 ) , or NatR ( p4339 ) . Double deletion strains were derived from the meiotic products of heterozygous diploids with at least three spores of each genotype being compared . All yeast cells expressing GFP fusion proteins were tagged endogenously in haploids unless otherwise indicated . C-terminally tagged GFP strains were constructed by standard methods involving single-step gene replacement using the pKT128 ( SpHIS5 ) plasmid [44] in the wild-type Y7043 background and crossed to the indicated single deletion mutants ( BY4741; kanMX ) . Mdm12–RFP was created similarly to C-terminal GFP fusions except using pMRFP–NAT in BY4741 . For PCA , the plasmids used for C-terminal genomic tagging were created as follows: Venus–YFP fragments F1 and F2 were amplified by PCR from p413–TEF–Zip linker–Venus YFP-F1 and p415–TEF–Zip linker–Venus YFP-F2 ( gift of S . Michnick [45] ) , and cloned into plasmids pKT128 and pKT209 [44] , respectively , replacing yEGFP and including a myc tag ( MEQKLISEEDL ) in the linker region to give pHVF1CT ( pFA6a–myc–VF1–HIS5 ) and pUVF2CT ( pFA6a–myc–VF2–URA3 ) . For N-terminal integrations , plasmid pHVF1NT ( HIS5–PHO5–VF1 ) was created by replacing eGFP in plasmid pTLHPG ( HIS5–PHO5–GFP; gift of T . Levine ) with the Venus F1 PCR fragment . To create the Tom5ΔTM–VF1 strain , the C-terminal 21 amino acids of endogenous TOM5 were replaced in frame by VF1 . SGA analysis was performed according to established protocols [46] essentially as previously described [47] using a Singer RoToR Colony Arraying robot ( Singer Instruments ) . Δcho2::URA3 and Δemc6::URA3 query strains were constructed using standard techniques in strain background Y7092 and crossed to the yeast haploid deletion mutant array ( DMA ) using a Singer RoToR HDA robot . Following diploid selection , spots were replicated three times and sporulated for 5 d . Haploids were germinated on SD-media lacking histidine , arginine , and lysine supplemented with thialysine and canavanine ( both at 100 µg/ml ) . Control sets of single deletion strains were generated by plating on media containing 5-FOA to counterselect for the Δcho2::URA3 or Δemc6::URA3 alleles and G418 sulfate ( 200 µg/ml ) to select for the DMA strain , whereas double mutants were selected for by plating on media lacking uracil and containing G418 sulfate . A further round of selection was performed on the same media . For the CHO2 SGA screen in the presence of choline , all plates additionally contained 1 mM choline . Arrays were imaged using a flatbed scanner . Balony software ( http://code . google . com/p/balony/ ) was used to measure spot sizes , determine cutoff values for genetic interactions , and define strains that showed statistically significant changes in growth rate [48] . Cutoff values for genetic interactions were defined for each screen by determining three standard deviations from the mean of the ratios of the double mutant to single mutant growth rates . Double mutant strains that met the cutoff and showed significant changes in growth relative to the corresponding single mutant control ( one-tailed student's t test; p<0 . 05; n = 3 ) were considered as genetic interactions . For the CHO2 SGA screen , aggravating genetic interactions identified in the screen done in the absence of choline were considered rescued if they were no longer identified as genetic interactions according to the above criteria in the screen done in the presence of choline . Gene ontology analysis was performed using Funspec ( http://funspec . med . utoronto . ca ) and Cytoscape ( http://www . cytoscape . org ) [49] , [50] . Log phase live yeast cells were imaged using a Zeiss LSM-5 Pascal confocal microscope and Zeiss Pascal software . Unless otherwise stated , all proteins were tagged at the C-terminus of the endogenous protein . Optical slices were taken through the center of each cell , and images being directly compared were captured with identical microscope settings on the same day . For the studies with cells expressing Mmm1–GFP , cells were imaged live at room temperature by using an Olympus BX61 microscope , a UPlanApo ×100/1 . 35 lens , a QImaging Retiga EX camera , and IVision software ( version v 4 . 0 . 5 ) . The Venus–YFP variant of PCA was used to examine protein–protein interactions in live yeast . Unless otherwise stated , endogenous proteins were tagged in haploid yeast by the PCR method with either VF1 or VF2 in the BY4741 and Y7043 strains , respectively . Correct integration and expression was confirmed by colony PCR and Western blot analysis with anti-myc antibodies ( Sigma ) for each fusion protein . The VF1 and VF2 strains to be assayed were then crossed , and haploid meiotic progenies with both alleles were recovered by random spore analysis or tetrad dissection . Finally , the PCA was visualized in log phase yeast by confocal microscopy . Cells were labeled with L-[3-3H]serine ( American Radiolabeled Chemicals ) as described [35] with the following modifications . About 2 OD600 units of cells from a saturated culture were added to 25 ml of SC medium and incubated at 30°C . When the cultures reached an OD600 of about 0 . 3 , 10 µg/ml myriocin ( SigmaAldrich , stock = 500 µg/ml in methanol ) was added to the medium , the cells were grown for 30 min , and 10 µg/ml cerulenin ( SigmaAldrich , stock = 5 mg/ml in dimethyl sulfoxide ) was added to the medium . About 5 min later , 50 µCi of [3H]serine was added to the medium , and the cells were grown for an additional 30 min . The culture was then added to an equal volume of ice-cold water , and it was washed once with ice-cold water . Cells were lysed in a Mini-BeadBeater-8 ( BioSpec ) . Lipids were extracted as described [51] , separated by HPLC [52] , and the fractions containing PS , PE , and PC were collected and analyzed by liquid scintillation counting . Crude mitochondria were prepared as described [18] . Briefly cells were grown in YPD medium to an OD600 of ∼0 . 3 , washed once with water , and incubated in 1 ml 0 . 1 M Tris-SO4 ( pH 9 . 4 ) containing 10 mM DTT for 10 min at 30°C . They were washed once with spheroplast buffer ( 1 . 2 M sorbitol , 20 mM Tris pH 7 . 4 ) and resuspended in 1 . 5 ml of the same buffer containing 1 mg/ml zymolyase 20T ( Seikagaku Biobusiness , Japan ) . After incubation for 60 min at 30°C , cells were pelleted ( 5 min , 500×g ) and washed twice with spheroplast buffer . Cells were resuspended in ice-cold lysis buffer ( 0 . 6 M mannitol , 2 . 0 mM Tris pH 7 . 4 , 1 mM EDTA , 1 mM PMSF and protease inhibitors , [Roche] ) and lysed with a dounce using a B-pestle . The extract was centrifuged twice for 5 min at 3 , 000×g to remove unlysed cells and debris . The supernatant was centrifuged at 9 , 600×g for 10 min and the pellet containing crude mitochondria was resuspended in lysis buffer using a dounce ( B-pestle ) . The method of labeling crude mitochondria with [3H]serine was adapted from [37] . We heated 1–2 mg of crude mitochondria in 1 ml of lysis buffer to 30°C , and 0 . 6 µa MnCl2 and 10 µCi of L-[3-3H]serine ( American Radiolabeled Chemicals ) were added . After 20 min , 0 . 5 mM serine and 5 mM EDTA were added . Samples of 200 µl were taken after 0 , 5 , 10 , and 15 min and added to 6 ml of chloroform:methanol ( 1∶2 ) . Lipids were extracted , separated by HPLC , and extracted as described in the previous section . Forty OD600 of yeast cells were grown to mid-log phase in SC media and collected by centrifugation . After washing once in 0 . 67% w/v Yeast Nitrogen Base , cells were resuspended in 12 ml of 0 . 67% yeast nitrogen base containing 2% dextrose and 12 µCi [3H]ethanolamine ( Perkin-Elmer ) . After labeling for 30 min at 37°C , the suspension was diluted to a final volume of 120 ml in synthetic defined media . The culture was placed in a shaking incubator at 30°C and 20 ml aliquots taken every 30 min . Lipids were extracted and separated by HPLC as described in the previous section . About 2 OD600 units of cells from a saturated culture were washed with water , resuspended in 50 ml of fresh SC medium containing 200 µCi of [3H]acetate ( American Radiolabeled Chemicals ) , and grown at 30°C for at least 3–4 generations . Crude mitochondria were purified as described in the previous section and further purified by equilibrium centrifugation using density gradients made from OptiPrep ( Axis-Shield , Oslo , Norway ) as described [53] . Lipids were extracted and separated by one-dimensional TLC as described by [54] . TLC plates were scanned on a RITA Star Thin Layer Analyzer ( Raytest ) . Mitochondria were purified as described in the previous section . Portions of the cell lysates and purified mitochondria were immunoblotted using antibodies against the ER-membrane proteins Dpm1 ( Invitrogen ) or Kar2 ( gift from R . Schekman ) and the mitochondrial membrane protein Por1 ( Invitrogen ) . The immunoblots were quantitated using the Odyssey infrared imaging system ( LI-COR Biosciences ) , and the percentage of each marker in the purified mitochondria was determined . Psd assays were performed as described [35] except that the concentration of the substrate , 1-oleoyl-2-[12-[ ( 7-nitro-2-1 , 3-benzoxadiazol-4-yl ) amino]dodecanoyl]-sn-glycero-3-phosphoserine ( Avanti Polar Lipids ) , was 500 µM . Yeast cultures were grown in YP+2% galactose or YP+2% dextrose to express or suppress TAP–Tom5 , respectively . Log phase cultures ( 30 OD600 units ) were collected by centrifugation and resuspended in 400 µl of lysis buffer ( 250 mM NaCl , 50 mM Tris-HCl pH 7 . 4 , 50 mM NaF , 5 mM EDTA , 1 mM DTT , 1 mM AEBSF , and 0 . 1% NP-40 ) . Cells were disrupted with glass beads for 5 min at 4°C . The resulting lysate was centrifuged at 16 , 100×g for 5 min , and a sample of the supernatant was collected . The supernatant was then incubated under rotation with 20 µl of prewashed GFP–Trap_A bead slurry ( Chromotek ) for 1 h at 4°C . The supernatant was removed and the beads were washed with lysis buffer ( 3×1 ml , then 3×500 µl with 5-min incubations ) . Beads were resuspended in 100 µl of sample buffer ( 50 mM Tris-HCl pH 6 . 8 , 10% glycerol , 2% SDS , 4% 2-mercaptoethanol , and 0 . 02% bromophenol blue ) and eluted by heating at 65°C for 20 min . The eluted protein was subjected to SDS-PAGE ( 10% polyacrylamide ) and immunoblotted with anti-TAP ( 1∶5 , 000 dilution , Pierce , CAB1001 ) or anti-GFP ( 1∶5 , 000 dilution , Roche , 11814460001 ) antibodies . Goat anti-rabbit IgG-HRP ( 1∶5 , 000 dilution , Bio-Rad , 172-1019 ) and goat anti-mouse IgG-HRP ( 1∶5 , 000 dilution , Bio-Rad , 172-1011 ) were used as secondary antibodies . Blots were imaged with either Supersignal West Pico or Femto ECL substrate ( Thermo Scientific ) . Cells were grown to mid-logarithmic growth phase , and 10 OD600 units of cells were fixed in 1 ml of fixative media ( 1% glutaraldehyde , 0 . 2% paraformaldehyde , and 40 mM potassium phosphate , pH 7 . 0 ) for 10 min at room temperature . The cells were pelleted and resuspended in 1 ml of fresh fixative media and incubated on ice for 50 min , washed twice with 0 . 9% NaCl and once with water . They were then incubated with 2% solution of KMnO4 for 5 min at room temperature , centrifuged , and resuspended with a freshly prepared solution of 2% KMmO4 for 45 min at room temperature for en-bloc staining . The cells were then dehydrated using a graded series of ethanol solutions ( 50% , 70% , 80% , 90% , 95% , and 100% ) for 10 min each , with two more incubations in 100% ethanol from a freshly opened bottle . The samples were subsequently embedded stepwise using Spurr's low viscosity resin ( EMS , Hatfield , PA ) . Samples were infiltrated for 2 h each with a 3∶1 , 1∶1 , and 1∶3 dehydrating agent–embedding media mixture . Cells were incubated overnight with 100% fresh resin . The next day , the cells were resuspended in fresh 100% resin for 2–3 h , transferred into BEEM embedding capsules , and polymerized at 70°C for 72 h . Semi- and ultrathin sections were produced with a diamond knife ( Diatome , Biel , Switzerland ) on an ultra-microtome ( Ultracut UCT , Leica-Microsystems ) , collected on 200 mesh copper grids ( EMS , Hatfield , PA ) , poststained with uranyl acetate and lead citrate , and visualized with a FEI Tecnai T12 TEM , operating at 120 kV . Pictures were recorded on a below mounted Gatan 2k×2k CCD camera . Contact sites between the ER and mitochondria were defined as regions where these organelles come within 30 nm of one another . The ER–mitochondria ratio was determined by measuring the length of contacts between the ER and mitochondria divided by the length of mitochondrial perimeter in each image . Cells were transformed with pMCZ-Y , a plasmid encoding the lacZ gene under the KAR2 promoter . Where indicated , 1 mM DTT was added to growing cells 1 h before the assay . β-Galactosidase assay was performed as described [55] . | Mitochondrial membrane biogenesis and lipid metabolism depend on the transfer of phospholipid from the endoplasmic reticulum to mitochondria . This transfer is thought to occur at regions where these organelles are in close contact , and , although the process is thought not to involve vesicles , the mechanism is not known . In this study , we found a complex of proteins in the endoplasmic reticulum that is required for the transfer of one phospholipid—phosphatidylserine—from the endoplasmic reticulum to mitochondria . Cells lacking this protein complex have nonfunctional mitochondria with an abnormal lipid composition . We show that the complex is required to maintain close contacts between the endoplasmic reticulum and mitochondria; the complex probably directly interacts with at least one protein on the surface of mitochondria . In addition , cells lacking this complex and a second previously identified tethering complex are not viable . Thus , our findings suggest that tethering of the endoplasmic reticulum and mitochondria is essential for cell growth , likely because it is necessary for lipid exchange between these organelles . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"biology",
"and",
"life",
"sciences"
]
| 2014 | A Conserved Endoplasmic Reticulum Membrane Protein Complex (EMC) Facilitates Phospholipid Transfer from the ER to Mitochondria |
Wolbachia is an intracellular bacterium that infects a remarkable range of insect hosts . Insects such as mosquitos act as vectors for many devastating human viruses such as Dengue , West Nile , and Zika . Remarkably , Wolbachia infection provides insect hosts with resistance to many arboviruses thereby rendering the insects ineffective as vectors . To utilize Wolbachia effectively as a tool against vector-borne viruses a better understanding of the host-Wolbachia relationship is needed . To investigate Wolbachia-insect interactions we used the Wolbachia/Drosophila model that provides a genetically tractable system for studying host-pathogen interactions . We coupled genome-wide RNAi screening with a novel high-throughput fluorescence in situ hybridization ( FISH ) assay to detect changes in Wolbachia levels in a Wolbachia-infected Drosophila cell line JW18 . 1117 genes altered Wolbachia levels when knocked down by RNAi of which 329 genes increased and 788 genes decreased the level of Wolbachia . Validation of hits included in depth secondary screening using in vitro RNAi , Drosophila mutants , and Wolbachia-detection by DNA qPCR . A diverse set of host gene networks was identified to regulate Wolbachia levels and unexpectedly revealed that perturbations of host translation components such as the ribosome and translation initiation factors results in increased Wolbachia levels both in vitro using RNAi and in vivo using mutants and a chemical-based translation inhibition assay . This work provides evidence for Wolbachia-host translation interaction and strengthens our general understanding of the Wolbachia-host intracellular relationship .
Insects are common vectors for devastating human viruses such as Zika , Yellow Fever , and Dengue . A novel preventative strategy has emerged to combat vector-borne diseases that exploits the consequences of vector-insect infection with the bacteria Wolbachia pipientis [1–4] . Wolbachia is a vertically transmitted , gram-negative intracellular bacterium known to infect 40–70% of all insects [5 , 6] . Wolbachia provides hosts with resistance to pathogens such as viruses [7–10] . Remarkably , Wolbachia infections can reduce host viral load enough to render insect hosts incapable of transmitting disease-causing viruses effectively [1 , 2 , 11–24] . The relationship between Wolbachia and a host is complex and dynamic . Understanding how bacterial levels can change is vital because it dictates how Wolbachia manipulates the host insect . For example , the antiviral protection provided by Wolbachia is strongest when Wolbachia levels within a host are high [10 , 25–27] . On the other hand , Wolbachia can become deleterious to the host when Wolbachia population levels are too high leading to cellular damage and reduced lifespan[28–30] . To apply Wolbachia as an effective tool to combat vector-borne viruses we need a better understanding of host influences on Wolbachia levels . Wolbachia infects a large host and tissue range suggesting interaction with various host systems and pathways for successful intracellular maintenance within a host [5 , 31] . To date , reports suggest that Wolbachia levels may be influenced in various contexts by interaction with host cytoskeletal components [32–35] , the host ubiquitin/proteasome [36] , host autophagy [37] , and by host miRNAs [16 , 38] . A comprehensive analysis of host systems that influence Wolbachia levels has not been carried out and will further our knowledge of this symbiotic relationship and reveal molecular mechanisms that occur between Wolbachia and the host to maintain it . Wolbachia-host interactions can be studied in the genetically tractable Drosophila melanogaster system which allows for the systematic dissection of host signaling pathways that interact with the bacteria using the wide array of genetic and genomic tools available . The Drosophila system enables rapid unbiased screening of host factors that impact Wolbachia at the cellular and organismal level . While some influences on the relationship , such as systemic effects , require studies in the whole organism , many aspects of molecular and cellular signaling can be studied in a Drosophila cell culture-based system . Drosophila cells are particularly amenable to genome-scale screens because of the ease and efficacy of RNAi in this system [39]-[40] . Previous cell culture-based RNAi screening has been a successful approach to study a wide range of intracellular bacteria-host interactions in Drosophila cell lines [41–44] . Thus , we reasoned that this was a feasible approach for studying Wolbachia-host interactions . Here we performed a whole genome RNAi screen in a Wolbachia-infected Drosophila cell line , JW18 , which was originally derived from Wolbachia-infected Drosophila embryos and has previously proven suitable for high-throughput assays [36 , 45] . The goal was to determine in an unbiased and comprehensive manner which host systems affect intracellular Wolbachia levels . The primary screen identified 1117 host genes that robustly altered Wolbachia levels . Knock down of 329 of these genes resulted in increased Wolbachia levels whereas 788 genes led to decreased Wolbachia levels . To characterize these genes , we generated manually curated categories , performed Gene Ontology enrichment analysis , and identified enriched host networks using bioinformatic analysis tools . The effects on Wolbachia levels were validated in follow-up RNAi assays that confirmed Wolbachia changes visually by RNA FISH as well as quantitatively using a highly sensitive DNA qPCR assay . We uncovered an unexpected role of host translation components such as the ribosome and translation initiation factors in suppressing Wolbachia levels both in tissue culture using RNAi and in the fly using mutants and a chemical-based translation inhibition assay . Furthermore , we show a decrease in overall translation in Wolbachia-infected JW18 cells compared to Wolbachia-free JW18DOX cells and that an inverse trend exists between Wolbachia levels and host translation levels in JW18 cells . This work provides strong evidence for a relationship between Wolbachia and host translation and strengthens our general understanding of the Wolbachia-host intracellular relationship .
Wolbachia is an intracellular bacterium that resides within a wide range of insect hosts . To identify host factors that enhance or suppress intracellular Wolbachia levels , we performed a genome-wide RNAi screen in Wolbachia-infected JW18 Drosophila cells that were originally derived from Wolbachia-infected embryos [45] . In order to visually detect Wolbachia levels we established a specific and sensitive RNA Fluorescence In Situ Hybridization ( FISH ) method consisting of a set of 48 fluorescently labeled DNA oligos that collectively bind in series to the target Wolbachia 23s rRNA ( Fig 1A ) . This enabled detection of infection levels ranging from as low as a single bacterium in a cell to a highly infected cell and could clearly distinguish Wolbachia-infected cells from Wolbachia-free cells ( Fig 1B ) . Thus , we were able to assess Wolbachia infection levels in the JW18 cell population and found that under our culturing conditions we could stably maintain JW18 cells with a Wolbachia infection level of 14% of the JW18 cells ( Fig 1C ) . Of the infected cells , 73% of the cells had a low Wolbachia infection ( 1–10 bacteria ) , 13 . 5% had a medium infection ( 11–30 bacteria ) , and 13 . 5% were highly infected ( >30 bacteria ) . Though Wolbachia levels may change in different culturing conditions , the JW18 cell line maintained Wolbachia levels stably for the duration of the screen . These experiments confirmed the feasibility and sensitivity of RNA FISH to detect different levels of Wolbachia infection in Drosophila cells in a highly sensitive manner . Prior to screening we characterized the JW18 cell line and its associated Wolbachia strain by generating a JW18 DNA library and sequencing it using DNAseq technology ( S1 Fig ) . This allowed for phylogenetic analysis of the Wolbachia strain and revealed that it clustered most closely with the avirulent wMel strain which is well characterized for its antiviral effect on RNA-based viruses in Drosophila as well as in mosquitos ( S1A Fig ) [1 , 2 , 7 , 8] . Further analysis included gene copy number variation of the Wolbachia genome and identified one deleted and one highly duplicated region ( 3–4 fold increased ) ( S1B Fig ) . The deleted region contained eight genes known as the “Octomom” region postulated to influence virulence [27 , 46] . The loss of “Octomom” has also been reported in wMelPop-infected mosquito cell lines after extended passaging over 44 months [47] . This suggests that loss of this region happened independently in two cases and may be related to passage in cell culture . A highly duplicated region spans approximately from positions 91 , 800–127 , 100 and contains 38 full or partial genes , including those with unknown function as well as genes predicted to be involved in metabolite synthesis and transport , molecular chaperones , DNA polymerase III subunit , DNA gyrase subunit , and 50S ribosomal proteins . For analysis of gene copy number variation in the JW18 cell line , the DNA library was aligned to the Release 6 reference genome of D . melanogaster . This revealed that the cell line is of male origin with an X:A chromosomal ratio of 1:2 and tetraploid in copy number ( S1C Fig ) . Bioinformatic analysis on genes of high or low copy number did not reveal an enrichment for any particular molecular or cellular functional class and the majority ( 72% ) of genes in the JW18 cell line were at copy numbers expected for a tetraploid male genotype ( 4 copies on autosomes , 2 copies on X ) . This made the JW18 cell line suitable for RNAi screening . As a first step to uncovering Wolbachia-host interactions , we asked whether gene expression changes occur in the host during stable Wolbachia infection . To do this , we used a control Wolbachia-free version of the JW18 cell line which was previously generated through doxycycline treatment ( JW18DOX ) ( S2A Fig ) . A comparison of host gene expression changes in the presence and absence of Wolbachia through RNAseq analysis revealed 308 and 559 host genes that were up- or down-regulated respectively by two-fold or more ( padj<0 . 05 ) ( S2B Fig ) . Of these genes , 21 displayed major expression changes of log2 fold >4 ( DptB , Wnt2 , SP1173 , bi , FASN3 , CG5758 , CP7Fb , beta-Man ) or log2-fold <-4 ( CG12693 , CG13741 , Tsp74F , esn , cac , CG4676 , CG42827 , CG18088 , CG17839 , 5-HT2A , CG43740 , CG3036 , aru ) ( also see S2C Fig ) . The presence of Wolbachia led to elevated gene expression of several components of the host immune response including the Gram-negative antimicrobial peptide Diptericin B ( DptB ) , which was the most highly upregulated gene in the presence of Wolbachia ( S2C Fig ) . Gene ontology ( GO ) analysis further confirmed a host immune response with enriched terms such as ‘response to other organism’ and ‘peptidoglycan binding’ that included genes for antimicrobial peptides ( attA , AttB , AttC , DptB , LysB ) and peptidoglycan receptors ( PGRP-SA , -SD , -LB , -LF ) as well as antioxidants such as Jafrac2 , Prx2540-1 , Prx2540-2 , Pxn , GstS1 with ‘peroxiredoxin’ and ‘peroxidase activity’ . Other expression changes included extracellular matrix components such as upregulation of collagen type IV ( Col4a1 and vkg ) and downregulation of genes for integral components of the plasma membrane including cell adhesion components ( kek5 , mew , Integrin , and tetraspanin 42Ed and 39D ) . Gene ontology analysis further identified a significant enrichment of ion transporters and channels that were downregulated as well as genes encoding several proteins such as myosin II , projectin and others associated with the muscle Z-disc that were downregulated . Finally , we observed an overall upregulation of host proteasome components at the RNA level in the presence of Wolbachia ( S3 Fig ) , which is in line with proteomics data of proteasome upregulation in the presence of Wolbachia [48 , 49] . In summary , these host factors may play an important role in the Wolbachia-host relationship however their specific roles in this interaction remain to be determined . The screening approach combined the visual RNA FISH Wolbachia detection assay ( Fig 1 ) with in vitro RNAi knockdown of host genes to ask which host genes influence Wolbachia levels ( Fig 2A ) . Prior to screening , we tested whether RNAi was a feasible approach in JW18 cells . First , we confirmed that RNAi had no adverse effects such as cytotoxicity on the cells using a negative control dsRNA targeting LacZ which was not present in our system ( Fig 2B and 2C ) . Second , we tested RNAi knockdown efficiency in the JW18 cell line . To do this a Jupiter-GFP transgene present in the cell line was targeted for knockdown using dsRNA to GFP . High knockdown efficiency was achieved using this RNAi protocol as seen by the efficient knockdown of the Jupiter-GFP transgene both visually by RNA FISH ( 90 . 2% reduction ) ( Fig 2D and 2E ) and by protein levels as shown by Western blot ( 97 . 9% reduction ) ( Fig 2F ) compared with either the ‘no dsRNA’ knockdown ( Fig 2B ) or ‘LacZ’ knockdown ( Fig 2C ) conditions . This confirmed the suitability of the JW18 cell line for an RNAi-based screening approach . For controls that alter Wolbachia levels , we identified a host ribosomal gene , RpL40 , from a pilot screen that consistently led to increased Wolbachia levels when depleted by RNAi ( Fig 2G ) compared to cells that were not treated by RNAi ( Fig 2B ) or treated with lacZ dsRNA treatment ( Fig 2C ) . We achieved 96 . 3% RNAi knockdown efficiency as confirmed by qPCR for RpL40 levels relative to a no knockdown control ( Fig 2H ) . At the time of the screen we did not know of any host protein whose knockdown would decrease Wolbachia levels . Therefore , as a Wolbachia-decreasing control , cells were incubated with 5μM doxycycline for 5 days which successfully reduced the Wolbachia levels in the cells by 91 . 9% as measured by RNA FISH ( Fig 2I and 2J ) . To quantify the effect of the controls on Wolbachia levels we isolated genomic DNA from each treated sample and used quantitative PCR DNA amplification to detect the number of Wolbachia genomes per cell by measuring Wolbachia wspB copy number relative to the Drosophila gene RpL11 ( Fig 2K ) . Relative to control cells , the RNAi treatment with RpL40 resulted in a 3 . 4-fold increase in Wolbachia , doxycycline decreased Wolbachia levels 6 . 3-fold , whereas LacZ and GFP RNAi had no significant effect confirming that our controls allowed us to manipulate Wolbachia levels in the JW18 cell line and that this cell line with its relative low infection rate ( Fig 1C ) provided a sensitive tool for detecting dynamic changes in Wolbachia levels through an RNAi screening approach . The layout of the whole genome screen is illustrated in Fig 2A . Briefly , Wolbachia-infected JW18 cells were incubated with the DRSC Drosophila Whole Genome RNAi Library version 2 . 0 which was pre-arrayed in 384 well tissue culture plates such that each well contained a specific dsRNA amplicon to target one host gene . The 5-day incubation period allowed for efficient host gene knockdown . Thereafter the cells were processed for RNA FISH detection of Wolbachia 23s rRNA . Total fluorescence signal was detected using automated microscopy and served as a readout for Wolbachia levels within each plate well . Host cells within each well were detected by DAPI staining . Finally , the Wolbachia fluorescence signal was divided by the total number of DAPI-stained host cells detected to provide an average Wolbachia per cell readout which was normalized to the plate average ( represented as a robust Z score ) . The library was screened in triplicate . The raw screening data were subjected to several quality control steps ( S4 Fig ) . Briefly , we realigned the DRSC Version 2 . 0 Whole Genome RNAi library dsRNA amplicons with Release 6 of the D . melanogaster genome using the bioinformatic tool UP-TORR [50] . This provided an accurate updated description of the gene target for each dsRNA amplicon . Initially the library included 24 036 unique dsRNA amplicons targeting 15 589 genes , however owing to updates in gene organization and annotation models of the reference genome since the initial release of the library we removed 1499 outdated amplicons from our subsequent analysis as they were no longer predicted to have gene targets ( S1 Table ) . We also excluded 1481 amplicons that were annotated in UP-TORR to target multiple genes ( S2 and S3 Tables ) . We further excluded 66 amplicons for a positional effect on the dsRNA library tissue culture plates at the A1 position ( S5 Fig , S4 Table ) . Thus , we effectively screened 20 990 unique dsRNA amplicons targeting 14 024 genes ( 80% of D . melanogaster Release 6 genome ) . A further quality control step to reduce false positive hits was to cross-reference potential hits with RNAseq gene expression data for the JW18 cell line to exclude genes with undetectable expression in the cell line ( S5 Table ) . To identify and select for hits from the primary data , we first analyzed the screen-wide controls . A plot of all controls included in the whole genome screen revealed that RpL40 knockdown increased Wolbachia levels ( median robust Z score of 2 . 2 ) , conversely doxycycline treatment decreased Wolbachia levels throughout the screen ( median robust Z of -3 . 5 ) , whereas a standard control included in the whole genome library , Rho1 , and GFP RNAi knockdown did not significantly affect Wolbachia levels ( Fig 3A ) . We used this range as a guide to set robust Z limits for primary hits at ≥ 1 . 5 or ≤ -1 . 5 . Every dsRNA amplicon was screened in triplicate . To be considered as a ‘hit’ amplicon at least 2 of the 3 replicates needed to satisfy the robust Z score limits ( S4 Fig ) . To categorize the primary screen hits , each gene was assigned to a ‘High’ , ‘Medium’ , and “Low’ bin based on the confidence level ( S4 Fig ) . This was determined based on the total number of different dsRNA amplicons representing a hit gene in the library and how many of these dsRNA amplicons had a significant effect on Wolbachia levels ( S4 Fig ) . In this manner , we were able to stratify the primary screen hits to assist in follow up analysis . The screen identified 1117 genes that when knocked down had a significant effect on the Wolbachia levels in JW18 cells ( S6 Table ) . Knock down of 329 of the 1117 genes resulted in increased Wolbachia levels , suggesting that these genes normally restrict Wolbachia levels within the host cell ( Fig 3B ) . Knockdown of 788 genes resulted in decreased Wolbachia levels , suggesting Wolbachia may be dependent on these host genes for survival within the host cell ( Fig 3B ) . For each of the two hit categories , genes were classified by confidence level ( described in S4 Fig , and Fig 3C ) . We found a higher proportion of low confidence hits ( 21% ) in the category of genes that decreased Wolbachia levels compared to genes that led to Wolbachia level increases which only contained 12 . 5% low confidence hits . To analyze the expression of the 1117 genes , the hits were distributed into 9 bins based on their gene expression level from JW18 RNAseq data ( S6A Fig ) . Hits displayed a wide range of expression and an enrichment of low expression for hits that decreased Wolbachia levels ( S6B Fig ) . We did not observe any biases for variation in gene DNA copy number based on DNAseq data for the JW18 cell line ( S6C Fig ) . Next , we asked whether changes in Wolbachia levels could be explained by effects on host cell proliferation or were independent of effects on host cell proliferation . We measured cell proliferation using the raw screen data by normalizing the number of cells scanned per well ( DAPI ) to the number of fields of view required to capture the cells . This allowed us to generate a robust Z score measure of cell proliferation effects for the 1117 genes identified as hits . For genes that increased Wolbachia levels , 12% ( 41 genes ) increased cell proliferation ( robust Z>1 ) , 45% ( 147 genes ) decreased cell proliferation ( robust Z<-1 ) , and 43% ( 141 genes ) had no effect on cell proliferation ( Fig 3D , S7 Table ) . These data suggest that a significant number of gene knockdowns ( 45% ) may indirectly lead to an increase in Wolbachia levels through slowed cell proliferation . Importantly , 43% of hits identified had no effect on cell proliferation whilst increasing Wolbachia levels . These results suggest that changes in Wolbachia levels are not strictly linked to host cell proliferation . For genes that decreased Wolbachia levels , the majority ( 82% , 644 genes ) did not affect cell proliferation and 2% ( 19 genes ) increased and 16% ( 125 genes ) decreased cell proliferation ( Fig 3D , S7 Table ) . To summarize , the screen identified 1117 host genes that act to support or suppress Wolbachia levels within the host Drosophila cell . To classify the 1117 gene hits identified in the whole genome screen , we first manually curated the hits using gene annotation available on FlyBase ( http://www . flybase . org ) relating to each gene such as gene family , domains , molecular function , gene ontology ( GO ) information , gene summaries , interactions and pathways , orthologs , and related recent research papers . We identified distinct categories of genes that when knocked down by RNAi increased ( Fig 4A ) or decreased ( Fig 4B ) Wolbachia levels . The largest gene category that led to decreased Wolbachia levels by RNAi knockdown contained genes for host metabolism and transporters suggesting that Wolbachia strongly relies on this aspect of the host ( Fig 4B ) . On the other hand , gene knockdowns that increased Wolbachia contained many components of the core ribosome network , translation factors , and the proteasome core and regulatory proteins network ( Fig 4A ) . Six of the broad gene categories could be further sub-classified for processes that either enhanced or suppressed Wolbachia levels . First , RNAi knockdown of members in the category containing cytoskeleton , cell adhesion and extracellular matrix components decreased Wolbachia , these included cadherins , formins , spectrin and genes involved in microtubule organization , whereas knockdowns that resulted in increased Wolbachia were actin and tubulin-related . Second , Wolbachia levels may be sensitive to disturbances in membrane dynamics and trafficking . Specifically , knockdown of SNARE components , endosomal , lysosomal and ESCRT components decreased Wolbachia , whereas knockdown of components of COPI , endosome recycling , and several SNAP receptors increased Wolbachia levels . Third , disruptions in several cell cycle-related components decreased Wolbachia levels , while Wolbachia levels increased upon disruption of cytokinesis , the separase complex and the Anaphase Promoting Complex . Fourth , the knockdown of components related to RNA helicases and the exon junction complex decreased Wolbachia , while disruption of many spliceosome components increased Wolbachia . Fifth , epigenetic changes influenced Wolbachia levels: knockdown of members involved in heterochromatin silencing , Sin3 complex and coREST decreased Wolbachia levels , whereas knocking down members of the BRAHMA complex resulted in increased Wolbachia levels . Finally , Wolbachia levels were sensitive to changes in host transcription . We observed that disruption of components in the mediator complex and regulators of transcription from Polymerase II promoters decreased Wolbachia , whereas knockdown of the BRD4pTEFb complex involved in transcriptional pausing and other transcriptional elongation factors resulted in increased Wolbachia levels . Together , this manual curation revealed that this whole genome screen yielded host genes that suppress or enhance Wolbachia levels and that these primary hits could be classified into distinct gene categories . Further GO term enrichment analysis using the online tool Panther ( http://www . pantherdb . org/ ) suggested that the 329 genes resulting in Wolbachia increases formed a robust dataset as many of the enriched terms overlapped with our manual curation ( S7 Fig ) . In contrast , there was a lack of enrichment for the 788 Wolbachia-decreasing genes even though manual curation had sorted many of these genes into categories . For this reason , further analysis focused on the 329 host genes that increased Wolbachia when knocked down by RNAi . To assess whether specific host networks were enriched within the 329 host genes identified as potential suppressors of Wolbachia we used two bioinformatic tools namely the Kyoto Encyclopedia of Genes and Genomes ( KEGG ) , and the protein complex enrichment analysis tool ( COMPLEAT ) with criteria for a network restricted to complexes with 3 or more components ( p<0 . 05 ) [51] . This analysis revealed enrichment of several host networks among the 329 genes whose knockdown resulted in Wolbachia increases including a striking 67 . 5% of the core cytoplasmic ribosome ( 56/83 expressed ribosomal proteins ) and 31 . 3% of all translation initiation components ( 10/32 expressed proteins ) as well as 70 . 1% the core proteasome ( 24/34 expressed proteins ) ( Fig 5A , and S8 Fig ) . These findings strongly suggested that perturbations in host translation components could alter Wolbachia levels . For both networks , the majority of components did not significantly affect cell proliferation within the duration of the RNAi screen assay ( circles ) , though some did have a negative impact ( robust Z<-1 ) ( square ) ( Fig 5A , see Fig 3D ) . Importantly , these data show that Wolbachia level fluctuations are independent of host cell proliferation changes because Wolbachia levels increased in RNAi knockdowns of network components regardless of the presence or absence of cell proliferation changes ( Fig 5A , S9 Fig ) . We chose to validate and characterize the novel Wolbachia-host translation interaction identified in the whole genome RNAi screen . We validated the influence of the ribosome , translation initiation complex , and proteasome on Wolbachia levels by knocking down representative members of each network using RNAi knockdown in JW18 cells ( Fig 5B ) . Each gene was validated using two different dsRNA amplicons that were designed to target different parts of the gene . Effects on Wolbachia levels were assessed quantitatively by DNA qPCR measuring the number of Wolbachia genomes ( wspB DNA copies ) relative to the number of host cell nuclei ( RpL11 DNA copies ) . Network validation is represented relative to untreated JW18 control cells ( No RNAi ) and the positive control RpL40 RNAi knockdown is included for reference . For the translation initiation network , we selected eIF-4a , eIF-2 subunit beta , eIF-3c , eIF-3i , and eIF-3ga . All ribosome components’ RNAi knockdown significantly increased Wolbachia levels by 5-fold or more ( Fig 5B ) . For the ribosomal network , we selected RpL10 , RpL36 , RpLP1 , RpS4 , RpLP2 , RpS3 and RpS26 for validation and each RNAi knockdown resulted in a significant increase of nearly 10-fold or higher Wolbachia levels relative to untreated JW18 control cells ( Fig 5B ) . We also validated the proteasome network using RNAi knockdown of three selected genes ( Rpn11 , Rpt2 , Rpn2 ) which resulted in significant Wolbachia increases ( S10 Fig ) . To summarize , we were able to validate that RNAi knockdown of ribosomal , translation initiation , and proteasomal networks leads to striking increases in Wolbachia levels in JW18 cells . To characterize the changes in Wolbachia levels in the JW18 cell line when ribosome or proteasome ( S10 Fig ) components are perturbed by RNAi , we visually classified the level of Wolbachia infection in cells using the Wolbachia-detecting 23s rRNA FISH probe combined with DAPI staining and the GFP-Jupiter transgene labelling microtubules to identify the cells . Each cell was classified according to its Wolbachia infection into the following categories: uninfected ( no Wolbachia ) , low ( 1–10 Wolbachia ) , medium ( 11–30 Wolbachia ) , and high ( >30 Wolbachia ) infection . Similar to Fig 1C , in a control LacZ knockdown JW18 control population 14% of the total number of cells were infected . In contrast , RNAi knockdown of the ribosome component RpS3 resulted in an overall dramatic increase in the total number of infected cells ( 73% ) ( Fig 5C ) . Of the infected cells in the control population , 73% had a low level of infection whereas 13 . 5% had a medium level infection and 13 . 5% had a high level of infection ( Fig 5D ) . A comparison of the extent of infection revealed a 1 . 6-fold increase in medium and highly infected cells after knockdown of network components compared to the control ( Fig 5D ) . Similar results were obtained in proteasome RNAi knockdown cells showing an increase in Wolbachia-infected cells to 87% ( S10 Fig ) . Together , our results show an increase in the total number of infected cells after ribosomal network knockdown and within this population a relative increase of medium to high infected cells , however the majority ( 57% ) of cells maintained a low level of infection . Next , we tested whether these networks could influence Wolbachia in the fly ( Fig 6 and S10 Fig ) . In Drosophila , Wolbachia are found abundantly in the ovary . To test the effect of perturbing the ribosome , females from a Wolbachia-infected stock were crossed to available ribosomal mutant alleles for RpL27A and RpS3 at 25°C . Then , the Wolbachia infection level in the ovaries of 5 day-old Wolbachia-infected heterozygous mutant and wild-type siblings were compared by RNA FISH for Wolbachia 23s rRNA . We observed dramatic increases in Wolbachia levels in the ribosomal mutants compared to the control sibling ovaries at early stages of oogenesis in the germarium as well as in maturing egg chambers ( Fig 6A and 6C ) . Quantification of the integrated density of the 23s rRNA Wolbachia FISH probe in Z-stack projections of germaria for ribosomal mutants confirmed a 2 . 94-fold ( RpL27A ) and 3-fold ( RpS3 ) increase in the mutant compared to control siblings ( Fig 6B ) ( Non-parametric Mann Whitney , RpL27A and RpS3 p<0 . 0001 ) . Further , quantification of stage 10 egg chambers revealed a 1 . 6-fold ( RpL27A ) and 1 . 27-fold ( RpS3 ) increase compared to their respective control siblings ( Fig 6D ) ( Non-parametric Mann Whitney , RpL27A p = 0 . 0002 , RpS3 p = 0 . 0089 ) . The fecundity of both ribosomal mutant lines was assessed by counting eggs laid per female as well as assessing the embryo viability . We found no significant difference between ribosomal mutant and control Wolbachia-infected siblings nor between Wolbachia-infected and uninfected flies , suggesting that the rate of oogenesis and viability of offspring are not affected by reducing the levels of ribosomal proteins nor by the level of Wolbachia infection ( S11 Fig ) . In conclusion , these results demonstrate that Wolbachia levels are sensitive to changes in the host ribosomal network in both early and late stages of Drosophila oogenesis and that under the conditions tested Wolbachia-infection does not impact fecundity of the animals . Similar results were obtained in proteasomal subunit Prosβ6 ( DTS5 ) mutant flies ( S10 Fig ) . Apart from the Drosophila ovary , we tested the effect of ribosomal and proteasome mutations on Wolbachia levels in other tissues including larval imaginal discs , adult male testes , and in the whole fly . To do this we processed RpS3 mutant and control larval imaginal discs for Wolbachia RNA FISH visualization ( S12 Fig ) . We found significantly increased levels of Wolbachia in haltere- , wing- , leg- discs ( 2 . 23-fold ( p<0 . 0001 ) , 1 . 15-fold ( p = 0 . 0229 ) , and 1 . 74-fold ( p<0 . 0001 ) Non-parametric Mann Whitney ) respectively . Similar results were obtained in proteasomal mutant ( DTS5 ) flies ( S12 Fig ) showing increased Wolbachia in haltere- , wing- , and leg- discs ( 2 . 55-fold ( p<0 . 0001 ) , 1 . 91-fold ( p = 0 . 0005 ) , 2 . 0-fold ( p<0 . 0001 ) Non-parametric Mann Whitney ) as well as in the larval brain ( 2 . 26-fold ( p = 0 . 0003 ) . These data suggest that increases in Wolbachia levels in RpS3 and Prosβ6 ( DTS5 ) mutants occur early in development and in a variety of tissue types ( S12 Fig ) . In addition , we found significantly increased Wolbachia levels ( 2 . 43-fold ( p = 0 . 0360 ) in the hub of adult RpL27A mutant testes compared to control siblings as well as a significant 2 . 8-fold increase in Wolbachia in proteasomal DTS5 mutant testes compared to sibling controls ( p = 0 . 0093 ) ( Non-parametric Mann Whitney ) ( S13 Fig ) . Finally , we assessed the Wolbachia level increase in whole flies using DNA qPCR and found increased Wolbachia levels in RpS3 mutants for males ( 1 . 22-fold ) and females ( 1 . 56-fold ) and RpL27A mutant females ( 2 . 74-fold ) compared to control siblings ( S14 Fig ) . Together these data suggest that Wolbachia level increases in ribosomal and proteasomal mutants occur in a wide range of tissue types and is not sex specific . Next , we asked whether a direct relationship exists between Wolbachia and host translation . To do this we asked whether chemical inhibition of host translation by cycloheximide would alter Wolbachia levels in host Drosophila . Wolbachia-infected D . melanogaster were fed cycloheximide-containing food or control food for 7 days prior to genomic DNA extraction of whole flies . We tested the Wolbachia-levels in individual whole flies using DNA qPCR and found increased Wolbachia levels in flies fed on cycloheximide compared to control flies ( Fig 6E ) . This suggested that Wolbachia levels are sensitive to host translation and that perturbation of host translation leads to increased Wolbachia levels . Having observed that Wolbachia levels are sensitive to host translation , we wanted to observe the relationship between Wolbachia and host translation levels in an unperturbed manner in Drosophila JW18 cells and in the fly . To correlate levels of host translation with levels of Wolbachia , we combined Wolbachia RNA FISH detection with a visual fluorescent ‘click’ chemistry-based method to assess global protein synthesis levels in host cells ( Fig 7 ) . This assay is based on a sensitive , non-radioactive method that utilizes ‘click’ chemistry to detect nascent protein synthesis in cells ( Fig 7A ) [52] . Detection of protein synthesis was based on the incorporation of a specialized alkyne-modified methionine homopropargylglycine ( HPG ) or alkyne-modified puromycin ( OPPuro ) ( S13A–S13C Fig ) into newly synthesized proteins in JW18 cells ( Fig 7 ) or Drosophila testes ( S13 Fig ) respectively . Labelled proteins were detected using a chemo-selective ligation or “click” reaction between the alkyne modified proteins and an azide-containing fluorescent dye which was added . This resulted in a fluorescent readout within each host cell correlating to the level of protein synthesis . Note that we assumed the majority of protein synthesis detected in this assay was host-related , however HPG can also be incorporated during bacterial protein synthesis , thus Wolbachia translation will have contributed to the overall fluorescent readout . We further processed the samples to detect Wolbachia by RNA FISH and then imaged cells using confocal microscopy ( Fig 7A and 7B ) . Quantification of this fluorescent readout of protein synthesis in the JW18 population and Wolbachia-free JW18DOX revealed that the median translation level in the Wolbachia-infected JW18 cell line was reduced 23 . 6% compared to Wolbachia-free JW18DOX ( Non-parametric Mann Whitney , p<0 . 0001 ) ( Fig 7C ) . This observation is consistent with previous observations of a global reduction in host translation [53] and translation machinery [49 , 53 , 54] in the presence of Wolbachia infection . To extend this observation , we binned cells into categories based on the level of Wolbachia infection and found a decreasing trend in the level of translation in cells as Wolbachia infection level increased ( Fig 7D ) . The median translation level decreased by 43% when comparing cells that did not contain Wolbachia to cells containing a medium-high level of infection ( Fig 7D ) . Furthermore , we found a statistically significant negative correlation ( r = -0 . 1344 , p = 0 . 006 , Pearson’s correlation ) between translation levels and Wolbachia levels in JW18 cells ( S15 Fig ) . Together , these results show a relationship between Wolbachia infection level and host translation in JW18 cells .
The recent applications of Wolbachia as a tool to lower the transmission of vector-borne viruses necessitates a comprehensive analysis of the relationship between Wolbachia and the vector host . In particular , the observation that increasing Wolbachia density leads to stronger antiviral effects in vectors [10 , 25–27] argues for a thorough examination of how intracellular Wolbachia levels are controlled . Here we focused on understanding which host systems influence Wolbachia levels . We performed a comprehensive unbiased whole genome RNAi screen that adapted RNA FISH for a high throughput approach . Traditionally , visual cell culture-based screens that investigate host-pathogen interactions use immunofluorescent staining , luminescent readouts , or fluorescently-tagged pathogens . The lack of tools for Wolbachia such as a commercially available antibody or a fluorescently-tagged Wolbachia strain necessitated our RNA FISH approach as a visual assay . This screen confirmed the feasibility of an RNA FISH detection approach as 1117 host genes were identified that alter Wolbachia levels . This accounted for approximately 8% of all screened genes . Knock down of 329 of these genes resulted in increased Wolbachia levels whereas 788 genes resulted in decreased Wolbachia levels . In summary , the screen successfully identified a comprehensive array of host genes that influence intracellular Wolbachia levels . Here we report that Wolbachia levels are sensitive to changes in host translation . When host translation components such as the ribosome or translation initiation complex are perturbed by RNAi we observe remarkable increases in Wolbachia levels ( Fig 5 ) . In support , Wolbachia levels increase in the Drosophila ovary ( Fig 6 ) , testis hub ( S13 Fig ) , larval imaginal discs ( S12 Fig ) , and in the whole fly ( S14 Fig ) for ribosomal mutants . Furthermore , Wolbachia levels increase upon global host translation inhibition when flies were fed with cycloheximide ( Fig 6 ) . Collectively , these results provide the first evidence that Wolbachia levels are sensitive to host translation level changes and suggests that host translation might normally play an inhibitory role in regulating intracellular Wolbachia levels . In addition to the sensitivity that Wolbachia displays towards host translation levels it is possible that host translation is also directly affected by Wolbachia . Quantification of protein synthesis in individual JW18 cells compared to JW18DOX cells revealed that JW18 cells had overall significantly lower translation level compared to the Wolbachia-free JW18DOX cell line ( Fig 7 ) . When JW18 cells were classified according to Wolbachia infection level , higher Wolbachia levels in JW18 cells correlated with significantly lower levels of host translation as measured by global protein synthesis levels ( Fig 7 , S15 Fig ) . We did not observe changes in translation components at the RNA level ( S3 Fig ) , however , recent proteomics studies revealed that over 100 host proteins with roles in host translation were suppressed in the presence of Wolbachia [49 , 55] . The mechanisms underlying this remain to be determined . One possibility is that the host translation is dampened by a stress response to Wolbachia . However , our gene expression analysis did not suggest any major alterations in stress response-related genes at the RNA level in response to Wolbachia ( S16 Fig ) . Yet , several significant changes in stress response were detected at the proteome level suggesting that stress could play a role in the Wolbachia-host intracellular relationship [49 , 55 , 56] . Host translation shutdown via metabolic stress pathways is a common mechanism employed by pathogens [57] . The other possible mechanism for dampening host translation is active manipulation of host translation machinery by Wolbachia perhaps at the post-translational level as our data do not suggest changes at the transcriptional level ( S3 Fig ) . Wolbachia encodes and expresses a fully functional type IV secretion system and many potential effector proteins [58–60] . Although the majority of Wolbachia effector proteins remains to be characterized , it is possible that Wolbachia encodes effectors that can manipulate the host’s translation machinery at a post-translational level as is the case for other intracellular bacteria such as Legionella [61 , 62] . Wolbachia interaction with host translation could be important in the context of positive-strand RNA virus infection in the host . Wolbachia-mediated suppression of viral replication in hosts is well described [7 , 8 , 11 , 12 , 63] . Multiple mechanisms may underlie this observation including interference with viral entry and very early stages of viral replication [38 , 53 , 56 , 63–70] . All viruses depend on host translation machinery for replication of their genomes . One intriguing possibility is that the interaction between Wolbachia and host translation could impact viral replication [53] . Wolbachia infection inhibits positive-strand RNA viral replication at very early stages of viral replication in the virus lifecycle [53 , 63] . Thus , changes in host translation could be one mechanism by which Wolbachia infection contributes towards viral replication interference . Future work to elucidate whether this is a contributing mechanism to the Wolbachia-mediated antiviral response in a wide range of Wolbachia-host-virus relationships may provide valuable field applications for combating vector-borne viruses . Our whole genome screen yielded a diverse range of host systems and complexes that influenced Wolbachia levels . Manual curation and bioinformatic analyses such as GO term enrichment and network analysis identified host pathways such as translation initiation , ribosome , cell cycle , splicing , immune-related genes , proteasome complex , COPI vesicle coat , polarity proteins and the Brahma complex . The GO term enrichment analysis and COMPLEAT network analysis suggested that the 329 genes resulting in Wolbachia increases formed a more robust dataset than the larger 788 gene category resulting in Wolbachia decreases owing to a lack of enrichment for specific networks and processes in this category . For this reason , we focused on the host networks that increased Wolbachia in this report . Nevertheless , future follow-up analysis on genes that decreased Wolbachia levels especially the larger categories such as metabolism & transporters , cytoskeleton , cell adhesion & extracellular matrix , as well as membrane dynamics and vesicular trafficking may yield rewarding results . We already appreciate that Wolbachia relies on several aspects of these broad categories . For example , Wolbachia can alter host iron , carbohydrate and lipid metabolism [71–75] . Further , Wolbachia interacts with host cytoskeleton such as microtubules for transport and host actin [32 , 34 , 35 , 76] . Finally , Wolbachia resides within a host-derived membrane niche , as such genes identified in the membrane dynamics and vesicular trafficking category would be of interest [77 , 78] . Further investigation of these Wolbachia-decreasing categories may provide comprehensive insights into Wolbachia-host interactions . Interestingly our study also revealed that knockdown of the core proteasome leads to increases in Wolbachia levels ( Fig 4A , S8 Fig , S10 Fig ) . A previous report suggested that Wolbachia require high levels of proteolysis for optimal survival [36] . An explanation for this discrepancy might be that previous observations of decreased Wolbachia levels were based on RNAi experiments that knocked down ubiquitin-related components not the core proteasome . Ubiquitination is known to function in many diverse contexts and pathways such as autophagy , cell cycle , immune response , DNA damage response and regulation of endocytic machinery [79] . Our screen also identified several ubiquitin-related components whose knockdown resulted in decreased Wolbachia levels ( Fig 4B ) . Perhaps Wolbachia relies on the host ubiquitination system for survival in an unknown but specific context , not simply for providing amino acids as nutrients from the degradation of proteins by the proteasome . Our gene expression data ( S3 Fig ) along with recent proteomics studies [49 , 55] suggest that the host proteasome is upregulated in the presence of Wolbachia . We propose that the host proteasome plays an inhibitory role in Wolbachia level regulation . Perhaps Wolbachia levels are controlled by degradation of effector proteins in the cytosol , thereby preventing Wolbachia from utilizing the host cell in an optimal manner . The results both in the JW18 cell line as well as in the ovary ( S10 Fig ) , testis hub ( S13 Fig ) , and larval imaginal discs ( S12 Fig ) of D . melanogaster strongly suggest that the host core proteasome normally plays a restrictive role in Wolbachia-host interactions that is separate from observations of ubiquitin pathway perturbation . In summary , here we presented a whole genome screen to identify host systems that influence Wolbachia levels . Our focus was on Wolbachia sensitivity to alterations of host translation-related components such as the ribosome and translation initiation factors . We report a novel relationship between Wolbachia and host translation and suggest a restrictive role for host translation on Wolbachia levels . Future work to identify whether Wolbachia is able to actively manipulate host translation will provide valuable insight into understanding this unique host-symbiont relationship .
A stable Wolbachia-infected Drosophila cell line ( JW18 ) and a doxycycline-treated Wolbachia-free cell line ( JW18DOX ) , where Wolbachia infection of the JW18 cell line was removed by treatment with doxycycline to generate a Wolbachia-free version were kindly provided by William Sullivan at UCSC . Cell lines were maintained in Sang and Shield media ( Sigma ) supplemented with 10% heat inactivated One Shot Fetal Bovine Serum ( Life Technologies ) . The doubling time of the JW18 cells was calculated from a growth curve using the formula ( t2-t1 ) /3 . 32 x ( logn2-logn1 ) where ‘t’ is time and ‘n’ is cell number . The JW18DOX cell line was originally treated with doxycycline in late 2010 and thereafter maintained in normal culturing media without doxycycline . D . melanogaster infected with the wMel Wolbachia strain was a gift from Luis Teixeira ( Instituto Gulbenkian de Ciência ) . To generate a Wolbachia-infected double balancer line , Wolbachia-infected virgins were crossed to Sp/CyO; MKRS/TM6B males . In the next generation Wolbachia-infected +/CyO;+/TM6B female virgins were crossed to males of the original double balancer stock . In the final generation , a stock of Wolbachia-infected Sp/CyO; MKRS/Tm6B double balancers was established . Ribosomal mutant fly stocks were ordered from the Bloomington Drosophila Stock Center . The following haploinsufficient lines were used: RpS32/TM2 ( stock no . 1696 ) and RpL27A1/CyO ( stock no . 5697 ) . Males from each line were crossed at 25°C to the Wolbachia-infected double balancer line described above . Siblings from each cross were matured for 5 days before tissues were dissected and stained for Wolbachia using the 23s rRNA Wolbachia-specific FISH probe . Tissues were imaged in Z-stacks using confocal microscopy . Quantification of the integrated density of the 23s rRNA Wolbachia FISH probe in tissue Z stacks were done using Fiji Image Processing software as described in the FISH section . A dominant temperature-sensitive ( DTS ) lethal mutant for proteasome component Pros26 , known as DTS5 , was a gift from John Belote , Syracuse University . Heterozygotes die as pupae when raised at 29°C , but are viable and fertile at 25°C . The mutant contains a missense mutation in the gene encoding the β6 subunit of the 20S proteasome . Males were crossed to Wolbachia-infected female double balancers at the permissive temperature of 25°C . Hatched offspring from the cross were matured at the non-permissive 29°C for 5 days prior to dissection of the tissues . Imaging and analysis of tissues of siblings were done as described above for the ribosomal mutant crosses . For fecundity testing we performed the following crosses: In the first generation , we crossed Wolbachia-infected Sp/CyO; MKRS/TM6B double balancer virgin females to males from RpS32/TM2 , RpL27A1/CyO , and DTS5 stocks . For control fecundity experiments that were Wolbachia-free , we set up the reciprocal crosses using virgin females from RpS32/TM2 , RpL27A1/CyO , and DTS5 stocks and males from the Wolbachia-infected Sp/CyO; MKRS/TM6B double balancer stock . In the next generation we collected the following virgin females for fecundity testing: RpS32/MKRS and TM2/MKRS ( control sibling ) , RpL27A1/CyO and Sp/CyO ( control sibling ) , and DTS5/TM6B and MKRS/TM6B ( control sibling ) . For fecundity testing , all virgins were 2–4 days old . These virgins were mated with Oregon R males for one day prior to setting up the cages for fecundity testing . For testing , 3 females and 1 OregonR male were allowed to mate together and lay eggs on agar plates for 6 hours each day for 3 days . For each genotype between 6–17 ‘3x1’-matings were set up . 24 hours later eggs were counted and scored for hatching . Genomic DNA was extracted from cells or Drosophila tissues using a DNeasy Blood & Tissue Kit ( Qiagen ) following manufacturer’s instructions . To quantify the level of Wolbachia in the sample a DNA qPCR assay was performed using SYBR Green I Master 2x ( Roche ) , using a Roche LightCycler 480 machine . Primer sets included a primer set to detect wspB which is a gene encoding a Wolbachia surface antigen ( F: 5’ ACA ACA GCT ATA GGG CTG AAT TGG AA 3’ , R: 5’ TCA GGA TCC TCA CCA GTC TCC TTT AG 3’ ) , as well as a primer set to detect the Drosophila gene RpL32 ( also known as RpL49 ) ( F: 5’ CGA GGG ATA CCT GTG AGC AGC TT 3’ , R: 5’GTC ACT TCT TGT GCT GCC ATC GT 3’ ) . Wolbachia levels were normalized by the host nuclear marker for each sample . RNA was extracted and DNase-treated using a RNeasy kit ( Qiagen ) according to manufacturer’s instructions . Total RNA was reverse transcribed using an RNA to cDNA EcoDry Premix ( OligodT ) or EcoDry Premix ( Random hexamer ) ( Clontech ) according to manufacturer’s instructions . Quantitative PCR was performed on 1/200 of the RT reaction using LightCycler 480 SYBR Green I Master 2x ( Roche ) and a Roche LightCycler 480 machine . Results were normalized to the housekeeping gene Rp49 . Primer sets used to validate RNAi knockdown were designed to amplify areas outside of the dsRNA amplicon . Gene knockdown was represented relative to expression levels in LacZ dsRNA-treated cells . Genomic DNA was extracted from JW18 cells in duplicate . Samples were quantified using a Qubit fluorometer ( Thermo Fisher Scientific ) . DNA libraries were prepared using a Nextera DNA Library Prep kit ( Illumina ) according to manufacturer’s instructions . DNA libraries were sequenced on an Illumina HiSeq2500 Sequencing platform in two lanes as paired-end reads 100 cycle lanes . Total RNA was extracted in triplicate from JW18 and JW18TET cells and DNase-treated . RNA was quantified by Nanodrop and 5μg of each sample was subjected to two rounds of rRNA depletion using a Ribo-Zero rRNA Removal Magnetic kit ( Epicentre , Illumina ) or NEBNext rRNA Depletion Kit ( Human/Mouse/Rat ) ( New England BioLabs , E6310L ) . After rRNA depletion libraries were prepared according to manufacturer’s instructions using the NEBNext Ultra Directional RNA Library Prep Kit for Illumina ( New England BioLabs , E7420L ) and NEBNext Multiplex Oligos for Illumina Index Primers Set I ( Illumina , E7335 ) . After adaptor ligation , the libraries were amplified by qPCR using the KAPA Real-time amplification kit ( KAPA Biosystems ) . Finally , libraries were purified using Agencourt AMPure XP beads ( Beckman Coulter ) as described in the NEBNext Ultra Directional RNA Library Prep Kit for Illumina ( New England BioLabs , E7420L ) protocol . Quality and quantity was assessed using a Bioanalyzer ( Agilent ) and a Qubit fluorometer ( Thermo Fisher Scientific ) . Libraries were sequenced on an Illumina HiSeq2500 Sequencing platform in single read 50 cycle lanes . Differential gene expression analysis was performed from one lane of high output , single end reads 50 , Illumina HiSeq run . The experiment consisted of 3 technical replicates each for JW18 and JW18TET cells . The alignment program , Tophat ( version 2 . 0 . 9 ) ( https://ccb . jhu . edu/software/tophat/index . shtml ) was used for reads mapping with two mismatches allowed . Featurecounts ( http://bioinf . wehi . edu . au/featureCounts/ ) was used to find the read counts for annotated genomic features . For the differential gene statistical analysis , DESeq2 R/Bioconductor package in the R statistical programming environment was used ( http://www . bioconductor . org/packages/release/bioc/html/DESeq2 . html ) . We mapped short reads generated from DNA-Seq with Bowtie2 version 2 . 2 . 9 [80] . We used default parameters and mapped to combined sequences of Drosophila genome release 6 [81] and Wolbachia pipientis wMel ( [58] , GenBank accession ID AE017196 . 1 ) . We determined basal ploidy level of JW18 cells by clustering normalized DNA-Seq read densities as in [82] . In doing that , we identified different copy number segments whose normalized read densities are between zero ( no DNA content ) to the mean density ( basal ploidy level ) . Clusters of such read densities indicate the minimum ploidy . From the determined basal ploidy , we called copy numbers of JW18 cell line genome using Control-FREEC version 5 . 7 [83] at 1 kb levels . We called copy numbers in an identical way to [82] but with this exception; we performed calling twice and combined the results . Control-FREEC performs GC contents-based normalization of DNA-Seq reads . Therefore , we set the minimum expected GC contents to be 0 . 30 for robust copy number calling of the cell line genome first . Then we underwent our analysis again with the minimum expected GC contents of 0 . 25 to increase sensitivity against the bacterial genome . In our reports , we combined copy number calls from the former for JW18 cells , and from the latter for Wolbachia . In S1 Fig , we used DNA-Seq results from [83] to call copy number calls on S2R+ and Kc167 cells . We re-analyzed the original data after mapping to the release 6 genome as above . The method described in [59] was used to analyze the genotype of the Wolbachia strain in the JW18 cell strain . Briefly , fastq sequences were mapped against a “holo-genome” consisting of the Release 5 version of the D . melanogaster genome ( Ensembl Genomes Release 24 , Drosophila_melanogaster . BDGP5 . 24 . dna . toplevel . fa ) and the Wolbachia wMel reference genome ( Ensembl Genomes Release 24 , Wolbachia_ endosymbiont_of_drosophila_melanogaster . GCA_000008025 . 1 . 24 ) [84 , 85] . Holo-genome reference mapping was performed using bwa mem v0 . 7 . 5a with default parameters in paired-end mode . Mapped reads for all runs from the same sample were merged , sorted and converted to BAM format using samtools v0 . 1 . 19 [86] . BAM files were then used to create BCF and fastq consensus sequence files using samtools mpileup v0 . 1 . 19 ( options -d 100000 ) . Fastq consensus sequence files were converted to fasta using seqtk v1 . 0-r76-dirty and concatenated with consensus sequences of Wolbachia-type strains from [27] . Maximum-likelihood phylogenetic analysis on resulting multiple alignments was performed using raxmlHPC-PTHREADS v8 . 1 . 16 ( options -T 12 -f a -x 12345 -p 12345 -N 100 -m GTRGAMMA ) [87] . Copy number variants were detected by visual inspection of read depth across the wMel genome . GO enrichment analysis were performed using PANTHER Version 12 . 0 ( release 2017-07-10 ) ( http://www . pantherdb . org/ ) . The entire set of screened genes was used as the experimental background . Protein complex enrichment analysis was performed using COMPLEAT ( http://www . flyrnai . org/compleat/ ) . As the experimental background we used the entire set of screened genes . Complex size was limited to ≥3 , with a p value filter of p<0 . 05 . Outdated amplicons from the Drosophila RNAi Screening Center ( DRSC ) whole genome library 2 . 0 were identified using Updated Targets of RNAi Reagents ( UP-TORR ) ( http://www . flyrnai . org/up-torr/ ) . For amplicons that could be transferred to Release 6 , we followed the dsRNA in vitro synthesis protocol as described by the DRSC . The DNA templates were generated by PCR on genomic DNA extracted from the JW18 cells , genomic DNA from wild-type flies or pBlueScript SK ( + ) plasmid DNA ( in the case of LacZ ) . All gene specific primer sequences were selected by the DRSC and the T7 promoter sequence ( TAATACGACTCACTATAGGG ) was added to the 5’ ends of all primer pairs . Gradient PCR reactions were performed with Choice Taq Mastermix ( Denville Scientific Inc . ) using 5ng of genomic DNA , 0 . 1ng of plasmid DNA , or 1:3 diluted PCR template DNA . PCR products were verified by electrophoresis on a 0 . 7% ( w/v ) agarose gel with the 1kb PLUS ladder ( Invitrogen ) and only products with a clear single band were selected for IVT . IVT was performed according to manufacturer’s instructions for the MEGAscript T7 Transcription Kit ( Ambion ) using 8μl of amplified T7-flanked PCR product per reaction . dsRNA products were DNase-treated using Turbo DNase ( Ambion ) and purified with Qiagen RNeasy Mini spin columns ( Qiagen ) according to manufacturer’s protocols . Quality of purified dsRNA was assessed by electrophoresis on a 0 . 7% agarose gel , and concentration was determined by Nanodrop ( Thermo Fisher Scientific ) [40] . RNAi in JW18 cells was done using a bathing method described by the DRSC [40] . dsRNA aliquots were prepared in serum-free Sang and Shield media ( Sigma ) . dsRNA was added to wells to yield a final concentration of 25nM . For the whole genome screen the pre-arrayed DRSC Drosophila Whole Genome Library Version 2 . 0 was used . For each RNAi experiment , sub-confluent JW18 cells were scraped , pelleted ( 1000rpm for 5–15 minutes ) , and re-suspended in serum-free Sang and Shield media ( Sigma ) to seed 40 000 cells in 384 well format . Cells and dsRNA were incubated together at room temperature for 30 minutes in serum-free conditions . Thereafter Sang and Shield media ( Sigma ) supplemented with 10% heat inactivated One Shot Fetal Bovine Serum ( Life Technologies ) was added to each well and incubated at 25°C for 5 days before analysis . Cells were plated in Poly-L-lysine-coated chambered cover-glass wells ( Thermo Scientific ) and allowed to settle . Medium was aspirated and cells were washed with 1xPBS before fixing with 4% PFA in 1xPBS ( Electron Microscopy Sciences ) . Cells were washed twice in 1xPBS followed by two washes in 100% methanol ( Fisher ) before finally adding 100% methanol to each chamber and sealing it with Parafilm M Film ( Sigma ) for storage at -20°C overnight or up to 1 month . Samples were rehydrated using the following washes: MeOH: PBT ( 1xPBS , 0 . 1% Tween-20 ) ( 3:1 ) , MeOH:PBS ( 1:1 ) , MeOH: PBS ( 1:3 ) , and a final wash in 1xPBS . Samples were then post-fixed for 10 minutes in 4% PFA at room temperature . In a pre-hybridization step , samples were incubated in 10% deionized formamide and 2x SCC for 10 minutes at room temperature . Pre-hybridization buffer was then removed , and a hybridization solution containing a Wolbachia-specific FISH probe was added and incubated overnight at 37°C . For each sample , the volume of hybridization buffer added was dependent on the type of well used , but enough should be added to cover the sample . Typically , 60μl of hybridization buffer comprised 10% Hi-Di deionized formamide ( Applied Biosystems Life Technologies ) , 1μl of competitor ( 5mg ml-1 E . coli tRNA ( Sigma ) and 5 mg ml-1 salmon sperm ssDNA ( Ambion ) ) , 10mM vanadyl ribonucleoside complex ( New England Biolabs ) , 2xSSC ( Ambion ) , 50μg nuclease-free BSA ( Sigma ) , 10ng Wolbachia-specific FISH probe , made up to 60μl with DEPC-treated water . The Wolbachia specific probe was designed to the Wolbachia 23s rRNA and labeled with Quasar670 ( Stellaris ) . After overnight hybridization samples were washed twice in pre-warmed pre-hybridization buffer for 15 minutes at 37°C . Followed by two washes in 1x PBS for 30 minutes each . Finally , samples were stained for 5 minutes in 1:500 DAPI:1xPBS followed by two washes in 1x PBS . Unless otherwise stated , samples were imaged as Z-stacks on a Zeiss LSM 780 confocal at 63x . For Wolbachia detection in dissected Drosophila ovaries and testes , the same protocol was followed from fixation onwards . Quantification of Wolbachia levels based on the intensity of the Wolbachia 23s rRNA probe for tissues was done in Fiji Image processing software . Z-stacks capturing entire tissues were projected as ‘sum slices’ . Each tissue was manually outlined using the Freehand tool . The measurements tool was set to capture the integrated density within the outlined tissue of the FISH probe channel stack as well as provide an area measurement of the outlined tissue . We normalized the integrated density reading for each by its total area . Large-scale RNAi screening was done using the DRSC Drosophila Whole Genome Library Version 2 . 0 that was seeded in Corning clear bottom , black 384 well plates with 0 . 25μg dsRNA pre-arrayed per well . This concentration of dsRNA was appropriate for the bathing method of RNAi [40] . JW18 cells were re-suspended in serum-free media at 4x106 cells/ml and an automated Matrix Wellmate dispenser ( Thermo Fisher Scientific ) was used to dispense 40 000 cells into each well of the 384 well plates in a sterile tissue culture hood . The cells were incubated with the dsRNA in serum-free media for 30 minutes before automatic dispensing of Sang and Shield media ( Sigma ) supplemented with 10% heat inactivated One Shot Fetal Bovine Serum ( Life Technologies ) into each well . Plates were incubated at 25°C for 5 days in a humidity chamber . After 5 days plates were drained followed by automated dispensing of 4% paraformaldehyde ( Electron Microscopy Sciences ) and incubation at room temperature for 10 minutes and automatically aspirated thereafter . An automated BioTek EL406 liquid handler ( BioTek ) was used throughout the protocol for all aspiration steps and a Matrix Wellmate ( Thermo Fisher Scientific ) was used for all dispensing steps . Next , plates were washed once with 1xPBS followed by three washes with 100% methanol ( Fisher ) , sealed with Parafilm M film , and stored overnight at -20°C ( or up to 1 month ) . All subsequent rehydration , post-fixation , pre-hybridization and hybridization steps of the RNA FISH protocol described above were carried out in an automated manner . After overnight hybridization at 37°C the plates were washed twice with pre-warmed pre-hybridization buffer followed by incubation at room temperature for 30 minutes with 1xPBS/DAPI . Finally , plates were washed once with 1xPBS and 40μl 1xPBS was dispensed into all wells and plates were sealed with aluminum foil and stored at 4°C . Plates were imaged with a 20x objective lens using an Arrayscan VTI Microscope ( Cellomics ) coupled with the automated image analysis software HCS Studio Cellomics Scan Version 6 . 6 . 0 ( Thermo Fisher Scientific ) . Image acquisition involved identification of DAPI stained cell nuclei as primary objects , followed by application of a ring mask around the primary objects to identify Wolbachia associated with each cell as secondary objects . Segmentation of the objects was optimized to exclude any areas containing cell clumps . For each well , 1500 primary objects ( DAPI cell nuclei ) were acquired . RNAi screen primary data analysis and criteria for hit selection is summarized in S4 Fig . Protein synthesis levels in JW18 cells were detected using a Click-iT HPG Alexa Fluor 594 Protein Synthesis Assay Kit ( Molecular probes , C10429 ) . Regular Sang and Shield media ( containing methionine ) ( Sigma ) was removed from JW18 cells . Cells were washed once in 1xPBS . A working solution of Click-iT HPG was prepared according to manufacturer’s instructions using methionine-free Grace’s Insect Medium ( Thermo Scientific ) . Cells were incubated for 30 minutes in 50μM Click-iT HPG working solution . After incubation , cells were washed once in 1xPBS followed by fixation in 5% formaldehyde . To combine the protocol with RNA FISH detection of Wolbachia in the cells , we next proceeded to wash the cells twice in methanol followed by storing the sample in 100% methanol at -20°C overnight . From this point , we followed the RNA FISH protocol described in the FISH section . After FISH hybridization and post-hybridization washes , we incubated cells with 0 . 5% Triton X-100 in 1xPBS for 20 minutes at room temperature . Cells were washed twice with 3% BSA in 1x PBS . The Click-iT Reaction Cocktail was added to the samples for 30 minutes at room temperature protected from light . Thereafter , samples were washed once with Click-iT Reaction Rinse Buffer before staining with 1x HCS NuclearMask Blue Stain working solution as per manufacturer’s instructions . Samples were imaged on confocal at 63x magnification . Controls included in the assay were as follows: incubation of cells with cycloheximide at a final concentration of 100μg/ml for 1 hour prior to the start of the experiment as well as during the 30-minute incubation with HPG; and a negative control sample that was not incubated with HPG . Quantification of the Click-iT fluorescent intensity signal within each cell was done in a similar manner as described for FISH signal in egg chambers in the FISH section . Briefly , projected Z-stacks were manually outlined in Fiji and integrated density and area measurements were captured for each cell using the measurements tool . This allowed for a normalized integrated density measurement for individual cells . These data could then be paired with the Wolbachia level within individual cells as measured by RNA FISH . Protein synthesis in the Drosophila testis was detected by the Click-iT Plus OPP Alexa Fluor 594 protein synthesis assay kit ( Molecular Probes ) as previously described [88] . Samples were incubated for 30 minutes in 1:400 Click-iT OPP reagent in fresh Shields and Sang M3 Insect medium ( Sigma ) . | Insects such as mosquitos act as vectors to spread devastating human diseases such as Dengue , West Nile , and Zika . It is critical to develop control strategies to prevent the transmission of these diseases to human populations . A novel strategy takes advantage of an endosymbiotic bacterium Wolbachia pipientis . The presence of this bacterium in insect vectors prevents successful transmission of RNA viruses . The degree to which viruses are blocked by Wolbachia is dependent on the levels of the bacteria present in the host such that higher Wolbachia levels induce a stronger antiviral effect . In order to use Wolbachia as a tool against vector-borne virus transmission a better understanding of host influences on Wolbachia levels is needed . Here we performed a genome-wide RNAi screen in a model host system Drosophila melanogaster infected with Wolbachia to identify host systems that affect Wolbachia levels . We found that host translation can influence Wolbachia levels in the host . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"rna",
"interference",
"pathology",
"and",
"laboratory",
"medicine",
"cell",
"processes",
"genomic",
"library",
"screening",
"animals",
"wolbachia",
"invertebrate",
"genomics",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"epigenetics",
"molecular",
"biology",
"techniques",
"bacteria",
"drosophila",
"research",
"and",
"analysis",
"methods",
"cell",
"proliferation",
"genetic",
"interference",
"animal",
"studies",
"gene",
"expression",
"pathogenesis",
"molecular",
"biology",
"insects",
"animal",
"genomics",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"arthropoda",
"biochemistry",
"rna",
"eukaryota",
"cell",
"biology",
"nucleic",
"acids",
"host-pathogen",
"interactions",
"genetic",
"screens",
"library",
"screening",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"organisms"
]
| 2018 | Whole genome screen reveals a novel relationship between Wolbachia levels and Drosophila host translation |
Chronic liver infection by hepatitis C virus ( HCV ) is a major public health concern . Despite partly successful treatment options , several aspects of intrahepatic HCV infection dynamics are still poorly understood , including the preferred mode of viral propagation , as well as the proportion of infected hepatocytes . Answers to these questions have important implications for the development of therapeutic interventions . In this study , we present methods to analyze the spatial distribution of infected hepatocytes obtained by single cell laser capture microdissection from liver biopsy samples of patients chronically infected with HCV . By characterizing the internal structure of clusters of infected cells , we are able to evaluate hypotheses about intrahepatic infection dynamics . We found that individual clusters on biopsy samples range in size from infected cells . In addition , the HCV RNA content in a cluster declines from the cell that presumably founded the cluster to cells at the maximal cluster extension . These observations support the idea that HCV infection in the liver is seeded randomly ( e . g . from the blood ) and then spreads locally . Assuming that the amount of intracellular HCV RNA is a proxy for how long a cell has been infected , we estimate based on models of intracellular HCV RNA replication and accumulation that cells in clusters have been infected on average for less than a week . Further , we do not find a relationship between the cluster size and the estimated cluster expansion time . Our method represents a novel approach to make inferences about infection dynamics in solid tissues from static spatial data .
Around 170 million people worldwide are chronically infected with hepatitis C virus ( HCV ) , representing a major public health problem [1] . Chronic HCV infection can lead to liver cirrhosis , hepatocellular carcinoma and liver failure , and it represents the leading cause for liver transplantation in Western countries [2] . Despite successful treatment options using mostly type I interferon-α ( IFN-α ) or newer direct-acting antiviral agents , several aspects of HCV infection dynamics are still unknown . For example , does the virus propagate preferentially by cell-to-cell transmission or via diffusion of viral particles ? Do innate immune responses interfere with the spatial spread of the infection ? And what fraction of hepatocytes are infected during the acute and chronic stages of infection ? To answer these questions , a better understanding of the in vivo infection process is needed . As appropriate animal models for HCV infection are lacking , inferring in vivo infection dynamics from clinical data has relied on mathematical models that describe the interaction of hepatocytes with viral particles [3]–[8] . Mathematical modeling of viral load dynamics in combination with data on treatment with IFN-α and direct-acting antivirals has helped to reveal and quantify aspects of the infection process , such as the half-life of viral particles and the loss rate of infected hepatocytes under treatment [3] , [6] , [9] , [10] . In addition , models have quantified the necessary treatment efficacy to clear the virus [5] , [11] . Existing models have been fit to HCV RNA levels measured in the serum of patients . Measurements of viral levels in the liver and in particular of HCV RNA levels within cells of the liver have generally been lacking . Advances in techniques , such as two-photon microscopy [12] , [13] and laser capture microdissection [14] , now allow one to visualize and analyze HCV infection in the liver at the cellular level . Using single cell laser capture microdissection ( scLCM ) , it is possible to determine the HCV RNA content in single hepatocytes from liver biopsies of HCV infected patients as well as the spatial relationships among infected cells [15] . Analyzing regular grids of hepatocytes we found that infected hepatocytes tend to occur in clusters [15] , in agreement with other studies reporting a focal distribution of HCV RNA in infected liver tissue [12] , [14] , [16] , [17] . However , patients differ in their individual viral load , as well as in the frequency of hepatocytes infected . To extend our previous observation of a spatial heterogeneous distribution of infected hepatocytes [15] , we now develop statistical methods to characterize properties of clusters of infected hepatocytes in more detail , e . g . in terms of cluster size and intracellular HCV RNA levels . Analyzing data from 4 chronically infected patients [15] , we find that clusters of infected cells comprise between 4 and 50 cells in the plane of the liver biopsies . These sizes are comparable to the range of cluster sizes observed in in vitro experiments on Huh-7 . 5 cells under conditions only allowing cell-to-cell transmission [18] . In addition , we find that the level of intracellular viral RNA declines in infected cells at increasing distance from the cell that presumably founded the cluster [12] , [19] . Using intracellular HCV RNA content as a proxy for the time since infection in a given cell , this suggests that cells closer to the founder cell of the cluster have been infected for a longer time than those in the periphery . Both of these observations suggest that viral infection once seeded spreads locally , supporting cell-to-cell transmission [12] , [20] or viral release from an infected cell with rapid binding to and infection of neighboring cells . We then used mathematical models to describe intracellular viral replication and accumulation of viral RNA . Applying these models to interpret the data , we do not find a relationship between the observed cluster size and the estimated time that the cluster has been expanding , suggesting that individual cellular factors might influence cluster growth . We also estimate that the cells in the detected clusters have been infected on average for less than a week . This finding is consistent with previous estimates of the mean lifetime of HCV infected cells [21] , [22] . Overall , our study presents a set of novel methods to infer in vivo viral dynamics of chronic HCV infection in the human liver based on liver biopsy samples .
In a previous analysis of two-dimensional grids of hepatocytes analyzed by scLCM , we obtained evidence for clustering of HCV infected cells in the liver [15] . Determining the size of individual clusters visually based on the actual grid data is difficult as we are only analyzing a small fraction of tissue . Infected cells at the edge of the sampling area might be part of a larger cluster that extends outside the sampling region . In this study , we apply enhanced methods of spatial statistics to the data in order to ( 1 ) estimate the size of clusters accounting for edge effects due to the limited sampling area , and ( 2 ) characterize the structure of clusters of infected cells in more detail . If hepatocytes in the liver were infected completely at random , for example due to rapid seeding from the blood , we would expect homogeneous infection and no clusters . Since we observe clusters of infection [15] , we make the next most parsimonious assumption that viral spread in vivo is a combination of random spatially scattered infection of some cells that seed the cluster ( possibly from virus in the blood ) followed by predominantly random local spread from these cells . We assume that seeding of the cluster centers follows a Poisson process , with the mean number of clusters per unit area equal to , and the number of cells in each cluster also following a Poisson process , with the mean number of cells in each cluster equal to . This compound Poisson spatial distribution is called a Matérn cluster process [23] , [24] . A Matérn cluster process assumes that the units of a cluster are distributed within a radial disc with domain radius ( Figure 1A ) . Assuming this regular cluster structure allows us to account for edge effects due to the small sampling area , i . e . , that only parts of a cluster were sampled on the grid [25] . To determine individual cluster sizes , we fit a Matérn cluster process to each of the two-dimensional grids of hepatocytes analyzed by scLCM to estimate the domain radius . We proceed by first randomly distributing the HCV RNA content of each cell over the space occupied by the cell in the grid and then fitting a Matérn cluster process to these spatial point patterns ( Figure 1B and Text S1 ) , rather than fitting to the spatial distribution of infected hepatocytes . This is done as ( i ) a Matérn cluster process assumes continuous space , and ( ii ) because we are also interested in the internal structure of the cluster ( see below ) . We note that the random distribution of HCV molecules within a cell is an artificial construct to allow us to estimate the radius of the cluster , when in fact HCV replication likely occurs in localized structures in the cell ( reviewed in [26] ) . To make sure that the particular distribution of HCV RNA inside the infected cell does not affect the results , we repeat the procedure of random allocation of HCV RNA inside the cell 10 , 000 times , thus obtaining that many bootstrap replicates of our data set . With this transformation , the algorithm detects regions with similar densities of HCV RNA . It is possible that cells with low amounts of HCV RNA that are adjacent to cells with higher levels of HCV RNA would not be counted as part of a cluster due to the difference in the density of viral RNA . To account for this potential artifact of the algorithm , we estimated the domain radius of a cluster based on the minimum level of HCV RNA in all cells assumed to belong to the cluster ( Figure 1B and Materials & Methods for details ) . Since we do not know a priori which cells belong to the cluster , our estimate is done iteratively . We start with the cell with the highest HCV RNA content and estimate the cluster radius for the obtained point pattern ( Figure 1B ) . In succession , we assume that the cell with the next highest content would be the cell with the minimal level of HCV RNA in the cluster , we reduce higher levels of HCV RNA to this level ( Figure 1B ) and again fit a Matérn Cluster process to the new point pattern . We stop when inclusion of the next cell would result in an apparent homogeneous distribution of infected cells as determined by the Quadrant-Count method ( see Supporting Information , Text S1 ) . At this point , we are not detecting individual clusters , because we included so many cells that there is only one large homogeneous cluster . Detection of the maximal cluster size , and thus the average number of clusters per sampled grid , is independent of the assumed founder cell as we are identifying regions of cells with similar densities of HCV RNA . Our algorithm allows all infected cells to be counted with equivalent weights , and because we estimate the cluster radius in iterative steps , we characterize the structure of each cluster analogous to the ring-structure of a tree trunk , where successive rings have similar spatial density of HCV RNA ( Figure 1 ) . Note that we perform this iterative procedure with all 10 , 000 bootstrap replicates . We applied the above procedure to each grid of each subject individually . In Figure 2 , we show , for subject 3 , the estimates of the cluster radius , , and the corresponding expected number of cluster centers per unit area , , as a function of the minimal HCV RNA content of cells in the clusters at each iteration of our algorithm ( from right-to-left ) . The maximal cluster extension , , is found when the mean number of cluster centers , , starts to decrease from its more or less constant value ( Figure 2 , reading right-to-left ) . A decreasing indicates fewer and larger clusters , a more homogeneous distribution , and in the limit indicates that all cells on the grid belong to one single cluster . Since we previously determined that infected cells are clustered [15] , such inference of a homogeneous distribution is unreasonable . For each iteration we analyze , in each of our bootstrap replicates , if the resulting spatial distribution still indicates clustering or not . The iterative process is stopped when we start finding a homogeneous spatial distribution , that is when less than 95% of the replicates show evidence of clustering . On grid 1 of subject 3 , the algorithm detects one large cluster , while on the other two grids , the existence of several smaller clusters of infected cells is predicted , with the estimated mean number of clusters , , on these grids being times larger than on grid 1 . The corresponding figures for all other subjects are shown in the Supporting Information ( Figures S1 and S2–S4 ) . The algorithm provides estimates of the cluster radius and the number of clusters in each grid , however it does not explicitly show where each cluster is located . Based on the results for and for subject 3 , in the right panel of Figure 2 , we infer a possible spatial distribution of clusters of infected cells on each grid for this subject as shown . Overall , with this method we find that our sections of liver biopsies contain between ∼1 and ∼45 clusters of infected cells , which have radii between 12 . 9 µm and 103 µm ( Table 1 ) . To determine the size of a cluster in terms of numbers of cells , we transform the cluster radius back into an actual number of hepatocytes , . The radius was estimated searching for areas with similar densities of HCV RNA and assuming a circular structure of the cluster , but hepatocytes themselves are better described by a square geometry ( Figure 1B ) . Thus for a given radius we estimate the number of cells inside a circle of radius and inside a square with sides of length ( where was estimated based on our algorithm above ) . Note that by definition most hepatocytes in a cluster are infected , however some uninfected cells could also be counted in a cluster . This is because clusters are defined as a circle around the seeding infected cell , and uninfected cells could be in close proximity to infected cells and by chance were not ( yet ) infected ( see Figure 1B , last step ) . The minimum number of cells belonging to a cluster is determined by , where the area of the cluster is given by a disc with radius , and denotes the area of a hepatocyte . The maximum number of cells in this cluster is estimated by assuming the area of the cluster is given by a square with edge length , hence , . For example , the estimated number of hepatocytes in a cluster with a maximal cluster extension of as determined on grid 2 of subject 3 ( Figure 2 ) ranges between and cells . We found that clusters comprise between 4–50 infected hepatocytes , not considering the two extreme cases ( grid 1 of subject 1 and grid 1 of subject 3 ) that have to be taken with care ( Table 1 ) . For the maximal cluster size of cells estimated for grid 1 of subject 3 ( Table 1 ) , the estimate of is close to the reliability threshold for ( see Materials & Methods ) , since a substantial part of the cluster is outside the sampled region . For grid 1 of subject 1 , we do not find evidence for a spatially heterogeneous distribution of infected cells ( see also Figure 5B of [15] ) . Therefore , the cluster detection algorithm is affected and determines a “cluster” for each infected cell ( see Figure S1 and Figure S2 ) . An important point is that we are only analyzing 2D sections , and most likely clusters are defined in 3D . Thus , the total number of cells in a cluster will be larger than our minimal estimates . After defining the clusters and their sizes , we analyzed the profile of the intracellular HCV RNA level within the clusters , i . e . , the viral landscape or viroscape [15] . The relationship between the maximal radius of a cluster of infected cells and the HCV RNA content in the cell that presumably founded the cluster is shown in detail in Figure 3A . For example , the hepatocyte with the highest amount of HCV RNA on grid 3 of subject 2 contained 8 . 8 IU/cell , and the corresponding cluster of infected cells reached a maximal radius of ( see Table 1 ) . The mean cluster radius among all patients was ( 95%-CI [26 . 7 , 57 . 9] ) with variability between and within patients . There was no significant relationship between the maximal cluster radius and the HCV RNA content in the assumed founder cell of the cluster , i . e . , the cell with the highest HCV RNA content ( Figure 3A , linear mixed effects model ) . This result does not change if grids 1 of subject 1 and 3 , which could be outliers ( see above ) , are neglected . The iterative algorithm that was used to determine the maximal cluster radius ( Figure 1 ) also describes the HCV RNA content in a cell as a function of the distance to the core of the cluster assuming radial spread . We find that the viroscape of a cluster of infected cells , i . e . , the topography of the amount of intracellular viral RNA , can be characterized by a biphasic decline in the amount of HCV RNA content of cells with increasing cluster extension ( Figure 3B ) . As no relation could be found between the maximal cluster extension and the HCV RNA content in the assumed founder cell of the cluster ( Figure 3A ) , we then asked if there was a relationship between the total amount of HCV RNA in a cluster and the actual size of a cluster ( i . e . , number of cells in the cluster ) . To this end , we have to convert continuous characteristics like the radii of the cluster , , back into discrete number of cells . For consistency with the method above , we constructed 10 , 000 random bootstrap replicates of the observed clusters . To maintain the properties in the data , these replicates were defined consistent with individual cluster characteristics , namely their maximal cluster extension in µm ( Table 1 ) and the observed biphasic decline in HCV RNA with increasing cluster extension ( Figure 3B ) ( see Materials & Methods ) . The results of these analyses do not indicate a significant correlation between cluster size ( in number of cells ) and the total amount of HCV RNA in a cluster ( Figure 3C , Linear-mixed effects model , , neglecting grids 1 of subject 1 and 3 ) . Previously detected correlations between cluster size and viral load ( see Fig . 6 of [15] ) are largely affected by grid 1 of subject 3 . Several aspects of viral replication and transmission dynamics should influence the observed HCV RNA profile within a cluster of infected cells . These include the rate at which neighboring cells get infected , the average time it takes for a cell to start viral replication , the rate at which the virus replicates , the rate at which intracellular HCV RNA is degraded , and the lifespan of infected cells ( Figure 4A ) . We assume that the amount of HCV RNA inside an infected hepatocyte can be used as a surrogate for the time a cell has been infected , the so called age of infection , . That is , hepatocytes with higher amounts of HCV RNA most likely have been infected for longer . After infection , a number of events have to occur , taking time , before viral RNA replication starts . Following our previous work [27] , we assume that after , each viral RNA can be copied in ∼6h , which would be the doubling time of the amount of viral RNA inside a cell , if there was no export of virions or degradation of viral RNA . However , we have estimated that on average 75% of new viral RNA is exported in virions [27] , although this fraction could vary over time . In our data , we find hepatocytes with up to ∼50 IU/cell , corresponding to ∼100 HCV RNA copies/cell ( 1 IU1 . 96 genome copies [15] ) . This value is consistent with other ex vivo studies [14] , [17] , [19] . With these parameters , we simulate the stochastic replication dynamics of HCV within a cell [27] following the accumulation of HCV RNA within the cell . Varying between , we estimate that cells reach a level of HCV RNA copies/cell on average days p . i . ( 95%-CI [3 . 5 days , 7 . 7 days] ) . Based on the observed HCV RNA content , we estimate that cells in the liver samples are infected for less than a week ( Figure 4B and Table 2 , see also Text S2 ) . The cells assumed to build the core of the individual clusters , i . e . , earliest infected and containing the largest amount of intracellular viral RNA , are estimated to be infected for on average ∼4 days ( Table 2 ) . Calculating the elapsed time between infection of the cells with the largest amount and the lowest amount of intracellular viral RNA within a cluster , , i . e . , the time the cluster has been expanding , we estimate that the observed clusters of infected cells were formed on average in 2 days ( Table 2 and Figure 4C ) . We do not find a relationship between the maximal cluster size and the estimated age of infection of the hepatocyte that presumably founded the cluster , nor between the maximal cluster extension and the time a cluster has been expanding , ( Figure 4B , C , , linear mixed-effects model ) . This could indicate variability in viral propagation compared between and within patients , possibly due to individual and locally available host factors as well as individual viral characteristics , such as the ability to spread cell-to-cell and the speed of spread . The ratio between the rate at which neighboring cells get infected and the rate at which HCV RNA accumulates within an infected cell influences the steepness of the viroscape with increasing cluster size . For example , if the virus has a high local transmission probability combined with a short time between cell infection and new virion production , the so-called eclipse phase , we would expect a more uniform distribution of HCV RNA content in individual cells , i . e . a flatter viroscape . In contrast , a long eclipse phase combined with a low transmission probability will lead to increased accumulation of HCV RNA within the cells that got infected early on , showing larger differences in the HCV RNA content between the cells within a cluster . We simulated cluster expansion by local spread within a two-dimensional lattice of hepatocytes to investigate how the transmission probability per intracellular viral RNA per minute and the duration of the eclipse phase would influence the viral profiles in a cluster ( Text S3 ) . Calculating the average fractional decrease between the viral RNA within the founder cell and the HCV RNA amount in the surrounding cells in a cluster of size 9 , we observe that the largest declines are obtained for long eclipse phases , , and low cell-to-cell transmission probabilities per positive strand intracellular HCV RNA per minute , ( Figure 4D ) . The viral profiles in our data show a biphasic decay ( Figure 3B ) , with an average first-phase fractional decrease of intracellular HCV RNA among all patients ( Figure 4D ) of ( range [0 . 30 , 0 . 91] ) relative to the maximal amount of HCV RNA in the founder cell of the cluster . This means that the HCV RNA level in direct neighbors of the core of the cluster is on average ∼40% , , of the HCV RNA level within the core , i . e . , that the intracellular HCV RNA level decreases by ∼60% . We can recover these observations in our simulations if the probability for virus transmission per intracellular viral RNA per minute , , lies approximately between and , where defines the rate at which infected cells become infectious , i . e . , start producing virus ( Figure 4D ) . For each patient , two to three grids of hepatocytes were analyzed by scLCM . Thus , only a small number of hepatocytes ( cells ) was sampled and analyzed in comparison to the estimated total number of hepatocytes in a human liver [28]–[31] . To determine if the infection patterns in the sampled sections of liver tissue are consistent with the infection dynamics in the entire liver , we compared the measured serum viral load for each patient to the expected serum viral load from the observed liver sections using an age-structured model of HCV infection ( Materials and Methods and [9] ) : With an export rate , , of intracellular viral RNA as virions and a clearance rate of extracellular virus , , the viral load per ml of serum produced by infected hepatocytes can be estimated by , where denotes the average amount of intracellular HCV RNA per infected hepatocyte and is a scaling constant , relating the total amount of virus in the body , , to virus per ml . We used our data to determine and , with denoting the observed fraction of infected hepatocytes and the total number of hepatocytes in a human liver . With the previously estimated values of and [10] , and with , the average total extracellular fluid volume for an individual , we estimate based on the sampled liver biopsies ( Table 3 ) . For each subject , estimates for are in a range consistent with , but slightly lower than , the measured serum viral load ( Table 3 ) . On average our estimates are 0 . 3 log ( or 5% ) lower than the measured viral load . A possible explanation for this lower estimate is a higher concentration of HCV RNA in plasma than in other body fluids [32] , as we assumed a homogeneous distribution throughout the extracellular fluid when calculating . The other possibility is that the value of could vary among patients , whereas we are using an average estimate for all of them . We note that although could also vary , recent estimates show that c∼22 day−1 with little inter-patient variability [10] . In summary , based on this analysis it seems reasonable to assume that the patterns of infection ( fraction of infected cells , density and level of intracellular HCV RNA ) in the liver sections are an appropriate and consistent representation of the infection in the whole liver .
In previous work , we showed that infected hepatocytes tend to occur in clusters [15] . However , our previous cluster detection methods [15] were not able to specify cluster characteristics , such as the cluster size and the dynamics of cluster formation . Here we developed statistical methods to determine the properties of such clusters of infected cells identified by scLCM . With our approach , which is based on assuming infected cells are distributed according to a Matérn cluster process , we are able to systematically determine the size and the internal structure of clusters . In comparison to other clustering methods [33] , [34] , our stepwise procedure allows us to address the dynamics of the cluster building process based on the varying but interdependent levels of intracellular viral RNA . Applying our method to liver biopsy samples from four different chronically infected HCV patients analyzed by scLCM , we find that clusters of infected hepatocytes in a biopsy section comprise between 4 and 50 cells ( Table 1 ) . While we show results based on a Matérn cluster process that assumes radial cluster shapes , we also fitted a Thomas process , which allows for normally distributed “offspring” around cluster centers , to the data [23] . This did not change our results in terms of cluster size or cluster frequency ( not shown ) . One caveat is that we are studying 2D sections , while in vivo the clusters of infection would be 3D . It is not easy to expand our results to that situation , because each 2D section could represent a different cross-section of the corresponding 3D cluster . Sequencing the intracellular HCV RNA in future studies could indicate whether viruses from nearby clusters are distinct , and therefore originate from separate clusters , or whether they represent cross-sections of irregular 3D clusters . It is interesting to compare our results to 2D infection in vitro . In an experiment using HCV infected Huh-7 . 5 cell lines , Lacek et al . [18] found cluster sizes ranging between 1–40 cells around 2 days after the initial infection , similar to the sizes and timing we report . However , when blocking local transmission by appropriate monoclonal antibodies , they found a significant reduction in cluster sizes [18] . At the highest antibody concentration used , the maximum cluster size was reduced to 5 cells ( see Fig . 3A in [18] ) . The agreement of the cluster sizes found in vitro with the cluster sizes found in our study , as well as the findings of others that HCV infected cells tend to occur in clusters [12] , [14] , [15] , [17] support the hypothesis that local spread of HCV is an important contributor to the patterns we observed . The hypothesis of local spread is also corroborated by our second result , indicating that the HCV RNA amount within infected hepatocytes of a cluster declines with increasing cluster extension ( Fig . 3B ) . Using the HCV RNA amount inside an infected hepatocyte as a surrogate for the time since infection , this observation supports the notion that HCV spreads locally to neighboring cells and then begins replicating within those cells . However , we cannot distinguish the hypothesis that the local spread is mediated by diffusing viral particles that rapidly bind to and infect neighboring cells from the hypothesis of cell-to-cell transmission . Most likely both mechanisms operate but to varying degrees . We note , however , that random seeding of the clusters from the blood , as we assumed , and the observed selective pressure of HCV specific antibodies [35] , [36] suggests that transmission by freely diffusing viral particles remains an important mode of transmission throughout chronic infection . Another possible explanation for the clustered distribution of infected hepatocytes could be that some cells in initially homogeneously infected regions are cleared by innate and adaptive immune responses . While we cannot formally reject these other hypotheses , we think that the sum of our data more likely supports additional local spread . Using mathematical models to describe the accumulation of intracellular viral RNA with time , we estimate that the hepatocytes we have analyzed have been infected for less than 7 days ( Table 2 and Text S2 ) . Although previous estimates for the turnover of infected hepatocytes based on the kinetics of plasma HCV viral load decrease under treatment have been highly variable among different patients [3] , [22] , [37] , a large cohort study of 2100 chronically HCV infected patients under treatment with peginterferon α-2a with or without ribavirin estimated the average lifetime of infected hepatocytes to be around 5–7 days [21] . Several studies examined the proliferation of hepatocytes during chronic HCV infection [38]–[41] . Staining for the cell proliferation marker Ki-67 shows that between 1%–2% [38]–[40] , and up to 10% ( for early stages of liver inflammation as here [41] ) of hepatocytes were Ki-67+ during chronic HCV infection . If we assume that an hepatocyte takes ∼24 h to undergo cell division in vivo , then these Ki-67 measurements imply 1%–10% of hepatocytes are dividing per day . We found 20%–40% of cells infected , with a turnover rate of ∼0 . 14/day ( corresponding to a 7 day average lifetime ) , thus we predict an overall turnover rate in these chronically infected individuals of between 3% and 6% per day , in agreement with the estimates based on Ki-67 . Our mathematical models calculating the age of infection assumed that each hepatocyte has the same maximum number of intracellular replication complexes and the same maximum level of intracellular viral RNA , copies/cell . However , individual hepatocytes might vary substantially in their replication dynamics , possibly due to local innate and adaptive immune responses , as well as cell specific factors [42] . In addition , the maximal amount of intrahepatic HCV RNA levels could be variable . We found infected hepatocytes with up to 50 IU/cell ( ∼100 HCV RNA copies/cell ) ( Table 1 ) ; other in situ studies found similar numbers . For example , Chang et al . found a maximum of 74 HCV RNA molecules per hepatocyte [19] and Stiffler et al . observed from less than one copy to a maximum of 10 copies per cell [14] , averaged over all cells . If one assumes that about 10% of cells are infected [12] , [17] , [19] , then their estimate is similar to our estimates for the number of HCV genomes per infected cell . In vitro observations indicate much higher numbers , with cells accumulating thousands of copies of viral RNA within 72 hours [43]–[50] . However , the systems used in vitro typically involve high multiplicity of infection or transfection and the cells used , i . e . , hepatoma cells , are highly permissive to infection , often with a very diminished or even absent endogenous interferon response , impairing the comparison to the situation in hepatocytes in vivo . If was substantial higher in vivo than what we observed , we would expect the infected cells in our biopsies to have been infected even more recently . Such a short lifespan seems unlikely . While it is appropriate to use the amount of intracellular HCV RNA in infected hepatocytes as a surrogate for the time since infection , without knowing all the relevant factors that might influence viral replication and accumulation ( e . g . type I IFN-responses , etc . ) , the accuracy of estimates for the time since infection of a cell based on intracellular HCV RNA content are difficult to judge . However , less than 7 days of infection for the cells in our samples seems a consistent average estimate using different mathematical methods . We found that the total size of a cluster did not correlate with the amount of HCV RNA in the cell that presumably founded the cluster ( Figure 3A ) . Such correlation would be expected if cell infection continued unimpeded , because higher HCV RNA levels in the founder cell would correspond to a cell infected for longer and , hence , to a potentially longer period of cluster expansion . The biphasic decline of the viroscape of a cluster with increasing cluster extension , particularly for large clusters ( Figure 3B ) , indicates that small sub-clusters characterized by cells with high levels of intracellular viral RNA are surrounded by larger areas of cells containing substantially lower viral load . Assuming a biphasic linear or biphasic exponential decline provided a better description of the underlying data than their monophasic analogons . A decline , although not the specific biphasic profiles that we find evidence for , has been reported before using other techniques [12] , [17] , [19] . Such a biphasic profile could indicate that local factors influence the progression of infection and the expansion of a cluster , such as local immune responses . One scenario could be that during establishment of the cluster , i . e . when the founding cell is infected , the viral landscape is mainly determined by the ratio between the rate of accumulation of viral RNA within a cell , including effects of the initial delay before viral RNA production starts , and the rate at which neighboring cells are infected . As infection progresses , infected cells and their neighbors start the production of antiviral factors , such as type I IFN , that protect uninfected cells from getting infected and interfere with the viral replication within infected cells . The “wave” of antiviral factors might overtake the propagation of infection , inhibiting viral production within infected cells at the border of a cluster and limiting cluster expansion . This would lead to a flatter viral landscape at the edges of a cluster . In this scenario , the first phase of decay in the amount of HCV RNA per cell with increasing cluster extension could be used to determine the ratio between viral transmission and viral replication , and the onset of the second phase might allow us to determine the effectiveness of the endogeneous antiviral response in the liver in these treatment naïve patients . Whether local infection stimulates a local immune response is still controversial [14] , [15] , [17] . We did not find any correlation between infection status of a cell and the expression of interferon in those cells or neighboring cells in our previous work assessing one interferon stimulating gene ( ISG ) [15] , in agreement with other results [14] . However , a recent report did find significant co-localization of ISG expression and HCV infection [17] . More details about the influence of endogeneous type I IFN responses on the spatial propagation of HCV infection within the liver , as well as about the dynamics of cluster propagation and IFN expression , are needed to determine the extent to which these response might limit viral transmission and replication [42] , [51] , [52] . Such factors could also explain why a correlation between cluster size and HCV RNA content could not be observed . At the edges of the clusters , the minimal amount of HCV RNA within cells is between 1–2 IU/cell ( Figure 3B ) , which is close to the limit of detection . However , the possibility that the second phase of decay might be due to an artifact of the experimental method , i . e . , measuring of extracellular viral RNA that sticks to the surface of the cell , is rather unlikely . Assuming that a hepatocyte has a cubic shape with a side of 20 , and that liver sinusoids have around the same diameter , the fluid volume above the apical surface of a cell can be approximated by a cylinder of radius and length , i . e . , . With a typical plasma viral load of , the chance that a virion would be in the fluid volume around a cell is less than 1 . 2% , suggesting that it is unlikely that the low HCV RNA levels at the edge of a cluster are due to extracellular virus . More precise estimates could be made if information on the rates of virus binding , dissociation and entry were available , but it seems likely that the shape of the viroscape is a characteristic of the dynamics of local cellular infection , rather than an artifact of the experimental method . A recent study reported a significant correlation between the proportion of infected hepatocytes and serum viral load [17] . Here we presented a mechanistic dynamic model that explains this result . Indeed , the picture of local infection described above , even with the caveats discussed ( low sample size , 2D sections ) , seems to be an appropriate and consistent representation of infection in the whole liver . In fact , we used this dynamic model of viral infection to show that the predicted serum viral load in the patients we analyzed is very similar to their measured plasma viral load , if we assume that the patterns of infection ( proportion of infected cells , clustering , level of intracellular HCV RNA ) observed in the biopsy samples are replicated throughout the liver . While this result indicates the consistency of using models on ordinary differential equations ( ODE ) to analyze viral dynamics on a systems level , the observed spatial heterogeneity of HCV infection within the liver suggests the importance of considering space when analyzing infection dynamics within solid tissue on a cellular level . Current HCV viral dynamics models are based on infection by cell-free virus , neglecting cell-to-cell transmission or local effects of IFN responses [8] , [53] . It will be interesting to incorporate these features in future models , and to examine how well these models agree with viral declines observed in patients on treatment . We based our analysis on a small number of cells , studied in unprecedented detail , assuming that they are a reasonable representation of the infection process . Due to the estimated level of heterogeneity , more data are needed to predict the frequency of infected hepatocytes in a chronically infected patient , as well as to analyze the influence of type I IFN responses on the observed spatial patterns . Still , our method represents a novel approach to infer infection dynamics from static spatial data .
In four chronically HCV infected patients ( Subject 1–4 ) up to three sections of liver tissue were analyzed by single cell laser capture microdissection ( scLCM ) . On each section , a grid of hepatocytes was analyzed for their HCV RNA content ( see Fig . S1 for the individual results ) as previously described [15] . The sensitivity level of the method to detect HCV RNA was set at 1 IU/cell . HCV RNA content was normalized to 7SL expression , a small ribosome-associated RNA which is abundant in the cytoplasm [15] . Using rigorous controls we ensured the precision of our method in determining single cell measurements ( see Supporting Information Text S3 ) . Cells in which the HCV RNA content could be detected but could not be normalized due to missing 7SL expression measurements were treated as infected throughout most of the analysis . None of the subjects had received treatment , and none were co-infected with HIV or HBV . All patients had genotype 1 infection and limited liver disease ( Metavir 0–1 ) . Full details of the experimental methods can be found in Kandathil et al . [15] . The experimental data are provided in the Supporting Information ( Dataset S1 ) . Based on our results on the characterization of clusters of infected cells ( Fig . 3B ) , we found that the amount of HCV RNA in each infected hepatocyte , , decreases with increasing cluster extension , . We used different models to describe this decline in HCV RNA viral load , assuming either monotonous or biphasic decay . Assuming biphasic linear decay , this decline in HCV RNA viral load can be described by the following function ( 1 ) Here , the parameter determines the maximal HCV RNA content measured in a single cell belonging to this cluster , and is the minimal radius of a cluster , where each cell has an HCV RNA content of ( Fig . 1A ) . The parameters and define the rates at which the HCV RNA amount within a cell is decreasing with increasing cluster size for phase 1 and 2 , respectively , while determines the cluster extension at which the rate of decrease changes . Equation ( 1 ) is fitted individually to the data of each of the different grids shown in Figure 3B , and the parameter values for , and are estimated . The estimated functions for are then used to sample the HCV RNA content in cells of simulated clusters in order to determine the relationship between the total size and the total intracellular viral load of infected cells . In addition to the assumption of a biphasic linear decay , we also fitted a model assuming an exponential decay , as well as a biphasic exponential decay of intracellular HCV RNA content with increasing cluster size to the data ( not shown ) . All these assumptions do not change our results with regard to the relationship between cluster size and total viral burden in the cluster . Models assuming a biphasic decay show better fits across all patients and samples judged by the mean residual sum of squares ( MRSQ ) , i . e . , the residual sum of squares divided by the difference between the number of data points and the number of free parameters , as well as the Akaike information criterion , than their monophasic analogons ( mono-linear , mono-exponential ) . Fitting a Matérn Cluster process to the spatial point patterns of HCV molecules as described above , we are able to determine the radius , , for discs of similar amounts of intracellular RNA , i . e . , each cell in this disc has at least this amount of intracellular viral RNA . To get a better impression on actual cluster sizes , we convert this radius based on HCV molecules back to cell numbers . As the Matern Cluster process assumes spherical cluster shapes and HCV RNA molecules are distributed randomly within each cell ( size = ) , the obtained radius might cut only a part of a cell , or include parts of uninfected cells ( see Fig . 1B ) . Therefore , we calculated a minimal and maximal cluster size in number of cells , assuming either a radial or quadratic cluster extension: The minimal number of cells belonging to a disc with radius is determined by , where denotes the area of a hepatocyte . The possible maximal number of cells in a disc is estimated by assuming the area of the cluster is given by a square with edge length , hence , . Our stepwise algorithm determines the cluster extension , , as a function of the minimal amount of intracellular HCV RNA , , in a cell belonging to a certain ( sub- ) cluster . By piling the different discs on top of each other ( Fig . 1B ) , the total clusters are assumed to be structured like the rings in a tree trunk with each “cell-ring” around the founder cell having similar intracellular HCV RNA content . When reconstructing individual clusters in terms of cluster size and intracellular viral RNA , we have to determine the number of cells and the range of the HCV RNA content in cells belonging to a ring at a fixed distance to the founder cell of the cluster . Therefore , for each bootstrap replicate , we sampled the number of cells , , belonging to a disc with radius , , with , based on a uniform distribution on . The actual number of cells in each ring is then defined by , with , i . e . , the assumed founder cell . In addition , each of the infected hepatocytes of ring was filled with an amount of HCV RNA uniformly sampled from , the range of the amount of HCV RNA estimated in the corresponding ring . The total number of cells belonging to the cluster , , as well as the total HCV RNA content in the cluster , , was then calculated based on the sum over all “rings” of the cluster , hence , , and . For each cluster , 10 , 000 bootstrap replicates were performed . We use a stochastic model of HCV replication dynamics developed recently [27] to estimate the age of infection of a cell based on its amount of intracellular HCV RNA . In brief , after an initial time post infection , replication complexes are formed over an average time of based on in vitro experimental data where negative strand RNA , assumed to represent replication complexes , is first detected post infection [50] . Each replication complex generates a new replication complex with a probability of per generation time of [27] . Positive strand viral RNA is produced by replication complexes with on average 75% of the newly generated viral RNA estimated to be exported in virions [27] . The remaining viral RNA accumulates within the cell , increasing the intracellular amount of positive strand HCV RNA . In our data , we observe a maximal intracellular HCV RNA amount of HCV RNA copies/cell , corresponding to the observed level of up to 50 IU/cell ( 1 IU 1 . 96 genome copies [15] ) , which is consistent with other studies [14] . For each cell , we run 10 , 000 instances of the stochastic model of viral replication and formation of replication complexes , varying between and [50] to estimate the age of infection based on the amount of intracellular HCV RNA . In addition to the stochastic model , we also analyze the data using an analytical model as developed in [5] ( see Text S2 ) . We used an age-structured mathematical model [9] to compare the intracellular HCV RNA content per hepatocyte ( in up to 300 hepatocytes analyzed per patient ) to the total serum viral load . This model explicitly describes the HCV viral load , , as a function of the intracellular HCV RNA content of single hepatocytes , . The model equation for is [9] , [10]: ( 2 ) where the intracellular amount of viral RNA in an hepatocyte , , is dependent on the time since infection , , called the age of infection of a cell . Intracellular HCV RNA is packaged into virions and exported from an infected cell at rate . The extracellular viral load , , increases due to the HCV RNA export at rate , from each infected hepatocyte of age , , and decreases due to viral clearance at rate per virion . For each patient , the viral load measured in HCV RNA copies per ( or international units ( IU ) to normalize among HCV RNA assays ) of serum was determined at the time of biopsy ( see Table 3 ) . As each patient was in the chronic stage of HCV infection , we assumed that viral load had reached steady-state , . According to Eq . ( 2 ) , the measured viral load , , would be related to the total number of infected hepatocytes in the liver , , by ( 3 ) where denotes the average HCV RNA content in an infected hepatocyte . The parameter is a scaling factor to account for the fact that is measured per ml of serum , while defines the total number of HCV viral particles in the human body . This scaling factor is set to , the average total extracellular fluid volume for a individual [58] . To compare the measured serum viral load , , to the individual liver sections , we estimate the viral load produced by infected cells , where denotes the estimated frequency of infected hepatocytes for each subject , and the total number of hepatocytes in a human liver , hence , . If , the inferred dynamics from the sampled liver sections are consistent with the average infection dynamics in the whole liver . To calculate , we first determined the frequency of infected hepatocytes in the individual liver samples , , as well as the observed mean HCV RNA content in infected hepatocytes , , combining all sections of one subject ( see Table 3 ) . The total number of hepatocytes in a human liver is in the range of cells [28]–[31] , and we used the upper limit of . The viral clearance rate and the viral export rate for HCV , , have been estimated recently as and [10] . The statistical dependency between different attributes of clusters of infected cells ( e . g . size of cluster and total amount of HCV RNA ) was analysed using Spearman's correlation coefficient and linear mixed effects models , with subject as the random effect . Linear mixed effects models take into account that several sections of liver tissue originate from the same patient . All analyses were performed using the language of statistical computing [59] . Core functions are provided in the Supporting Information ( Protocol S1 ) . | Around 170 million people worldwide are chronically infected with the hepatitis C virus ( HCV ) . Although partly successful treatment options are available , several aspects of HCV infection dynamics within the liver are still poorly understood . How many hepatocytes are infected during chronic HCV infection ? How does the virus propagate , and how do innate immune responses interfere with the spread of the virus ? We developed mathematical and computational methods to study liver biopsy samples of patients chronically infected with HCV that were analyzed by single cell laser capture microdissection , to infer the spatial distribution of infected cells . With these methods , we find that infected cells on biopsy sections tend to occur in clusters comprising 4–50 hepatocytes , and , based on their amount of intracellular viral RNA , that these cells have been infected for less than a week . The observed HCV RNA profile within clusters of infected cells suggests that factors such as local immune responses could have shaped cluster expansion and intracellular viral replication . Our methods can be applied to various types of infections in order to infer infection dynamics from spatial data . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"hepatitis",
"c",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"infectious",
"hepatitis",
"simulation",
"and",
"modeling",
"population",
"modeling",
"gastroenterology",
"and",
"hepatology",
"biology",
"and",
"life",
"sciences",
"infectious",
"disease",
"modeling",
"computational",
"biology",
"viral",
"diseases",
"liver",
"diseases",
"mathematical",
"modeling",
"research",
"and",
"analysis",
"methods"
]
| 2014 | Inferring Viral Dynamics in Chronically HCV Infected Patients from the Spatial Distribution of Infected Hepatocytes |
The C-terminal domain ( CTD ) of RNA polymerase II ( RNAPII ) is composed of heptapeptide repeats , which play a key regulatory role in gene expression . Using genetic interaction , chromatin immunoprecipitation followed by microarrays ( ChIP-on-chip ) and mRNA expression analysis , we found that truncating the CTD resulted in distinct changes to cellular function . Truncating the CTD altered RNAPII occupancy , leading to not only decreases , but also increases in mRNA levels . The latter were largely mediated by promoter elements and in part were linked to the transcription factor Rpn4 . The mediator subunit Cdk8 was enriched at promoters of these genes , and its removal not only restored normal mRNA and RNAPII occupancy levels , but also reduced the abnormally high cellular amounts of Rpn4 . This suggested a positive role of Cdk8 in relationship to RNAPII , which contrasted with the observed negative role at the activated INO1 gene . Here , loss of CDK8 suppressed the reduced mRNA expression and RNAPII occupancy levels of CTD truncation mutants .
The largest subunit of RNA polymerase II , Rpb1 , has a unique C-terminal domain ( CTD ) composed of the repeated sequence Tyr-Ser-Pro-Thr-Ser-Pro-Ser ( Y1 S2 P3 T4 S5 P6 S7 ) [1] , [2] . Although the CTD is highly conserved across species , the number of repeats varies in a manner resembling genomic complexity , with 25/26 repeats in Saccharomyces cerevisiae and 52 in humans [3] . Deletion of the entire CTD is lethal in budding yeast , while strains carrying 9–13 repeats are viable but display conditional phenotypes [4] , [5] . While not required to support basal transcription in vitro , the CTD is critical for the response to activator signals in vivo [6] , [7] . For example , CTD truncation mutants exhibit reduced activation of INO1 and GAL10 upon switching to inducing conditions [7] . The CTD is a scaffold for the recruitment of RNA processing and chromatin remodeling factors , a function linked to its differential phosphorylation at specific residues of the heptapeptide repeat [3] . Transcription begins with the recruitment of RNAPII with an unphosphorylated CTD to promoters , where it interacts with components of the transcription pre-initiation complex ( PIC ) [8] , [9] . Following , it is phosphorylated at S5 and S7 by the general transcription factor TFIIH , facilitating recruitment of capping enzymes and release of RNAPII from promoter-bound elements [10]–[13] . Elongation is characterized by phosphorylation of S2 by Ctk1 and Y1 and T4 by yet unidentified kinases [14] , [15] . S2 and Y1 phosphorylation play a role in the temporal recruitment of elongation and termination factors [15] . Subsequently , termination entails removal of all phosphorylation marks by Fcp1 and Ssu72 to regenerate an initiation competent RNAPII molecule [16]–[18] . While early work aimed at understanding CTD function uncovered a set of SRB ( Suppressor of RNA Polymerase B ) genes , a comprehensive genetic network governing CTD function has yet to be fully elucidated [19] . Of the identified SRB genes many encode members of a large multisubunit complex known as Mediator [20] . Mediator was first identified in vitro as a cellular fraction that stimulates RNAPII transcription , and is now known to not only physically interact with the CTD , but also to be important for the response to up-stream regulatory signals [21] . Although primarily associated at RNAPII gene promoters , Mediator also resides at open reading frames ( ORFs ) [22] , [23] . Furthermore , Mediator is organized into four functionally distinct submodules: head , middle , tail and Cdk8 module [24] . The head module interacts with the CTD while the tail and middle modules interact with gene-specific and general transcription factors [25] , [26] . The Cdk8 kinase module likely associates transiently with the core Mediator complex and has roles in both transcriptional activation and repression [27] , [28] . This dual activity is in part mediated by Cdk8's ability to phosphorylate multiple regulatory components of the transcription machinery . These include several transcription factors as well as factors more generally required for transcription such as the CTD itself [27] , [29]–[31] . While the mechanistic role of some of these phosphorylation events is unclear , CTD phosphorylation by Cdk8 prior to promoter association inhibits RNAPII recruitment and transcription initiation in vitro [29] . In contrast , CTD phosphorylation by Cdk8 and Kin28 following promoter association promotes RNAPII release from the PIC and thus stimulates transcription activation [30] . The work here highlighted the functional circuitry between the RNAPII-CTD and Mediator in the regulation of cellular homeostasis , gene expression , and the transcription factor Rpn4 . Our data uncovered a length-dependent requirement of the CTD for genetic interactions and mRNA levels of genes expressed under normal growth conditions . Truncating the CTD primarily resulted in increased expression and RNAPII association at a subset of genes , in part mediated by changes to transcription initiation . These genes had preferential association of Cdk8 at their promoters and were regulated by the transcription factor Rpn4 . The expression and RNAPII binding defects of the majority of this subset of genes were suppressed by deleting SRB10/CDK8 , suggesting that in CTD truncation mutants , Cdk8 functioned to enhance transcription and RNAPII association at a subset of genes . Conversely , our data also revealed that deletion of CDK8 suppressed the activation defects of CTD truncation mutants at the INO1 locus thus indicating that Cdk8 also functioned to repress transcription and RNAPII association in CTD truncation mutants .
To broadly determine the requirement of CTD length for cellular function , we used Epistasis Mini Array Profiling ( E-MAP ) to generate genetic interaction profiles of CTD truncation mutants containing 11 , 12 , 13 or 20 heptapeptide repeats ( rpb1-CTD11 , rpb1-CTD12 , rpb1-CTD13 and rpb1-CTD20 respectively ) against a library of 1532 different mutants involved principally in aspects of chromatin biology and RNA processing [32] ( Table S1 ) . CTD truncations were created at the RPB1 locus by addition of a TAG stop codon followed by a NAT resistance marker . As a control for the genetic integration strategy we also generated RPB1-CTDWT , which contained a NAT resistance marker following the endogenous stop codon . While the minimal CTD length for viability is 8 repeats , we focused on strains starting at 11 repeats as mutants bearing shorter CTDs were significantly unstable in our hands , consistent with previous findings [33] . Overall our data revealed a greater number of significant genetic interactions as the CTD was progressively shortened , an effect consistent with increasingly disrupted function ( Figure 1A ) . Furthermore , while hierarchical clustering based on Spearman's rho correlation delineated two major clusters , the first including rpb1-CTD11 , rpb1-CTD12 and rpb1-CTD13 and the second consisting of rpb1-CTD20 and RPB1-CTDWT ( Figure 1B ) , individual genetic interactions revealed more nuanced CTD length-dependent genetic interaction patterns ( Figure S1 ) . For example , aggravating interactions were observed with strains lacking ASF1 , RTT109 and DST1 when the CTD was truncated to 13 repeats or shorter , while truncation to 11 repeats was required for aggravating interactions with SET2 , RTR1 and SUB1 . Collectively , this data revealed significant and specific functional alterations to the CTD as a result of shortening its length and suggested that individual pathways required different CTD lengths for normal function . Finally , given that we identified significant genetic interactions with genes involved in a variety of processes , we compared the E-MAP profile of our shortest CTD truncation with all previously generated profiles to determine which pathways were principally affected by truncating the CTD . This analysis revealed that four of the ten most correlated profiles belonged to loss of function alleles of genes encoding subunits of TFIIH and Mediator ( RAD3 , MED8 , MED31 and MED20 ) suggesting that shortening the CTD results in genetic interaction patterns most similar to mutants affecting transcription initiation ( Figure 1C ) . Although the CTD plays a major role in the response to activator signals in vivo , its general involvement in transcription is less well defined . To investigate this important aspect , we generated gene expression profiles of CTD truncation mutants in normal growth conditions ( Table S2 ) ( Complete dataset can be found in array-express , code E-MTAB-1431 ) . Similar to the E-MAP data , the expression data revealed a length-dependent requirement for CTD function , with the severity and number of transcriptional changes increasing as the CTD was progressively shortened ( comparison of E-MAP vs . expression profiles Pearson's rho 0 . 57 ) ( Figure 2A and 2B ) . This gradient effect was clearly visible in the group of genes whose transcript levels decreased upon truncation of the CTD ( Figure 2A groups A , B and C constitute genes requiring greater than 13 , 12 , and 11 repeats for normal transcription respectively ) , and thus provided strong evidence of a gene-specific CTD length requirement for normal transcription . Surprisingly , given the central role of the CTD in RNAPII function , our microarray data identified only 127 genes with significant increases in mRNA levels and 80 genes with significant decreases ( p value <0 . 01 and fold change >1 . 7 compared to wild type ) , in strains carrying the shortest CTD allele , rpb1-CTD11 . Functional characterization of the set of genes with increased and decreased mRNA levels suggested that the transcriptional alterations were not affecting a random group of genes . Instead , using previously published transcription frequency data , we found that the genes with decreased mRNA levels tended to be highly transcribed with short mRNA half-lives , while the genes with increased mRNA levels were mostly lowly transcribed with long mRNA half-lives ( Figure 2C and 2D ) [34] . In addition , these genes belonged to different functional gene ontology ( GO ) categories . The genes with increased mRNA levels were enriched for proteasome and proteasome-associated catabolism processes while the genes with decreased levels were enriched for iron homeostasis , purine metabolism and pheromone response ( Table S3 ) . Finally , these genes were differentially regulated by transcription factors ( Figure 2E ) . The genes whose expression levels decreased were principally bound by Ste12 , while those with increased expression were bound by Ume6 , Met31 , Gcn4 and most significantly by Rpn4 which bound 46% of these genes ( p value 1 . 46E-41 ) . The measured gene expression changes in CTD truncation mutants could result from either effects on the synthesis or stability of the mRNA . To differentiate between these two possibilities , we measured RNAPII occupancy genome-wide and determined if the changes in gene expression correlated with alterations in RNAPII occupancy ( Complete dataset can be found in array-express , code E-MTAB-1341 ) . Specifically , we measured RNAPII in rpb1-CTD11 and wild type cells by chromatin immunoprecipitation followed by hybridization on a whole genome tiled microarray ( ChIP-on-chip ) using an antibody specific to the RNAPII subunit Rpb3 . Despite the use of different platforms , antibodies and normalization methods , the obtained genome-wide Rpb3 occupancy profiles obtained in wild type cells were highly correlated with those previously published by several groups ( Figure S2 ) [35]–[39] . Furthermore , the occupancy maps revealed highly correlated profiles between rpb1-CTD11 and wild type cells ( Spearman's rho 0 . 85 ) , agreeing with the limited transcriptional differences detected by the expression analysis . Nonetheless , our Rpb3 occupancy plots showed clear RNAPII occupancy differences along genes that were identified as either having increased or decreased mRNA levels in the rpb1-CTD11 mutant ( Figure 3A and B ) . Accordingly , plotting the average Rpb3 occupancy scores of the differentially regulated genes in rpb1-CTD11 versus wild type cells revealed that the genes with increased mRNA levels had a significant increase in Rpb3 binding levels along their coding regions while the genes with decreased mRNA levels had a significant decrease ( one-tailed t-test p value 2 . 98e-22 and 3 . 36e-7 , respectively ) , thus suggesting a direct effect of truncating the CTD on RNAPII levels and mRNA synthesis at specific loci ( Figure 3C ) . To better understand the effect of truncating the CTD on transcription , we generated genome-wide association profiles of representative transcription associated factors . These factors included the initiation factor , TFIIB which is encoded by the SUA7 gene , the capping enzyme Cet1 , the elongation factor Elf1 , and the Set2-dependent elongation associated chromatin mark histone H3 lysine 36 trimethylation ( H3K36me3 ) ( Complete dataset can be found in array-express , code E-MTAB-1379 ) . We note that with the exception of CET1 ( which was not present on our E-MAP array ) , the genes encoding these factors had negative genetic interactions with our shortest CTD truncation allele . Our genome-wide occupancy profiles under wild type conditions were highly correlated to those previously reported ( Figure 4 and Figure S3 ) [35] , [40] . Overall , genome-wide occupancy was independent of CTD length for TFIIB , Elf1 and H3K36me3 , despite the latter having decreased bulk levels in CTD truncation mutants ( Figure S3 ) [41] . In contrast , Cet1 chromatin association decreased primarily in genes with lower transcriptional frequencies , perhaps reflective of its decreased binding to RNAPII with a shortened CTD ( Figure S3B ) [42] . Focusing on only the genes whose expression levels were altered in the CTD truncation mutants , we observed several interesting patterns . First , the levels of H3K36me3 correlated well with the transcription changes as its occupancy was decreased in genes whose expression decreased and increased in genes whose expression increased in the rpb1-CTD11 mutant ( paired t-test p value 8 . 68e-6 and 9 . 34e-23 respectively ) ( Figure 4A ) . Second , the levels of Cet1 were greatly reduced at the promoters of genes whose expression increased in rpb1-CTD11 while only slightly reduced at those whose expression decreased ( Figure 4B ) ( paired t-test p value 7 . 82e-25 and 2 . 72e-7 respectively ) . Lastly , both TFIIB and Elf1 had statistically significant CTD-length dependent occupancy changes , although the overall magnitude of change was minor compared to that of H3K36me3 and Cet1 ( Figure 4C and D ) . The genetic similarity of CTD truncation mutants with mutants encoding initiation factors along with the ChIP-on-chip profiles of RNAPII and transcription associated factors suggested that possible changes to transcription initiation in the CTD truncation mutants might mediate some of the effects on gene expression . Using a LacZ reporter gene strategy we tested if the promoter elements of a set of exemplary genes sufficed to recapitulate the observed changes in expression . These assays revealed significant increases in β-galactosidase activity when the promoter regions of a subset of genes with increased mRNA levels were tested in the rpb1-CTD11 mutant compared to wild type . These data confirmed that alterations to promoter-directed initiation events were in part responsible for the increased expression observed for these genes at their native loci ( Figure 5 ) . In contrast , the promoters of the genes with decreased mRNA levels in rpb1-CTD11 mutants showed no significant differences in β-galactosidase as compared to wild type cells . We next expanded our characterization of the CTD to explore the well-established connection to Cdk8 in more detail . First , we showed that in addition to suppressing the cold sensitive phenotype of CTD truncation mutants , loss of CDK8 could also suppress other known CTD growth defects ( Figure S4 ) [19] . Second , despite Cdk8 being able to phosphorylate the CTD , its loss had only very minor effects on the bulk CTD phosphorylation defects seen in CTD truncation mutants [43] , [44] ( Figure S4 ) . Third , we found that loss of CDK8 had striking effects on the mRNA levels of genes whose expression was dependent on the CTD . Specifically , comparison of mRNA expression profiles for rpb1-CTD11 cdk8Δ and rpb1-CTD12 cdk8Δ double mutants to the single mutants revealed wide-spread and robust restoration of most of the genes with increased mRNA levels in rpb1-CTD11 , while only a few of the genes with decreased mRNA levels appeared to be suppressed ( Figure 6A ) . The restoration of mRNA levels in the genes with increased expression in the rpb1-CTD11 mutant was mediated by regulation of RNAPII levels , as Rpb3 occupancy changed from an elevated state in the rpb1-CTD11 mutant to close to wild type levels in the rpb1-CTD11 cdk8Δ mutant ( Figure 6B ) . Accordingly , the average Rpb3 binding scores at these genes in the rpb1-CTD11 cdk8Δ mutant were significantly lower than the scores of the rpb1-CTD11 mutant and were not statistically different from the scores of wild type cells ( one-tailed t-test p value 7 . 17e-18 and 0 . 159 respectively ) ( Figure 6C ) . Consistent with fewer genes being suppressed in the set of genes with decreased mRNA levels in the rpb1-CTD11 mutant , a restoring effect on RNAPII levels was not observed at this set of genes ( Figure 6C ) . A previously characterized phenotype of CTD truncation mutants is reduced activation of INO1 and GAL10 upon switching to inducing conditions . Therefore , we investigated if loss of CDK8 could also suppress these expression defects of CTD truncation mutants [7] . Focusing on INO1 , a gene important for the synthesis of inositol and survival in response to inositol starvation , we measured INO1 mRNA levels in wild type , rpb1-CTD11 , cdk8Δ and rpb1-CTD11 cdk8Δ mutants before and after induction . In agreement with previous work , rpb1-CTD11 mutants had an impaired ability to activate INO1 expression upon induction ( Figure 7A ) [7] , [45] . Upon deletion of CDK8 , INO1 mRNA levels were robustly and reproducibly restored . This effect was corroborated with the suppression of the growth defect of CTD truncation mutants in media lacking inositol upon removal of CDK8 ( Figure 7B ) . Consistent with this being a direct effect on mRNA synthesis , Rpb3 levels throughout the INO1 gene in rpb1-CTD11 mutants were significantly lower as compared to wild type . Furthermore , upon deletion of CDK8 , the levels of RNAPII associated with the INO1 gene were restored ( Figure 7C ) . While not statistically significant , we nevertheless observed a tendency for increased Rpb3 occupancy at the 3′ end of the gene in cdk8Δ and rpb1-CTD11 cdk8Δ mutants . To understand the mechanism underlying the restoration of the transcription and RNAPII recruitment changes in the rpb1-CTD11 mutant upon loss of CDK8 , we first tried to understand the role of Cdk8 in regulating these genes . To determine if Cdk8 played a direct regulatory role at these genes , we generated a genome-wide map of Cdk8 occupancy under wild type conditions ( Complete dataset can be found in array-express , code E-MTAB-1379 ) . The average gene occupancy of Cdk8 showed clear enrichment at promoters , although we did identify Cdk8 binding to a small number of ORFs ( Figure S5 ) [22] , [23] , [46] . Focusing on CTD-length dependent genes , we observed Cdk8 occupancy at the promoters of genes with increased mRNA levels in the rpb1-CTD11 mutant ( Figure 8A ) , while very little Cdk8 was observed at the set of genes with decreased levels ( data not shown ) . Importantly , Cdk8 occupancy was not significantly altered in strains with a truncated CTD ( Figure 8A ) . In both situations , the preferential association of Cdk8 with the genes having increased expression was significant even when compared to all genes in the genome ( one-tailed , unpaired t-test p-value 0 . 0001079 for wild-type and 0 . 005898 for rpb1-CTD11 , respectively ) , thus supporting a direct regulatory role for Cdk8 at these loci ( Figure 8B ) . However , despite its significant association and robust effect on normalizing the expression levels of this set of genes , our gene expression analysis clearly showed that Cdk8 was not the sole regulator of these genes as these were generally normal in cdk8Δ mutants ( Figure 6A ) [47] . Using strict criteria , our profiles of rpb1-CTD11 and rpb1-CTD11 cdk8Δ mutants revealed robust restoration of mRNA levels at 45% of the genes with increased expression levels in the rpb1-CTD11 mutant and 24% of the genes with decreased levels when CDK8 was deleted ( Figure 6A ) . Among the genes with increased expression , those suppressed were involved in proteasome assembly and proteasome catabolic processes ( Table S4 ) . Consistently , these genes were primarily regulated by Rpn4 ( Bonferroni corrected p value of hypergeometric test 1 . 06E-26 ) . Of the genes with decreased expression , the suppressed set were mainly involved in iron transport , assimilation and homeostasis , however , no significantly associated transcription factors were identified . Given that our data thus far suggested that the restoring effect was at the level of initiation and mediated by Cdk8 , we concentrated our efforts in determining if Rpn4 , the only transcription factor found to be significantly involved in regulating the expression of the suppressed set of genes , contributed to the suppression . First , we determined if RPN4 was genetically required for the suppression of CTD truncation phenotypes by loss of CDK8 by generating rpb1-CTD11 , cdk8Δ and rpn4Δ single , double and triple mutants and testing their growth on different conditions . To test for specificity we also investigated whether the suppression was affected by GCN4 , which encodes for a transcription factor involved in the regulation of the genes whose expression increased in the rpb1-CTD11 mutant but not on those suppressed by deletion of CDK8 . Deletion of RPN4 in the rpb1-CTD11 cdk8Δ background abolished the suppression , indicating that RPN4 was genetically required ( Figure 8B; compare rpb1-CTD11 cdk8Δ to rpb1-CTD11 cdk8Δ rpn4Δ ) . In contrast , deletion of GCN4 in the rpb1-CTD11 cdk8Δ background had no effect on the suppression , suggesting that the genetic interactions with RPN4 were specific ( Figure S8 ) . Considering that Rpn4 is a phospho-protein , we also tested the involvement of two previously identified phosphorylation sites that are important for its ubiquitin-dependent degradation [48] . Introduction of the RPN4 S214/220A mutant restored the suppression in a rpb1-CTD11 cdk8Δ rpn4Δ strain in most of the conditions tested , thus demonstrating a general lack of involvement of these phosphorylation sites in the suppression ( Figure S8 right panel: compare rpb1-CTD11 cdk8Δ and rpb1-CTD11 cdk8Δ rpn4Δ ) [48] . Despite our inability to link Rpn4 phosphorylation to the suppression mechanism , the genetic analysis showed that the growth of rpb1-CTD11 rpn4Δ double mutants was more compromised than that of rpb1-CTD11 mutants alone , indicating a clear dependence on Rpn4 function for maintaining rpb1-CTD11 cell fitness ( Figure 8B compare rpb1-CTD11 and rpb1-CTD11 rpn4Δ mutants ) . This phenotypic pattern contrasted the apparent increase in Rpn4 function in a rpb1-CTD11 mutant as suggested by our gene expression analysis , and indicated that mutating CDK8 normalized , rather than abolished Rpn4 activity in rpb1-CTD11 mutants . To test this hypothesis , we measured the levels of Rpn4 fused to a hemagglutinin ( HA ) tag in rpb1-CTD11 and cdk8Δ single and double mutants . Consistent with an increase in Rpn4 function , Rpn4 protein levels were increased in rpb1-CTD11 mutants compared to wild type cells ( Figure 8D ) . Surprisingly , Rpn4 protein levels were reduced upon deletion of CDK8 in the rpb1-CTD11 mutant , consistent with the observed restoration in gene expression of Rpn4 target genes . In addition , the initial gene expression analysis as well as detailed RT-qPCR analysis of the RPN4 locus did not detect significant alterations in RPN4 mRNA levels in rpb1-CTD11 and CDK8 single and double mutants , suggesting that the effect of the CTD and Cdk8 on Rpn4 was most likely at the protein level ( data not shown ) . In support of this and consistent with the slightly elevated level of Rpn4 in the cdk8Δ strain ( Figure 8D ) , loss of CDK8 increased the half-life of Rpn4 ( Figure 8E ) . This suggested that Cdk8 was a regulator of Rpn4 stability in vivo .
Our genetic interaction , mRNA profiling , and RNAPII binding studies illuminated key linkages between CTD function , gene expression , mediator function , and the transcription factor Rpn4 . We found distinct CTD- length dependent genetic interactions and gene expression alterations during steady state growth . The majority of the expression changes in the CTD mutants were in genes whose mRNA levels increased and these were accompanied by increased RNAPII binding across their coding regions . CTD truncation mutants were primarily defective in transcription initiation as suggested by our E-MAP profile of the rpb1-CTD11 mutant and further supported by reporter assays . Removal of the mediator subunit , Cdk8 , in cells with shortened CTD restored the original mRNA levels and RNAPII occupancy profiles at a subset of genes whose expression was increased in the CTD truncation mutant , highlighting an activating role for Cdk8 in gene expression regulation . In contrast , loss of CDK8 also restored the reduced activation of the INO1 gene exemplifying the more established repressive role for Cdk8 . Finally and highly consistent with the expression results , shortening the CTD resulted in increased cellular amounts of the transcription factor Rpn4 , which was normalized upon concomitant removal of CDK8 . Underscoring its role , we found that RPN4 was genetically required for the suppression of CTD truncation phenotypes by loss of CDK8 . The mRNA analysis identified genes whose expression levels during normal growth were dependent on CTD length , thus expanding the existing knowledge of CTD function in vivo , which has been derived from a primary focus on genes activated in response to specific conditions including INO1 and GAL10 [7] . Despite the CTD being essential for viability in vivo , we detected a seemingly low number of genes with altered expression levels in rpb1-CTD11 mutants . We reconcile this with the fact that our shortest allele was four repeats above the minimum required for viability in S . cerevisiae , suggesting that we were predominantly assaying those genes most sensitive to changes in CTD length rather than the essential function of the CTD . Nonetheless , using stringent criteria our data identified a set of over 200 genes whose transcription was CTD length-dependent . As expected from the well-documented role of the CTD in transcription activation , about 40% of CTD-dependent genes had decreased expression . Surprisingly , we found that about 60% of CTD-dependent genes had increased expression . Functional analysis of the genes with increased or decreased expression upon CTD truncation revealed key differences in mRNA stability , transcriptional frequency , GO categories and associated transcription factors , suggesting differential effects on groups of genes with distinct properties . In addition , for both groups there was a high correlation between mRNA levels and RNAPII occupancy suggesting a direct effect on RNAPII function rather than changes in posttranscriptional RNA processing . Furthermore , truncating the CTD also caused changes in the association of Cet1 and H3K36me3 at genes whose expression was altered in the rpb1-CTD11 mutant . Finally , our data linked the alterations observed at the genes with increased mRNA levels to changes in transcription initiation using promoter-fusion experiments . How this latter finding can be reconciled with the minor changes in TFIIB association at the promoters of these genes remains to be determined . The increased mRNA levels and concurrent increase in occupancy of RNAPII in rpb1-CTD11 mutants presents an interesting conundrum . Seemingly , these results pointed to a previously unreported inhibitory function of the CTD , as shortening it relieved the inhibition and resulted in higher RNAPII occupancy . However , we favor a model in which these relationships are reflective of a cellular stress response elicited by impairing CTD function . Consistent with this hypothesis , CTD truncation mutants displayed heightened sensitivity to a variety of stressors , as shown by others and us [4] , [19] , [49] . Furthermore , CTD truncation mutants had increased levels of Rpn4 protein and the genes that had increased mRNA levels tended to be regulated by Rpn4 , consistent with their important contributions to the cellular stress response [50]–[52] . In addition , we investigated the molecular underpinnings of the well-established connection between Cdk8 and the RNAPII CTD . To this end , we found that deletion of CDK8 normalized the expression of genes with increased mRNA levels in the CTD truncation alleles . This observation is consistent with the less-understood role for CDK8 as an activator of transcription , likely acting by enhancing recruitment of RNAPII with a shortened CTD to its target genes . Given that Cdk8 was found to be preferentially associated with the promoters of these genes regardless of CTD length , it is likely that this represents a direct mechanism . Importantly , our data clearly showed that Cdk8 was not the sole regulator of this subset of genes as a single deletion of CDK8 does not alter their expression . Thus , in wild type cells Cdk8 associated at these genes' promoters but it only enhanced transcription when CTD function was disrupted . This observations are in agreement with Cdk8's well-established role in the response to environmental signals [31] , [53] , [54] . Furthermore , we show that Cdk8's role in activating CTD-dependent genes with increased mRNA levels was in part mediated by increasing the protein levels of the transcription factor Rpn4 , which we found to be genetically required for the suppression . Accordingly , the levels of Rpn4 protein correlated with the mRNA levels of Rpn4 targets genes in rpb1-CTD11 and cdk8Δ single and double mutants . This is consistent with the known role of Cdk8 in regulating protein levels of transcription regulatory proteins and the established function of Rpn4 in activating gene expression as a result of stress [55] . Reminiscent of recent work by several groups showing that loss of Cdk8 stabilizes Gcn4 protein levels , our data on Rpn4 protein stability provided further support of a close linkage between Cdk8 and Rpn4 , although the mechanistic details remain to be determined [56]–[58] . In addition , we note that not all suppressed genes are known targets of Rpn4 , suggesting that it is likely not the only factor linking the RNAPII CTD and Cdk8 function . The fact that removal of Cdk8 also suppressed defects in activated transcription suggested an entirely different relationship between the RNAPII-CTD and Cdk8 from the one described above , this time involving a negative role for Cdk8 . This is exemplified by the INO1 locus , where rpb1-CTD11 mutants have decreased mRNA expression and RNAPII association when grown in inducing conditions , a defect that was restored upon deletion of CDK8 . While reminiscent of the model postulating that Cdk8-catalyzed phosphorylation of the CTD prevents promoter binding of RNAPII and thus results in transcriptional repression , we do not think this is the mechanism of suppression described here [29] . First , deletion of CDK8 had no alleviating effects on the bulk phosphorylation status of either full-length or truncated CTD . Second , deletion of CDK8 alone under non-inducing conditions did not result in de-repression of INO1 , in contrast to well-characterized Cdk8 target genes [47] . Lastly , despite our genome-wide Cdk8 occupancy data showing a reproducible , albeit slight , enrichment of Cdk8 at the INO1 promoter , it does not meet our enrichment criteria , making it unclear if Cdk8 directly associates and functions at this locus ( data not shown ) . In conclusion , our data revealed a tight link between Cdk8 and the RNAPII-CTD in transcription regulation , where Cdk8 can both enhance and repress transcription , the former in part mediated by regulating the levels of the transcription factor , Rpn4 .
Strains and plasmids are listed in Supplementary materials . Partial , complete gene deletions or integration of a 3XFLAG tag was achieved via the one-step gene replacement method [59] . CTD truncations were created at the RPB1 locus by addition of a TAG stop codon followed by a NAT resistance marker and confirmed by sequencing . As a control for E-MAP and gene expression analysis we used RPB1-CTDWT . This strain contained a NAT resistance marker following the endogenous stop codon . pRS314 [RPN4] and pRS314 [rpn4 S214/220A] were obtained from Dr . Youming Xie ( Wayne State University School of Medicine ) . Reporter plasmids were generated by cloning 450 bp of the desired promoter into the Sal1 BamH1 sites of pLG669-Z [60] . E-MAP screens were performed and normalized as described previously [32] . Strains were screened in triplicate . Complete E-MAP profiles can be found in Supplementary Table S1 . Microarrays were performed in duplicate as previously described [61] , [62] . Cultures were grown with a 24-well plate incubator/reader . Spiked in controls were used to determine global changes in mRNA levels . As no such changes were detected , the expression profiles were normalized to total mRNA levels , a more reproducible measure . Differentially expressed genes were determined by p value <0 . 01 and fold change >1 . 7 compared to wild type . Complete expression profiles can be found in Supplementary Table S2 . Suppressed genes were determined as those having fold changes <1 . 1 in the rpb1-CTD11 cdk8Δ mutant . The Yeast Promoter Atlas database was used for transcription factor enrichment by performing a Hypergeometric test with Bonferroni correction ( p value 0 . 05 ) [63] . “Biological Process” ontology annotated in the Bioconductor package org . Sc . sgd . db was used for GO enrichment using the conditional Hypergeometric test ( adjusted p value <0 . 05 ) described in the following reference [64] , [65] . Supplementary Table S3 and S4 contain a full list of significant GO terms . Yeast cultures were grown in media containing 200 µM of inositol ( uninduced ) and switched to media lacking inositol for 4 hrs ( induced ) [45] . Cross-linking was done with 1% formaldehyde for 20 min . Chromatin was prepared as described previously [66] . 5 µl of anti-Rpb3 ( Neoclone ) was used . Crosslinking reversal and DNA purification were followed by qPCR analysis of the immunoprecipitated and input DNA . cDNA was analyzed using a Rotor-Gene 600 ( Corbett Research ) and PerfeCTa SYBR Green FastMix ( Quanta Biosciences ) . Samples were analyzed from three independent DNA purifications and normalized to an intragenic region of Chromosome V [67] . Primers are listed in Supplementary materials . ChIP-on-chip cultures were grown overnight in YPD , diluted to 0 . 15 OD600 and grown to 0 . 5–0 . 6OD600 units . Cross-linking and chromatin isolation were performed as above . 5 µl of anti-Rpb3 ( Neoclone ) , 4 . 2 µl of anti-FLAG ( Sigma ) or 4 µl of anti-H3K36me3 ( Abcam ab9050 ) were coupled to 60 µl of protein A magnetic beads ( Invitrogen ) . DNA was amplified using a double T7 RNA polymerase method , labeled and hybridized as previously described [66] . Samples were normalized as described previously using the rMAT software [68] . Relative occupancy scores were calculated for all probes using a 300 bp sliding window . Rpb3 and H3K36me3 experiments were normalized to input while Flag-tagged factors were normalized to untagged controls . Samples were carried out in duplicate , quantile normalized and averaged data was used for calculating average enrichment scores . For ORFs , we averaged probes whose start sites fell within the ORF start and end positions , and for promoters we averaged probes mapping to 500 bp upstream of the ORFs . Enriched features had at least 50% of the probes contained in the feature above the threshold of 1 . 5 . Enriched features were identified for each replicate and the overlap was reported as the significantly enriched set . CHROMATRA plots were generated as described previously [69] . In detail , relative occupancy scores for each transcript were binned into segments of 150 bp . Transcripts were sorted by their length and transcriptional frequency and aligned by their TSSs . Transcripts were grouped into five classes according to their transcriptional frequency as per Holstege et al 1998 . Average gene profiles were generated by averaging all probes that mapped to genes of interest . For averaging , probes corresponding to ORFs were split into 40 bins while probes corresponding to UTRs were split into 20 bins . Reporter plasmids were transformed into wild type and rpb1-CTD11 mutants and assayed as previously described [70] . Measurements were obtained from three independent cultures . Overnight cultures grown on YPD or –TRP media were diluted to 0 . 5 OD600 , 10-fold serially diluted and spotted onto YPD or –TRP plates with or without the indicated amounts of hydroxyurea ( Sigma ) , formamide ( Sigma ) , or on plates lacking inositol . Plates were incubated at the indicated temperatures for 2–4 days . Whole cell extracts were prepared from logarithmic growing cells by glass bead lysis in the presence of trichloroacetic acid . Immunoblotting was carried out with 3E10 , 3E8 , 4E12 , 8WG16 ( Millipore ) , YN-18 ( Santa Cruz ) , Rpb3 ( Neoclone ) , HA-Peroxidase ( Roche ) and Pgk1 ( Molecular Probes ) antibodies [43] . Immunoblots were scanned with the Odyssey Infrared Imaging System ( Licor ) or visualized with SuperSignal enhanced chemiluminescence ( Pierce Chemical ) . RNA was extracted and purified using the Qiagen RNeasy Mini Kit . cDNA was generated using the Qiagen QuantiTect Reverse Transcription Kit . cDNA was analyzed by qPCR as described above . INO1 mRNA levels were normalized to ACT1 mRNA [7] . Samples were analyzed in triplicate from three independent RNA preparations . Overnight cultures were diluted to 0 . 3 OD600 and grown to 1 . 0 OD600 . 10 OD600 units were collected to constitute time 0 and a final concentration of 100 ug/ml of cycloheximide ( Sigma ) was added to the remaining culture . 10 OD600 units were collected at the indicated time points . Proteins were extracted using trichloroacetic acid . | RNA Polymerase II ( RNAPII ) is the enzyme responsible for the transcription of all protein-coding genes . It has a unique extended domain called the C-terminal domain ( CTD ) . This domain is highly conserved across species and is composed of repeats of a seven amino acid sequence . The CTD functions as a recruiting platform for regulatory and RNA processing factors , making the CTD a master orchestrator of transcription . Previous work revealed a critical role for CTD length in the transcription of induced genes . However , how CTD length is generally required for transcription is currently unclear , as is the mechanism underlying the observed suppression of CTD truncation phenotypes by loss of the SRB10/CDK8 gene . Here , using gene expression microarrays , we determined the set of genes most sensitive to alternations in CTD function and uncovered unexpected links between RNAPII-CTD and Cdk8 . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2013 | High-Throughput Genetic and Gene Expression Analysis of the RNAPII-CTD Reveals Unexpected Connections to SRB10/CDK8 |
Failure to establish an appropriate balance between pro- and anti-inflammatory immune responses is believed to contribute to pathogenesis of severe malaria . To determine whether this balance is maintained by classical regulatory T cells ( CD4+ FOXP3+ CD127−/low; Tregs ) we compared cellular responses between Gambian children ( n = 124 ) with severe Plasmodium falciparum malaria or uncomplicated malaria infections . Although no significant differences in Treg numbers or function were observed between the groups , Treg activity during acute disease was inversely correlated with malaria-specific memory responses detectable 28 days later . Thus , while Tregs may not regulate acute malarial inflammation , they may limit memory responses to levels that subsequently facilitate parasite clearance without causing immunopathology . Importantly , we identified a population of FOXP3− , CD45RO+ CD4+ T cells which coproduce IL-10 and IFN-γ . These cells are more prevalent in children with uncomplicated malaria than in those with severe disease , suggesting that they may be the regulators of acute malarial inflammation .
The clinical spectrum of P . falciparum infection ranges from asymptomatic parasite carriage to a febrile disease that may develop into a severe , life-threatening illness . The factors that determine disease severity are not completely understood but are likely to include both parasite and host components [1]–[3] . Ultimately , the interplay between the parasite and the immune response likely determines the outcome of the infection [4] . Although sterile immunity - completely preventing re-infection - is hardly ever seen and protection against clinical symptoms of uncomplicated disease is only acquired after repeated infections [5] , immunity to severe disease and death may be acquired after as few as one or two infections [6] suggesting that different immune mechanisms underlie these different levels of immunity . While there is a growing consensus that killing of malaria parasites or malaria-infected red blood cells requires the synergistic action of antibodies and cell-mediated immune responses [5] , [7] , the mechanisms conferring protection against severe disease are less clear . Given that pathology of severe disease has repeatedly been linked to sustained and/or excessive inflammatory responses [4] , acquiring the ability to regulate these responses adequately may be a key determinant of immunity that protects against severe disease [8] . Thus , while an early inflammatory response is needed to control parasite replication in human P . falciparum malaria [9]–[11] , excessive levels of pro-inflammatory cytokines such as TNF-α [12]–[15] , IFN-γ , [16] , [17] , IL-1β and IL-6 [14] , [18] , [19] are associated with severe pathology . Conversely , low levels of regulatory cytokines such as TGF-ß have been associated with acute [20] and severe malaria [21] , [22] , a relative deficiency in IL-10 was seen in those who succumbed to severe malaria [23] , significantly lower ratios of IL-10 to TNF-α were found in patients with severe malarial anaemia [24] , [25] , and high ratios of IFN-γ , TNF-α and IL-12 to TGF-β or IL-10 were associated with decreased risk of malaria but increased risk of clinical disease in those who became infected [26] . In summary , therefore , immunity against severe malaria may depend upon the host's ability to regulate the magnitude and timing of the cellular immune response , allowing the sequential induction of appropriate levels of inflammatory- and anti-inflammatory cytokines at key stages of the infection . Given these associations between severe disease and exacerbated immune pathology , a number of studies have explored the role of CD4+CD25hiFOXP3+CD127−/lo regulatory T cells ( Tregs ) in determining the outcome of malaria infection . Induced and/or activated in response to malaria infection [27] , Tregs may be beneficial to the host in the later part of the infection - when parasitaemia is being cleared - by down-regulating the inflammatory response and thereby preventing immune-mediated pathology . On the other hand , if Tregs mediate their suppressive effects too early , this could hamper the responses required for initial control of parasitaemia , permitting unbridled parasite growth which may also lead to severe disease . Malaria-specific induction of Tregs has been observed in a variety of experimental malaria infections in mice [28]–[31] , but their role in preventing severe malarial pathology is unclear . Thus , in BALB/c mice infected with a lethal strain of P . yoelii , ablation of Treg activity by depletion of CD25+ cells either allowed mice to control parasitaemia and survive [32] or had no impact on the course of disease [33] . Depleting CD25+ cells of BALB/c mice infected with either P . berghei NK65 [34] or P berghei ANKA [35] reduced neither parasitaemia nor mortality , but increased the severity of symptoms in the diseased mice , suggesting at least some benefit from Tregs in this model . Rather oddly , infection of CD25+ T cell-depleted BALB/c mice with P . chabaudi adami DS led to increased parasitaemia and more severe anaemia [30] . Finally , CD25+ T cell depletion around the time of parasite inoculation reduced the incidence of experimental cerebral malaria in C57BL/6 mice infected with P . berghei ANKA in two independent studies [29] , [31] , but not when CD25+ T cells were depleted 30 days prior to infection [31] . Whilst various explanations have been offered for these discrepant results , including differences in the various strains of mice and parasites employed , the microbial microenvironment in which the mice are kept , and the precise CD25 depletion protocols employed , these studies are currently not very helpful when trying to understand the role of Tregs in human malaria infections . Malaria naïve individuals undergoing experimental P . falciparum sporozoite infection showed an increase in FOXP3 mRNA expression and expansion of Tregs 10 days after infection; Treg induction correlated with high circulating levels of TGF-β , low levels of pro-inflammatory cytokines and rapid parasite growth [27] suggesting - but not proving - that Treg activation early in infection may inhibit the development of effective cellular immunity . More recently , we have observed that Treg populations appear to be transiently expanded and activated during the malaria transmission season in individuals from a malaria endemic community [36] , again suggesting that naturally acquired malaria infection can drive the expansion and activation of Tregs . However , although Tregs have been implicated in IL-10-mediated down-regulation of Th1-like responses in the placenta of malaria-infected women [37] and reduced Treg frequencies and function have been linked to enhanced anti-parasite immunity in certain ethnic groups in West Africa [38] the potential for Tregs to influence the clinical outcome of malaria infections is still unclear . To investigate the role of Tregs during clinical malaria infection , we have compared cellular immune responses of children with either severe or uncomplicated malaria . Interestingly , although we did not observe any significant differences in Treg numbers or function between severe and uncomplicated malaria cases , our data do indicate that malaria-induced Tregs may limit the magnitude of malaria-specific Th1 memory responses and thus moderate pro-inflammatory responses to subsequent infections , providing a possible explanation for the very rapid acquisition of immunity to severe malaria . Moreover , we have identified a population of FOXP3− , CD45RO+ , CD4+ T cells which co-produce IL-10 and IFN-γ and which are more prevalent in children with uncomplicated malaria than in those with severe disease . We suggest that these IL-10 producing effector T cells may contribute to clearing malaria infection without-inducing immune-mediated pathology .
We hypothesized that children with severe malaria would have fewer circulating Tregs than children with uncomplicated malaria , or that Tregs of severely ill children would be less active than those of children with uncomplicated disease . However , the proportion of cells expressing a Treg phenotype ( defined by flow cytometry as CD3+CD4+ lymphocytes being FOXP3+and CD127−/low; Figure 1A , 1B , and 1C ) was similar in the acute ( D0 ) and the convalescent phase ( D28 ) for both uncomplicated and severe cases ( SA+SB ) and in healthy control children; on average , 2–3% of CD4+ T cells expressed the regulatory phenotype ( Figure 1D ) . However , when the number of cells expressing a Treg phenotype was calculated using lymphocyte and monocyte counts from the differential WBC , we found that the absolute numbers of Tregs ( per litre of blood ) were significantly and similarly elevated in both severe and uncomplicated malaria cases during convalescence when compared to the acute phase ( p<0 . 001 ) or when compared to the control group of healthy children ( p = 0 . 037 ) ( Figure 1E ) . A similar kinetic was observed for FOXP3 mRNA levels ( Figure 1F ) . Although not supportive of our original hypothesis , these observations are consistent with the notion that acute malaria infection drives expansion of Treg populations which then persist for some weeks to maintain immune homeostasis during the contraction phase of the effector response [36] . In accordance with this notion , and in agreement with our previous observation that increased levels of Tregs were associated with faster parasite growth during the early stages of blood stage infection [27] , we observe here in - children with either severe or uncomplicated malaria infections - a significant positive correlation between parasite density and the frequency of Tregs within the CD4+ T cell population ( p = 0 . 002 , Figure 2 ) . Since fully differentiated Tregs predominantly express an activated/memory phenotype [44] T cells from children with severe and uncomplicated malaria were analysed for expression of CD45RO ( Figure 3A and 3B ) . In both uncomplicated and severe cases , the proportion of all T cells expressing CD45RO was significantly higher ( p<0 . 001 ) during acute infection than during convalescence ( data not shown ) . Irrespective of disease severity , more than 90% of Tregs expressed CD45RO during the acute phase of infection but expression of this marker decreased significantly ( to approx 70% ) during convalescence ( Figure 3C ) . Likewise , the median fluorescence intensity ( MFI ) of CD45RO staining was 1 . 5 fold higher ( p = 0 . 0025 ) during acute disease than during convalescence ( Figure 3D ) . Taken together , these data indicate that in both uncomplicated and severe cases of malaria Tregs are predominantly of a memory phenotype and are activated during acute malaria infection . Three different indicators were used to assess the regulatory potential of Tregs during acute malaria infection . Firstly , using a classical anti-CD25 depletion assay , we assessed the ability of Tregs to suppress P . falciparum shizont extract ( PfSE ) -driven lymphocyte proliferation . Anti CD25 treatment removed approximately half ( geometric mean 48 . 8%; CI95%: 41–58% ) of the CD4+ T cells that were FOXP3+CD127−/low and this was associated with a 1 . 76 fold and 1 . 57-fold ( geometric means ) increase in PfSE-induced lymphoproliferation in severe and uncomplicated cases , respectively , with no significant difference between the groups ( p = 0 . 343 , Figure 4A ) . Next , since reduced expression of SOCS-2 , a member of the suppressors of cytokine signaling family confined to Tregs [45] , has been linked to impaired Treg function in Africans [38] we compared SOCS-2 mRNA levels among severe and uncomplicated cases . While SOCS-2 levels were found to be significantly reduced during acute disease compared to convalescence ( p = 0 . 0076 ) , no difference was observed between those with severe and those with uncomplicated malaria ( data not shown ) . Finally , since high concentrations of TNF-α have been reported to impair Treg activity ( by upregulating and then signaling via TNFR2 , leading to decreased FOXP3 mRNA and protein expression [46] ) , and the functional impairment of Tregs observed in rheumatoid arthritis patients can be reversed by anti-TNF-α antibodies [47] , we considered the hypothesis that the high levels of TNF-α seen in severe malaria patients [13] , [15] , might upregulate TNFR2 and impair Treg function . TNFR2 expression on Tregs was assessed by flow cytometry ( Figure 4B and 4C ) . However , although a significantly higher proportion of Tregs expressed TNFR2 ( Figure 4D ) - with higher MFI ( data not shown , p = 0 . 028 ) - during acute disease compared to convalescence , no difference was seen in TNFR2 expression on Tregs between severe and uncomplicated cases ( Figure 4D ) . Moreover , there was no correlation between plasma levels of TNF-α and TNFR2 expression on Tregs , neither among severe or uncomplicated cases nor among all cases combined; neither did we observe any inverse correlation between TNF-α concentration and FOXP3 expression . Rather , the MFI of TNFR2 on Tregs was positively correlated with the MFI of FOXP3 in Tregs ( r: 0 . 476; p<0 . 0001 ) . Thus , our data seem to be more in line with data from mice suggesting that the interaction of TNF-α with TNFR2 on Tregs promotes their expansion and upregulation of FOXP3 [48] than with the data from studies of human rheumatoid arthritis . Since the balance of T-effector to Treg responses is likely to be as important , or more important , than the absolute levels of either [26] , we compared the ratio of the levels of mRNA for the Th1 transcription factor T-BET with those for FOXP3 , currently considered the best marker for Tregs , and the ratio of T-effector cells ( defined as CD3+CD4+CD25+FOXP3− , T-effector ) over Tregs among the various groups . As shown in Figure 5A , in all groups the T-BET/FOXP3 ratio was significantly higher during acute disease than during convalescence and a similar , albeit not significant , trend was observed for the ratio of T-effector/Tregs ( data not shown ) . Moreover , since the absolute number of circulating T-effector cells was significantly higher in severe cases than in uncomplicated cases ( p = 0 . 01 , data not shown ) , the T-effector/Treg ratio tended to be higher among severe cases than uncomplicated cases on day 0 ( p = 0 . 039 ) and a similar trend was seen for the T-BET/FOXP3 ratio ( p = 0 . 058 ) . The ratio of FOXP3 to GATA-3 ( Th2 lineage factor ) mRNA was similar for both time points in all groups ( data not shown ) but the Th1/Th2 ratio ( T-BET/GATA-3 mRNA ) was significantly higher during acute disease in children with CM , SA or SRD compared to those suffering from severe prostration ( p = 0 . 0075 ) , indicating that the expansion of the T-effector population is biased towards Th1 responses . These data confirm previous studies indicating a shift towards a more inflammatory response during acute and severe malaria but , significantly , our data extend the previous observations by revealing that this inflammation is not balanced by a commensurate increase in Treg function . Indeed , our data strongly suggest that the potent inflammation induced during an acute malaria infection overwhelms the normal homeostatic capacity of the immune system and , in particular , that the Treg response in children with severe malaria is insufficient to balance a much stronger Th1 effector response . To investigate further the dynamics of pro-inflammatory/regulatory responses during clinical malaria , plasma concentrations and mRNA transcripts of inflammatory ( IFN-γ , TNF-α ) and regulatory cytokines ( IL-10 ) were assayed . In accordance with previous observations , plasma concentrations of IFN-γ , TNF-α and IL-10 were all significantly higher during acute disease than during convalescence , with significantly higher levels in severely ill children compared to uncomplicated cases ( Figure 5B , 5C , and 5D ) . Levels of mRNA transcripts for IL-10 and IFN-γ were also significantly elevated in all groups during acute disease , but there was no significant difference between severity groups ( Figure S1A and S1B ) . For both severe and uncomplicated cases , levels of IFN-γ mRNA were highly correlated with levels of IL-10 mRNA during the acute phase ( severe: r = 0 . 833 p<0 . 001 , uncomplicated: r = 0 . 693 p<0 . 001 ) , suggesting that IFN-γ production is being balanced by IL-10 production . Interestingly , IFN-γ mRNA levels on day 0 correlated with FOXP3 mRNA on day 0 for both severe ( r: 0 . 39 p = 0 . 003 ) and uncomplicated cases ( r: 0 . 44 p = 0 . 0001 ) , suggesting that IFN-γ may also be driving FOXP3 expression . The balance of pro-and anti inflammatory cytokine responses clearly changed with time , but somewhat surprisingly , there were no marked differences in cytokines ratios between children with differing levels of disease severity . Thus , the ratios of TNF-α or IFN-γ to IL-10 on day 0 were similar in all disease severity groups ( Figure 5E and 5F ) and ratios of IFN-γ to IL-10 mRNA were similar in all disease severity groups both on day 0 and day 28 ( Figure S1F ) . However , IFN-γ mRNA levels were on average only 3 . 2-fold higher on day 0 than day 28 but IL-10 mRNA levels were 29-fold higher on day 0 , resulting in a significantly lower IFN-γ/IL-10 mRNA ratio on day 0 than day 28 ( Figure S1F ) . IL-10 is a crucial immunoregulatory cytokine in both human [23] and murine [33] , [49] malaria; we have recently identified CD4+ T effector cells as a major source of IL-10 [33] , but the source of IL-10 in human malaria infection is unknown . In other protozoal infections of mice CD4+ effector T cells that co-produce IFN-γ and IL-10 have been identified [50]–[52] . We therefore cultured freshly isolated PBMCs from 30 children with acute malaria ( 17 severe , 13 uncomplicated ) and 20 healthy control children , with or without PMA and Ionomycin ( PI ) , for 5 hours and analyzed them for the presence of intracellular IL-10 and IFN-γ by flow cytometry ( Figure 6A ) . No cytokine production was observed in unstimulated cells ( data not shown ) , and PBMC from healthy children failed to produce any IL-10 in response to PI ( data not shown ) , indicating that stimulation with PI predominantly induces cytokine production from recently activated cells . By contrast , distinct populations of IL-10+ and IFN-γ+ cells were seen among the PI-stimulated cells from children with acute severe or uncomplicated malaria , with a small but easily distinguishable population of cells ( approx 1% of all PBMC ) producing both cytokines simultaneously ( Figure 6A , right plot ) . In both severe and uncomplicated cases , IL-10 producing cells were predominantly CD45RO+ CD4+ T cells ( Figure 6B and 6C ) and were almost exclusively FOXP3− and CD25− ( Figure 6D ) . Moreover , although a transient increase of FOXP3 in activated human T-effector cells has been reported [53] , in our hands less than 1% ( median 0 . 97% , CI95%: 0 . 67–1 . 27% ) of IFN-γ producing cells were FOXP3+ ( Figure 6E ) . Overall , among children with acute malaria , approx 4% of PI-stimulated CD4+ T cells produced IL-10 and approx 8% produced IFN-γ and neither the proportions of cells producing one or the other cytokine ( Figure 6F ) nor the ratio of IFN-γ/IL-10 producing cells ( data not shown ) differed significantly between severe and uncomplicated cases . However , intriguingly , the proportion of CD4+ T cells simultaneously producing IL-10 and IFN-γ was three fold higher in uncomplicated cases than severe cases ( geometric mean 5 . 2% vs 1 . 6% , p = 0 . 041 , Figure 6F ) . Moreover , the proportion of IFN-γ+ CD4+ T cells that also produce IL-10 was almost twice as high among uncomplicated cases as among severe cases ( p = 0 . 045 , Figure 6G ) . Taken together , these data indicate that during acute , uncomplicated or severe , malaria infections IL-10 producing cells are overwhelmingly T effector cells and that Th1 effector cells that also produce IL-10 are more prevalent in children with uncomplicated malaria than in children with severe malaria . It has been reported that Tregs present at the time of infection [54] or vaccination [55] may restrict the development of subsequent Th1 memory responses . To determine whether Tregs present during acute malaria infection might similarly affect the induction of immunological memory , we compared FOXP3 mRNA levels on day 0 with malaria specific IFN-γ memory responses ( as assessed by PfSE-specific cultured ELISPOT ) among PBMC collected from 34 of our convalescent malaria patients ( 19 severe and 15 uncomplicated ) on Day 28 . As shown in Figure 7A , cells from uncomplicated and severe cases mounted similarly strong IFN-γ memory responses following culture with PfSE . When plotted against FOXP3 mRNA levels measured on day 0 , a linear by linear hyperbolic fit revealed that higher levels of FOXP3 mRNA on day 0 were highly significantly ( p = 0 . 009 ) associated with lower malaria-specific IFN-γ memory responses on Day 28 , suggesting that Tregs induced during the acute infection may limit the magnitude of subsequent Th1 responses ( Figure 7B ) . For neither group could a significant effect of parasitaemia on the memory response be observed ( r = −0 . 12 p = 0 . 962 for severe and r = 0 . 015 , p = 0 . 957 for uncomplicated cases ) .
We hypothesized that the balance of inflammatory to regulatory immune responses would be biased towards a more inflammatory response in children with severe malaria than in children with uncomplicated malaria , that this balance would be restored during convalescence and – crucially – that this would be associated with differences in the proportion , absolute number or function of circulating classical ( CD4+ CD127−/lo FOXP3+ ) regulatory T cells . In partial support of these hypotheses , the number of cells expressing a Treg phenotype and FOXP3-mRNA levels were both significantly higher during convalescence than during the acute clinical episode and the ratio of the Th1 transcription factor T-BET to the Treg transcription factor FOXP3 was significantly higher during acute disease than during convalescence in both severe and uncomplicated cases , compatible with the notion that Tregs fail to sufficiently regulate pro-inflammatory responses which might contribute to the onset of symptomatic malaria infection . Given our previous observation of Treg expansion during the pre-patent phase of malaria infection [27] , we suggest that Tregs are induced/activated shortly after parasite emergence from the liver , that their numbers in peripheral blood then decline as a result of sequestration of CD4+ T cells during acute disease [40] , [41] , [56] and then , as has been described for other T cell subsets [39] , [57] , Tregs regain access to the circulation after malaria is cured . The significant positive correlation of Treg numbers with parasitaemia , as well as the correlation between FOXP3 mRNA and IFN-γ mRNA levels in acute samples , further supports the notion that the initial infection induces a proportional increase in Tregs , attempting to balance the effector T cell response , and is in line with the recently proposed concept that antigenic challenge will give rise to an antigen specific Treg response , proportional in size to the inflammatory response [58] . Moreover , the Tregs circulating during acute malaria infections almost exclusively expressed an activated memory phenotype suggesting that they have expanded from a pre-existing pool of memory T-cells . This interpretation would be in line with recent elegant work in humans demonstrating that Tregs are derived by rapid turnover of memory populations in vivo [59] , and with data from murine studies where , after CD25-depletion , malaria infection very rapidly drives differentiation of Tregs from circulating mature CD4+ T cells [60] . Obviously , it would be of interest to study the relationship between Tregs and effector T cell kinetics and parasite biomass , which is not readily measurable . Future studies may explore the usefulness of P . falciparum Histidine Rich Protein 2 in this context , which has recently been suggested as a surrogate marker for parasite biomass [61] . However , despite clear evidence of Treg induction and reallocation during acute malaria infection , we could not find any robust differences in Treg parameters between children with severe and uncomplicated disease . Thus , neither Treg numbers nor FOXP3 mRNA levels differed significantly between children with uncomplicated malaria and those with severe malaria , and three different indicators of Treg function - their capacity to suppress lymphoproliferation , their expression of SOCS-2 [45] and TNFR2 [46] , [48] were all similar in severely ill children and children with uncomplicated disease . Furthermore , the similar distortion in the T-BET/FOXP3 mRNA ratio during acute disease and the lack of any marked differences between the two groups in ratios of inflammatory to anti-inflammatory cytokines , as well as the close correlation between IFN-γ and IL-10 in both groups which is in line with previous observations in experimental human malaria infections [62] , suggests that the systemic shift towards a pro-inflammatory immune response is similar in children with either severe or uncomplicated disease . At first glance , these data do not appear to support the hypothesis that deficiencies in Treg function underlie the tendency of some children to develop severe , life threatening malaria . However , we did observe significantly higher Th1 effector responses ( more T-effector cells , higher concentrations of IFN-γ and TNF-α ) in severely ill children than in children with uncomplicated disease , suggesting that the classical FOXP3+ Treg response that develops during acute malaria infection may be insufficient to balance the florid effector T cell response that develops particularly in children with severe disease . This would be in line with evidence showing that as the strength of the inflammatory stimulus increases , the suppressive capacity of human Tregs declines and the resistance of T-effector cells to regulation increases [63] . The situation observed during an acute , clinical malaria infection is thus in clear contrast to the situation in healthy , malaria-exposed individuals where Treg numbers closely track numbers of T-effectors , precisely maintaining an apparently optimal T-effector∶Treg ratio [36] . IL-10 is well-established as a vital homeostatic regulator of malaria-induced inflammation that prevents immune-pathology in mice [49] , [64] , promotes the necessary switch from early Th1 to subsequent Th2 responses [65] , [66] , and has been linked to protection from severe malaria anaemia [24] , [25] , and death [23] in humans . However , the cellular source of IL-10 in human malaria cases was , until now , ill defined . Contrary to our expectations , but in striking agreement with observations in P . yoelii-infected mice [33] , CD45RO+ CD4+ T cells ( that are CD25− and FOXP3− ) and not classical Tregs are the only substantial source of IL-10 during acute malaria infection . This observation is reminiscent of that made by Nylen [67] in patients with acute visceral leishmaniasis . Moreover , in our patients , a significant proportion of IL-10 producing CD4+ T cells were simultaneously producing IFN-γ , identifying them as Th1 cells . Although IL-10 secreting Th1 cells have been described recently in two murine models of toxoplasmosis [50] , and cutaneous leishmaniasis [51] , as far as we are aware , this is the first demonstration of IL-10 producing Th1 cells during human infections . Intriguingly , the proportion of these cells within the total CD4+ T cell population was significantly higher in children with uncomplicated malaria than in children with severe malaria suggesting that in human P . falciparum infection , as in murine T . gondii infections [50] , IL-10 producing Th1 cells , activated by a strong inflammatory stimulus , may act as anti-parasitic effector cells with a “built in” control mechanism to prevent the onset of immune pathology . If so , then the ability of these self-regulating effector cells to localize to sites of parasite sequestration in tissues , where they mediate parasite killing whilst simultaneously blocking tissue damage , may be key to clinical immunity to malaria . Thus , our data strongly suggest that the percentage of IL-10-producing Th1 effector cells , rather than the cocktail of circulating cytokines , may be the most relevant biomarker of effective immunity to severe malaria . Although Tregs may not seem to determine the outcome of current P . falciparum infections we did find evidence that they affect the magnitude of the malaria specific memory response induced by the current infection . A similar observation has been made in P . berghei ANKA-infected mice; animals that were depleted of CD25+ cells prior to infection and drug-cured on day 5 developed significantly stronger IFN-γ memory responses on day 14 than did intact infected/cured mice , and these mice also developed much more severe , and frequently fatal , clinical symptoms upon reinfection , despite more efficient parasite clearance [35] . Thus , malaria specific Tregs acquired during a primary infection may limit the magnitude of Th1 effector responses to subsequent infections to a level that allows parasite clearance without causing immunopathology . Future studies should be designed to test the hypothesis that Tregs may contribute to the very rapid development of resistance to severe malaria . In summary , our data indicate that classical FOXP3+ Tregs are unable to control the florid inflammation that accompanies acute malaria infections and this component of the immunoregulatory arsenal is rapidly overwhelmed in children with either mild or severe malaria . Importantly however we have identified , for the first time in an acute human infection , a population of IL-10 producing Th1 effector cells which appear to be a major source of this key anti-inflammatory cytokine during acute malaria infection , and which are associated with development of uncomplicated as opposed to severe malaria . We propose that IL-10-producing Th1 cells may be the essential regulators of acute infection-induced inflammation and that such “self-regulating” Th1 cells may be essential for the infection to be cleared without inducing immune-mediated pathology . Moreover , we have found evidence in support of the hypothesis that Tregs limit the magnitude of the Th1 memory response raising the intriguing possibility that they may play an important role in the rapid evolution of clinical immunity to severe malaria .
A case-control study was conducted in Gambian children with severe or uncomplicated malaria , resident in a peri-urban area within a 40 km radius south of the capital , Banjul , with low levels of malaria transmission [68] , [69] . Patients were enrolled at Brikama Health Centre , the MRC Fajara Gate Clinic or the Jammeh Foundation for Peace Hospital in Serekunda between September 2007 and January 2008 , after written informed consent was obtained from the parents or guardians . Uncomplicated disease was defined as an episode of fever ( temperature >37 . 5°C ) within the last 48 hours with more than 5000 parasites/µl detected by slide microscopy . Severe disease was defined using modified WHO criteria [70]: SA , defined as Hb<6 g/dl; SRD defined as serum lactate >7 mmol/L; CM defined as a Blantyre coma score ≤2 in the absence of hypoglycaemia , with the coma lasting at least for 2 hours . To avoid the confounding effects of other pathogens in children with concomitant systemic bacterial infections [71] , children with clinical and/or laboratory evidence of infections other than malaria were not enrolled into the study . For some experiments , healthy children of the same age and recruited from the same area at the same time of the year were enrolled as controls . In total , 59 severe , 65 uncomplicated and 20 control cases were enrolled . On admission ( D0 ) and after 4 weeks ( D28±3 days ) one ml of blood was collected in RNA stabilizing agent ( PAXgene™ Blood RNA system , Pre-AnalytiX ) and a maximum of 4 mls of blood ( mean: 3 . 2 mls CI 95%: 3 . 1–3 . 3 mls ) were collected into heparinized vacutainers® ( BD ) . All patients received standard care according to the Gambian Government Treatment Guidelines , provided by the health centre staff . The children's health was reviewed 7 days after admission . The study was reviewed and approved by the Joint Gambian Government/MRC Ethics Committee and the Ethics Committee of the London School of Hygiene & Tropical Medicine ( London , UK ) . P . falciparum parasites were identified by slide microscopy of 50 high power fields of a thick film . Full differential blood counts were obtained on days 0 and 28 using a Medonic™ instrument ( Clinical Diagnostics Solutions , Inc ) ; the presence of intestinal helminths was assessed by microscopy from stool samples collected into BioSepar ParasiTrap® diagnosis system , following the manufacturers' instructions . Sickle cell status was determined by metabisulfite test and confirmed on cellulose acetate electrophoresis [72] . Blood samples were processed within 2 hours of collection . Plasma was removed , stored at −80°C and replaced by an equal volume of RPMI 1640 ( Sigma-Aldrich ) . PBMC were isolated after density centrifugation over a 1 . 077 Nycoprep ( Nycomed , Sweden ) gradient ( 800 g , 30 min ) and washed twice in RPM 1640 . Cells were either stained for flow cytometry directly ex-vivo , or cultured in RPMI 1640 containing 10% human AB+ serum , 100 µg/ml streptomycin , 100 U/ml penicillin ( all Sigma-Aldrich ) , and 2 mM L-glutamine ( Invitrogen Life Technologies ) , referred to as complete growth medium ( GM ) . Fresh PBMC were stained using the following fluorochrome labeled mouse or rat anti-human antibodies: FITC anti-TNF-receptor II ( R&D ) , PE anti-FOXP3 ( clone PCH101 ) , Pacific Blue anti CD3 , APC-Alexa Fluor 750 anti-CD127 ( all Ebioscience ) , APC anti-CD25 , PerCP anti CD4 ( BD systems ) , ECD anti-CD45R0 ( Beckman-Coulter ) , and appropriate isotype controls . IL-10 and IFN-γ production by PBMC from 30 children with acute P . falciparum ( D0 ) was assessed after 5 hours stimulation in GM containing PMA ( 50 ng/ml ) and Ionomycin ( 1000 ng/ml ) or GM alone . Cells were stained with FITC anti-IFN-γ , PE anti-FOXP3 , PE-Cy7 anti-CD25 , APC-AF750 anti-CD8 , Pacific Blue anti-CD3 ( all Ebioscience ) , PerCP anti-CD4 , APC anti-IL-10 ( both BD ) , and ECD anti-CD45R0 ( Beckman-Coulter ) . To ascertain specificity of the intracellular cytokine staining , aliquots of some samples were incubated with saturating amounts of purified non-labelled antibody of the same clone prior to staining with the fluorochrome labeled ICS antibody . The FOXP3 staining buffer set ( Ebioscience ) was used following the manufacturer's protocol . Samples were acquired on a 3 laser/9 channel CyAn™ ADP flowcytometer using Summit 4 . 3 software ( Dako ) . Analysis was performed using FlowJo ( Tree Star Inc . ) . All flowcytometric analysis was performed at the MRC laboratories , The Gambia on freshly isolated cells . Plasma concentrations of IFN-γ , TNF-α and IL-10 were determined for each subject and time point on the Bio-Plex® 200 system , using X-Plex™ assays ( both Bio-Rad Laboratories ) , according to the manufacturer's instructions . Data were analysed using the Bio-Plex® Manager software . The detection limit was defined as the concentration corresponding to a fluorescence value above the mean background fluorescence in control wells plus 3 SD , being 8 . 76 pg/ml for IFN-γ , 5 pg/ml for TNF-α and 0 . 57 pg/ml for IL-10 . Values below this threshold were set to these levels . P . falciparum parasites ( 3D7 strain ) were cultured in vitro as described [73] and were routinely shown to be mycoplasma free by PCR ( Bio Whittaker ) . Schizont-infected erythrocytes were harvested from synchronized cultures by centrifugation through a Percoll gradient ( Sigma-Aldrich ) . PfSE was prepared by two rapid freeze-thaw cycles in liquid nitrogen and a 37°C water bath . Extracts of uninfected erythrocytes ( uRBC ) were prepared in the same way . PBMC from 10 severe and 10 uncomplicated malaria cases collected on day 0 were depleted of CD25hi cells or mock depleted using magnetic beads ( Dynal Biotech , UK ) , at a bead to PBMC ratio of 7∶1 , and cell proliferation was determined by [3H]-thymidine ( Amersham , UK ) incorporation after 6 days in culture with PfSE , uRBC ( RBC∶PBMC ratio equivalent to 2∶1 ) , GM , or 2 days culture with PMA ( 10 ng/ml ) + Ionomycin ( 100 ng/ml ) , as described [27] . Cultured ELISPOTs were performed to assess malaria specific IFN-γ memory responses , adapting an established method [74] . Up to 1 million PBMCs collected on day 28 were cultured in 24 well plates for 6 days in 1 ml GM and stimulated with either PfSE , uRBC ( RBC∶PBMC ratio equivalent to 2∶1 ) , or GM respectively . At day 3 , half the medium was exchanged and rIL-2 ( final concentration 20 IU/well ) was added . On day 6 cells were harvested , washed three times , and 1 . 5×105 cells seeded into duplicate wells onto Millipore MAIP S45 plates and restimulated overnight with PfSE , uRBC ( concentrations as above ) , GM or PHA-L ( 5 µg/ml ) . IFN-γ ELISpot was performed using MabTech antibodies according to the manufacturer's instructions . Spot forming cell numbers were counted using an ELISPOT plate reader ( AutoImmuneDiagnostica , Vers . 3 . 2 ) . Results are expressed as spot forming units ( SFU ) per million PBMC after subtraction of individual background values ( GM for PHA-L , uRBC for PfSE ) being deducted . Assays were discounted if the positive control ( PHA-L ) was <50 SFU , or the negative control was >30 SFU . For quantitative reverse transcription-polymerase chain reaction ( RT-PCR ) , total RNA was extracted from PAX tubes following the manufacturer's instructions and reverse transcribed into cDNA using TaqMan® reagents for reverse transcription ( Applied Biosystems ) , following the manufacturer's protocol . Gene expression profiles for FOXP3 , IL-10 , SOCS-2 and IFN-γ were measured by RT-PCR on a DNA Engine Opticon® ( MJ Research ) with QuantiTect SYBR Green PCR kits ( Qiagen Ltd ) using primers ( all Sigma Genosys ) previously described: IFN-γ , IL-10; FOXP3 designed by [75] , and SOCS-2 designed by [76] . T-BET and GATA-3 gene expression was determined using the TaqMan® Probe kit using the primers ( all Metabion ) designed by [77] . 18S rRNA , amplified using a commercially available kit ( rRNA primers and VIC labeled probe , Applied Biosystems ) , was used as an internal control . Data were analysed using Opticon Monitor 3™ analysis software ( BioRad ) and are expressed as the ratio of the transcript number of the gene of interest over the endogenous control , 18S rRNA . Genomic DNA from each parasite isolate was genotyped by sequencing the highly polymorphic block 2 region of the msp1 gene to assess the number of clones infecting each patient [78] . Analysis was performed using linear regression , with a random effect to allow for the within subject measurements over time , where the response variables were log transformed to improve the normality and constant variance assumptions . Significance ( measured at the 5% level ) tests for the effects of malaria group ( uncomplicated , SA or SB ) , time ( day 0 and day 28 ) and their interaction were adjusted for the possible confounding effects of age , sex , duration of prior symptoms and numbers of clones causing the infection . Where there was no significant malaria group and time interaction , p-values for the overall comparison of day 0 vs . day 28 are given . Within day 0 , comparisons of severe vs . uncomplicated and the two groups of severely ill patients ( SA vs SB ) were adjusted for any malaria group and time interactions . To allow for the multiplicity of tests resulting from multiple responses and multiple comparisons within a response performed in the model , a false discovery rate ( FDR ) of 5% was assumed . Using the Benjamini and Hochberg approach [43] only tests with a p-value below 0 . 012 have an FDR of ≤5% . Due to the large number of tests family-wise error rate correction methods were too conservative . Analyses were performed using Stata version 9 and Matlab version R2008a . | While Tregs have been implicated in regulation of the immune response to chronic infections , their potential in determining disease outcome in acute infections is unclear . In this study we have found that Tregs are unable to control the florid inflammation during acute , severe P . falciparum malaria infections , suggesting that this component of the immunoregulatory arsenal may be rapidly overwhelmed by virulent infections . Further , we identified , for the first time in an acute human infection , a population of IL-10-producing Th-1 effector cells and found that IL-10-producing Th-1 cells were associated with development of uncomplicated as opposed to severe malaria , leading us to suggest that such “self-regulating” Th-1 cells may contribute to clearing malaria infections without inducing immune-mediated pathology . In addition , we found evidence that malaria-induced Tregs may limit the magnitude of malaria-specific memory responses detectable 28 days later , which may reduce the risk of immune-mediated pathology upon reinfection and may explain how immunity to severe disease can be gained after as little as one or two infections . We conclude that vaccines designed to induce cell-mediated responses should be assessed for their ability to induce IL-10 producing Th-1 cells and Tregs . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"immunology/immunomodulation",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"immunology/immunity",
"to",
"infections",
"infectious",
"diseases/protozoal",
"infections"
]
| 2009 | Distinct Roles for FOXP3+ and FOXP3− CD4+ T Cells in Regulating Cellular Immunity to Uncomplicated and Severe Plasmodium falciparum Malaria |
Exome sequencing is becoming a standard tool for mapping Mendelian disease-causing ( or pathogenic ) non-synonymous single nucleotide variants ( nsSNVs ) . Minor allele frequency ( MAF ) filtering approach and functional prediction methods are commonly used to identify candidate pathogenic mutations in these studies . Combining multiple functional prediction methods may increase accuracy in prediction . Here , we propose to use a logit model to combine multiple prediction methods and compute an unbiased probability of a rare variant being pathogenic . Also , for the first time we assess the predictive power of seven prediction methods ( including SIFT , PolyPhen2 , CONDEL , and logit ) in predicting pathogenic nsSNVs from other rare variants , which reflects the situation after MAF filtering is done in exome-sequencing studies . We found that a logit model combining all or some original prediction methods outperforms other methods examined , but is unable to discriminate between autosomal dominant and autosomal recessive disease mutations . Finally , based on the predictions of the logit model , we estimate that an individual has around 5% of rare nsSNVs that are pathogenic and carries ∼22 pathogenic derived alleles at least , which if made homozygous by consanguineous marriages may lead to recessive diseases .
Since the first successful application of exome sequencing in finding the causal mutation for a Mendelian disease [1] , many such studies have been conducted to identify other Mendelian disease-causing ( or pathogenic ) variants . Compared to genetic linkage studies of Mendelian diseases , exome sequencing requires a smaller number of affected individuals , who may even be unrelated . With decreasing sequencing costs , exome sequencing is becoming a standard tool for mapping causal genes for human Mendelian diseases . The most common cause of Mendelian disease is a non-synonymous single-nucleotide variant ( nsSNV ) that results in a single amino acid change in the encoded protein [2] . With the large number ( typically around 8 , 000–10 , 000 ) of nsSNVs in an individual genome and the small number of ( usually affected and unrelated ) individuals available , standard methods for genetic linkage and association do not work for exome sequencing studies of Mendelian diseases . In order to narrow down the list of candidate nsSNVs , most exome sequencing studies rely on a hard-filtering approach , in which the causal mutation is assumed to be rare ( with minor allele frequency ( MAF ) ≤1% ) and so polymorphisms ( with MAF>1% ) found in public databases ( e . g . , dbSNP and 1000 Genomes Project ) as well as in-house control datasets are discarded [3] . Moreover , variants are rejected if they are not found in multiple cases or if they conflicts with the known disease inheritance mode . This approach has successfully reduced the number of mutations to look at in numerous studies , and several tools [3]–[5] are therefore developed to automate this process . However , hard-filtering in exome sequencing of Mendelian diseases still leaves a large number ( typically ∼100 to 1 , 000 ) of candidate nsSNVs . A method must , therefore , be used to predict which of the remaining ones have serious functional consequences and prioritize them for validation . For a comprehensive review of these methods , see Ng and Henikoff [6] . These different methods have their complementary strengths and combining multiple methods has been suggested to increase prediction accuracy [3] , [6] . Recently a combined predictive model ( known as CONDEL [7] ) has been developed . CONDEL is based on a Weighted Average of the normalized Scores ( WAS ) [7] for combining scores from different algorithms and is available in Ensembl's Variant Effect Predictor . Another combined model ( known as CAROL ) is based on a weighted Z method of each individual score [8] . As hard filtering is usually done before prediction methods are applied in exome sequencing studies of Mendelian diseases [1] , [3] , it is important for prediction methods to distinguish pathogenic nsSNVs from other rare variants . However , a number of individual methods ( including PolyPhen2 [9] ) and both CONDEL and CAROL only use common variants as negative controls for assessing their predictive performance and determining their optimal cut-offs for variant classification . Since rare and common variants in the human genome have clearly distinct properties [10] , we argue that such benchmarking may not be appropriate to exome sequencing studies . In this paper , we first proposed the use of a logit model to combine prediction scores from multiple methods and tailored it to compute an unbiased estimate of the probability of a rare nsSNV being pathogenic . Then , we assessed the performance of five popular prediction methods ( HumVar-trained PolyPhen2 , SIFT [11] , LRT [12] , MutationTaster [13] , PhyloP [14] ) and two combined models ( CONDEL and logit ) in distinguishing pathogenic nsSNVs from other rare variants . As a comparison , we also examined the predictive powers of these methods in discriminating between pathogenic and common nsSNVs using HumVar [9] as a benchmark dataset . In addition , we saw if these prediction methods could discriminate between autosomal dominant and autosomal recessive disease mutations . Furthermore , we estimated the proportion of pathogenic rare variants and total load of pathogenic derived alleles an individual carries using high coverage exome sequencing data from the HapMap project . Finally , we applied the logit prediction model to three in-house exome sequencing subjects to demonstrate its performance in real data . For clarity , throughout this paper , being deleterious means that an nsSNV is under purifying selection; being damaging means that an nsSNV leads to a loss of protein function; and being pathogenic means that an nsSNV has an effect on a Mendelian disease phenotype .
Figure 1 shows the Receiver Operating Characteristic ( ROC ) and Precision-Recall ( PR ) curves of the five individual methods ( HumVar-trained PolyPhen2 [10] , SIFT [11] , LRT [12] , MutationTaster [13] , PhyloP [14] , see a description of each method in Table S1 ) and two combined methods ( the proposed logit model and CONDEL [7] , based on the scores from all five individual methods ) evaluated on ExoVar ( a dataset composed of pathogenic nsSNVs and nearly non-pathogenic rare nsSNVs ) using a 10-fold cross-validation ( see Materials and Methods ) . The logit model clearly outperforms all other methods in terms of the Areas Under the Curve ( AUCs ) of ROC and PR . The averaged maximal Matthews correlation coefficient ( MCC ) [15] of the logit model and CONDEL in a 10-fold cross-validation are 0 . 615 and 0 . 558 respectively . Contrary to our intuition , we found that the predictive power of CONDEL combining five individual methods is slightly inferior to that of one individual method ( i . e . , MutationTaster ) in terms of ROC AUC , although it is better than those of all five individual methods in terms of PR AUC . Among the five individual prediction methods , MutationTaster , which considered multiple resources such as evolutionary conservation , splice-site changes and loss of protein features , outperforms the others in classifying pathogenic variants in this dataset . In contrast , PhyloP , which considers only evolutionary conservation , has the smallest AUCs , indicating that information other than evolutionary conservation is also important in classifying pathogenic nsSNVs . Nonetheless , these results suggest that the logit model is able to take advantage of the complementarily between predictions of different individual methods ( which are only weakly and moderately correlated , see Figure S1 ) to achieve a better prediction power . In addition , we investigated whether combining a subset of the five individual methods in the logit model has similar predictive power compared to combining all five methods using the same validation procedure . Interestingly , we found that some reduced models have similar or slightly better predictive power than the full model ( Figure 2 and Table S2 ) . Among all possible combinations , a logit model using the scores from SIFT , Polyphen2 and MutationTaster performs the best , in terms of ROC AUC and that using the scores from SIFT and MutationTaster only has similar performance to the full model , in terms of PR AUC ( so later , we will use this model in estimating the proportion of pathogenic rare nsSNVs and the total load of pathogenic derived alleles per individual as it requires non-missing scores for two methods only ) . In contrast , combining PhyloP and LRT , which do not incorporate any protein specific features for prediction , has the worst performance among all possible combinations . We also examined the predictive powers of the methods in discriminating between pathogenic and common nsSNVs evaluated on HumVar ( a popular benchmark dataset composed of pathogenic and common nsSNVs ) using a 10-fold cross-validation . Of note , the HumVar dataset was also used by PolyPhen2 and CONDEL to benchmark their prediction models for pathogenic nsSNVs . Figure 3 shows the ROC and PR curves . All combined models outperform the individual methods in terms of ROC and PR AUC , but the predictive power of a logit model is still better than that of CONDEL . The averaged maximal MCC of the logit model and CONDEL in a 10-fold cross-validation are 0 . 664 and 0 . 616 respectively . More importantly , all methods have less predictive power , in terms of ROC and PR AUC , in distinguishing pathogenic nsSNVs from other rare nsSNVs than from common nsSNVs . Figure 4 shows the ROC and PR curves of the methods evaluated on a dataset composed of autosomal dominant and autosomal recessive disease-causing nsSNVs , DomRec , using a 3-fold cross-validation . Although there is a significant difference in prediction scores between the two classess of disease mutations ( with the largest difference observed in PhyloP , Mann-Whitney U test p-value = 4 . 56×10−4 ) ( see Table 1 ) , no method can confidently discriminate between the two classes of mutations ( ROC and PR AUCs of all methods ≈0 . 5 , i . e . , random prediction ) ( Figure 4 ) . These results suggest that there is not enough information to characterize autosomal dominant and recessive disease mutations in these prediction tools . To calculate an unbiased ( posterior ) probability of a rare nsSNV being pathogenic in a logit model ( See Materials and Methods ) , we need to estimate the prior probability , i . e . , the proportion of pathogenic rare nsSNVs in an individual genome . Therefore , we downloaded the high coverage exome sequencing data of 8 HapMap individuals [1] and estimated the proportion of pathogenic rare nsSNVs in each individual to get an estimate of the true prior with adjustment for sensitivity and specificity of the prediction ( See Materials and Methods ) . Also , we obtained the total load of pathogenic derived alleles ( See Materials and Methods ) . We used a reduced logit model that combines the scores from SIFT and MutationTaster ( with 80 . 2% averaged sensitivity and 82 . 8% averaged specificity ) since it has similar performance to the full model ( in terms of PR AUC ) and allows nsSNV with missing scores for any other three methods to be used . The proportion of predicted pathogenic rare nsSNVs ranged from 17% to 24% at individual genomes . After the adjustment of sensitivity and specificity , the true proportion of rare nsSNVs being pathogenic per individual genome is estimated to be around 0 . 6–12 . 2% , with a mean of 5% . Consequently , the total load of pathogenic derived alleles per individual genome varies from 3 to 51 , with a mean of 22 ( Table 2 ) . Based on the estimated average load of pathogenic derived alleles per individual ( which is around 22 at least ) , we calculated the expected load of homozygous pathogenic variants in an offspring from consanguineous mating . Given the theoretical inbreeding coefficient ( F ) of a child of a consanguineous union [16] , the corresponding number of homozygous pathogenic variants in the child equals F•N/2 , where N is the total load of pathogenic derived alleles in each common ancestor of the consanguineous couple ( Table 3 ) . For example , brother–sister marriages , which constitute 20–30% of all marriages in Roman Egypt [17] , would produce offspring homozygous for the pathogenic derived alleles at 3 loci . The offspring of first cousins with no family history of Mendelian diseases would theoretically have 1 . 4 homozygous pathogenic variants . However , in reality , there is a large variability in the values of F for the children of a given relationship . The offspring of first cousins can have a value of F as little as 3% or as much as 12% [18] , which corresponds to 0 . 3–1 . 3 homozygous pathogenic variants in an individual genome . Therefore , for counseling purposes , it would be important for genetic counselors to obtain the actual proportion of genome that the consanguineous couple shares by descent from common ancestors , say by genome-wide genotyping using commercial arrays , when assessing the risk to the offspring of a couple seeking information . We found that , on average , around 5% of rare nsSNVs an individual carries are pathogenic ( i . e . , the prior probability of a rare nsSNV being pathogenic ≈5% ) , but this does not mean that every rare nsSNV has the same ( posterior ) probability of being pathogenic given the scores from individual prediction methods . Figure 5 illustrates how both prior probability and prediction scores determine the posterior probability of a rare nsSNV being pathogenic . When the prior is large ( >0 . 9 ) , moderate prediction scores ( ∼0 . 6 ) can already lead to a posterior as large as 0 . 8 . But when the prior is small ( ∼0 ) , only a small posterior is obtained , regardless of the prediction scores . Within our estimated range of the prior , the posteriors range from 6 . 8×10−4 to 0 . 20 when the prediction scores are increased from 0 to 1 . So when all prediction scores of an nsSNV reach their maximal values of 1 , the probability that the variant is pathogenic is only around 20% , indicating that a logit model alone still cannot declare a variant as pathogenic even with the strong supports from several individual methods . So additional information ( like regions shared among multiple affected family members ) is needed to confidently isolate the causal variant ( s ) [3] . In contrast , when all prediction scores of an nsSNV are close to their minimal values of 0 , such probability is close to 0 , indicating that a logit model is confident to declare a variant as non-pathogenic even with no or little support from individual prediction methods . We applied the ExoVar-trained logit model using SIFT , PolyPhen2 and MutationTaster to prioritize nsSNVs in 3 Mendelian-disease patients with in-house exome sequencing data . Table 4 shows the numbers and percentages of nsSNVs removed by the hard-filtering approach and functional prediction using the logit model in these patients . Although hard-filtering ( MAF>1% in dbSNP , HapMap and 1000 Genomes ) can exclude ∼7 , 000 ( more than 80% of ) nsSNVs , there are still ∼1 , 000 variants left for examination in each patient . Prediction using different logit models can further exclude more than 500 ( 55% of ) nsSNVs . Of note , the causal mutations in these patients ( i . e . , c . 1528G>C of TGM6 for spinocerebellar ataxia patients [19] in patients 1 and 2 as well as compound heterozygosity for IL10RA mutations: c . 251C>T and c . 301C>T [20] in patient 3 with neonatal onset Crohn's disease ) were all predicted to be pathogenic by the logit model .
In this paper , we propose to use a logit model to combine multiple prediction methods to increase the performance in predicting pathogenic nsSNVs and to compute an unbiased ( posterior ) probability of an nsSNV being pathogenic in exome sequencing after hard-filtering . Also , we examine the predictive power of five popular prediction methods ( PolyPhen2 , SIFT , LRT , MutationTaster , PhyloP ) and two combined models ( CONDEL and logit ) in discriminating between pathogenic nsSNVs and other rare nsSNVs ( which is strictly relevant to exome sequencing studies ) . Contrary to our intuition , we found that the combined approach CONDEL is not necessarily better than individual methods , as demonstrated by its lower ROC AUC compared to that of MutationTaster . However , the logit model using multiple individual methods consistently outperforms other methods examined and the model combining SIFT and MutationTaster has comparable or even slightly better performance than that combining all of the five individual methods . Unfortunately , no method is able to discriminate between autosomal dominant and autosomal recessive disease mutations . Finally , based on the predictions of the logit model , we estimate that an individual has around 5% of rare nsSNVs being pathogenic and carries at least ∼22 pathogenic derived alleles , which if made homozygous by consanguineous marriages may lead to recessive diseases . We found that prediction methods are less powerful in predicting pathogenic variants from other rare variants than from common variants . This is consistent with the fact that most rare alleles ( no matter whether they are pathogenic or not ) of nsSNVs are subject to strong purifying selection and therefore have similar structural and functional properties , whereas common nsSNVs ( in which the minor alleles are also found at high frequency in populations ) are subject to weak purifying selection and so have nearly different properties compared to rare nsSNVs . Thus , as expected , it is more difficult to separate pathogenic mutations from other rare nsSNVs than from common nsSNVs using prediction methods . There is a significant difference in PhyloP conservation scores between dominant and recessive disease mutations . This may be because dominant disease genes are more conserved than recessive disease genes which can be “hidden” from purifying selection while heterozygous [21] . However , none of the prediction methods we examined , which mainly use the genomic features at variant level , are able to distinguish autosomal dominant mutations from autosomal recessive disease-causing mutations . Presumably , the genomic features at gene level may help distinguish them . Some studies have analyzed the difference in functional classification of the two classes of disease mutations and found that mutations in genes coding for enzymes and transporters are most likely to cause recessive diseases , whereas mutations in transcription regulators , structural molecules , nucleic acid binding genes and signal transducers have a higher chance to cause dominant diseases [22] , [23] . Also , genes involved in recessive diseases have less conserved paralogs than dominant disease genes [21] , [23] , as recessive diseases are often caused by loss-of-function mutations [22] , [24] ( which create a defective protein product with little or no biologic activity , and/or interfere with the normal expression of the gene ) . If a close paralog of a recessive disease gene is present , the paralog is likely to compensate for the loss of function due to a mutated recessive disease gene and so the disease is not observed [25] . On the contrary , dominant diseases are usually caused by gain-of-function mutations ( which confer a new activity on the gene product , or lead to its inappropriate spatial and temporal expression ) and so the presence of wild-type proteins encoded by functionally similar paralogs may not suppress the new functions acquired by the mutant proteins [23] . We found that the correlations among the scores from several complementary prediction methods are mostly weak to moderate ( see Figure S1 ) . This can occur for two possible reasons . First , the set of species used by one method for measuring conservation may be significantly different from those used by another , and thus this may lead to a big difference between the prediction scores calculated by the different methods for the same site . Second , the set of perfectly conserved sites used for training by one method may also be different from the ones used by others due to the variation in sequence alignments adopted by each method . Nevertheless , to our knowledge , MutationTaster uses the largest amount of resources for training and this may explain its excellent predictive performance among the five individual methods examined . But some redundancy among prediction scores from multiple methods also explains why combining a subset of the five individual methods ( i . e . , PolyPhen2 , SIFT , and MutationTaster ) has similar predictive performance to combine all five individual methods in a logit model . We note that our estimates of the number of pathogenic alleles per individual from the human genome data are higher than those from the data on consanguineous marriages ( which suggest a much smaller number , usually less than 10 [26]–[30] ) . But comparison of our estimates with those from inbreeding studies is difficult since we use totally different method and data . A similar situation has also occurred in quantifying the number of lethal equivalents per individual , in which inbreeding studies suggest that each individual carries 2-6 lethal equivalents [31] , [32] whereas Kondrashov [33] found the number could be as high as 100 . Anyway , estimation from inbreeding studies typically relies on an implicit assumption that all recessive alleles are completely penetrant and expressive , but examples that violate this assumption have recently been found . For example , the presence of a dominant modifier DFNM1 leads to normal hearing in an individual homozygous for the DFNB26 mutation [34]; high expression of actin-binding protein plastin 3 ( PLS3 ) protects individuals carrying homozygous SMN1 deletions from developing spinal muscular atrophy ( SMA ) [35]; and among the two siblings affected by autosomal recessive polycystic kidney disease ( ARPCKD ) , one died at 18 hr but the other still had no symptom when presented at 16 [36] . So the numbers from inbreeding studies are likely to be an underestimate . However , mapping errors may also inflate our estimates . For example , it was found that sequencing variants in the inactive gene copy of CDC27 gene ( i . e . , pseudogene ) were wrongly mapped to the active gene copy of CDC27 [37] and we observed that the active gene copy of CDC27 had as many as 11 nsSNVs at 2 out of 8 HapMap subjects examined . It is likely that some of these nsSNVs actually came from CDC27 pseudogene ( s ) and that can therefore inflate our estimates . But missing scores at sequencing variants could , on the other hand , deflate our estimates . Around 13–17% of rare nsSNVs in an individual have missing scores at SIFT and/or MutationTaster and so pathogenic alleles in these variants cannot be counted . We also observed a marked variability in our estimates of the number of pathogenic alleles per individual . The statistical fluctuation of specificity is the major reason for the large variability . As shown in Table 2 , a standard error of 1 . 3% in specificity can already lead to a standard error of 1 . 9% in the estimated proportion of pathogenic rare variants and finally results in a standard error of ∼10–15 in the estimated total load of pathogenic rare variants . We showed that , after MAF filtering , the prior ( i . e . , the proportion of variants left being pathogenic ) is low ( which is around 5% and leads to a posterior probability of ∼20% at most ) and so it is still difficult for prediction methods to pinpoint the pathogenic mutation ( s ) in exome sequencing studies of Mendelian diseases . One way to increase the prior is to use additional information , including genomic regions shared by multiple affected family members and known biological pathways , to reduce the number of candidate pathogenic variants and therefore we have implemented these functions in KGGSeq . We also found that even low prediction scores can lead to a posterior that can help exclude non-pathogenic variants . Using the exome sequencing data of three patients with Mendelian diseases , we observed that a logit model could exclude more than 55% of rare nsSNVs . Moreover , these posterior probabilities can be used as weights of the nsSNVs for other analyses .
We obtained , from dbNSFP database v2 . 0 [40] , the prediction scores from four prediction algorithms ( SIFT , HumVar-trained Polyphen2 , LRT and MutationTaster ) and a conservation score ( PhyloP ) for each nsSNV in the human genome ( hg19 ) . Some methods ( e . g . MutationTaster ) reported predictions for alternative ( or non-reference ) alleles while some ( e . g . PolyPhen2 ) conducted predictions for derived ( or non-ancestral ) alleles . To avoid inconsistency in predictions , we removed variants whose alternative alleles are not derived alleles . We also downloaded the CONDEL perl script from the authors' website ( see Data Access ) and used it to compute a CONDEL WAS score for each nsSNV based on the scores from all five individual methods . For all methods , the scores were standardized to range from 0 to 1 and , in general , the larger the score , the higher the probability of causing diseases . Given a vector , X , of prediction scores from multiple prediction methods for a particular nsSNV , the ( posterior ) probability that the nsSNV is pathogenic ( Y = 1 ) is given by: ( 1 ) where α and β are , respectively , the constant and vector of coefficients of X from logistic regression on a population ( or random sample ) of pathogenic ( positive control ) and other ( negative control ) variants . However , when the sample ( benchmark dataset ) is selected , the probability in Eq . 1 would be biased . An unbiased estimate of ( posterior ) probability can be obtained by ( see Text S1 ) : ( 2 ) where R is the odds of a variant being pathogenic in the selected sample divided by that in the population ( i . e . , an individual genome ) , i . e . , ( 3 ) where Pdisease denotes the proportion of pathogenic nsSNVs . Note that the value of R is the same for all variants and affects only the probability calculated , but not variant classification . Also , in our case , Pdisease in a selected sample is simply the proportion of positive control variants among all variants included in the benchmark dataset . We will demonstrate how Pdisease in an individual genome can be obtained in the section below . We evaluated the performance of five individual methods ( SIFT , HumVar-trained Polyphen2 , LRT , MutationTaster , and PhyloP ) and two combined methods ( CONDEL WAS and logit ) in discriminating between The parameters α and β in the logit model , as well as the complementary cumulative distributions and optimal cutoffs ( that maximizes MCC ) for each individual prediction method required by CONDEL were obtained from the training dataset in each fold in a K-fold cross-validation . When we validated the trained models in a test dataset , the probability cutoff was increased from 0 to 1 and the corresponding true positive prediction rate ( sensitivity ) and true negative prediction rate ( specificity ) of different prediction methods were recorded . We built the ROC and PR curves and reported the MCC [15] , ROC AUC and PR AUC of each method evaluated on each benchmark dataset . The program AUCCalculator [43] was used to calculate the AUCs . The sensitivities and specificities from K folds in a K-fold cross-validation were averaged for ROC and PR curve plotting . For the logit model , the quantity R in Eq . 3 is assumed to be 1 as this affects only the biasness of the probability calculated but not classification . Furthermore , we evaluated the effect of using a subset instead of all five individual methods in the logit model on the power in distinguishing pathogenic nsSNVs from other rare variants using a 10-fold cross-validation on the ExoVar dataset . To obtain an estimated range of Pdisease , i . e . , the proportion of pathogenic rare nsSNVs in an individual genome , in Eq . 3 , we need to get an estimated range of P*disease , i . e . , the proportion of predicted pathogenic rare nsSNVs in an individual genome . To estimate P*disease , we downloaded the high coverage exome sequencing data of 8 HapMap individuals and removed common nsSNVs annotated in dbSNP and 1000 Genomes using KGGSeq . Then , prediction was done in these filtered datasets using a logit model ( with a cutoff that maximizes the MCC evaluated on the ExoVar dataset ) to obtain P*disease for each of the individuals . Given the specificity and sensitivity of the prediction model used , we have ( 4 ) Since the number of homozygous genotypes for the derived allele in an individual is small compared to that of heterozygous genotypes , result from the prediction and estimation on all genotypes is similar to that on heterozygous genotypes only . So , for simplicity , we reported the latter results . That also simplifies our calculation of the total load of pathogenic derived alleles in each individual presented as the value just equals the number of pathogenic rare variants . To quantify the variability in the estimates of the total load of pathogenic derived alleles per individual , we randomly repeated the 10-fold cross validation 200 times . Each round of 10-fold cross validation generates a pair of averaged sensitivity and specificity values and this pair of values is then used to calculate Pdisease according to equation ( 4 ) and the total load of pathogenic derived alleles . We obtained the 95% confidence interval ( CI ) of the total load of pathogenic derived alleles by taking the 2 . 5% and 97 . 5% quantiles in the distribution of 200 such estimates . We implemented the logit model as one of important functions in our KGGseq software package , which was designed to conduct knowledge-based downstream analyses in sequencing studies . KGGSeq also gives a posterior probability of being pathogenic for each nsSNV by combining the five individual prediction scores available in the dbNSFP database v2 . 0 [40] . These features should facilitate the ranking and/or filtering of nsSNVs in exome sequencing studies of Mendelian diseases . We applied the logit model to the in-house exome sequencing data of three patients . KGGSeq was used to exclude variants and genotypes with low quality ( read coverage ≤4× , Phred-scaled base sequencing quality ≤50 , and Phred-scaled genotype calling quality score ≤20 ) , map variants onto the reference genome ( hg18 ) , extract nsSNVs as well as remove known common nsSNVs ( annotated in dbSNP and 1000 Genomes ) . Finally , for each of the remaining rare nsSNVs , KGGSeq assigned a logit prediction score , a posterior probability of being pathogenic and decided whether it is pathogenic or not using a cutoff that maximizes the MCC in ExoVar dataset . The URLs for data presented herein are as follows: KGGseq software tool , http://statgenpro . psychiatry . hku . hk/kggseq/ ExoVar dataset , http://statgenpro . psychiatry . hku . hk/limx/kggseq/download/ExoVar . xls dbNSFP , http://sites . google . com/site/jpopgen/dbNSFP Galaxy , http://main . g2 . bx . psu . edu/library Polyphen2HumVar dataset , http://www . nature . com/nmeth/journal/v7/n4/suppinfo/nmeth0410-248_S1 . html Online Mendelian Inheritance in Man , http://www . ncbi . nlm . nih . gov/Omim 1000 Genomes Project data , http://www . sph . umich . edu/csg/abecasis/MACH/download/ Polyphen-2 website , http://genetics . bwh . harvard . edu/pph2/ UniProt website , http://www . uniprot . org/ Picard website , http://picard . sourceforge . net/; High coverage exome sequencing data of 8 HapMap individuals , http://krishna . gs . washington . edu/12_exomes/; CONDEL , http://bg . upf . edu/condel/home | Sequencing the coding regions of the human genome is becoming a standard approach in identifying causal genes for human Mendelian diseases . Researchers often rely on multiple functional prediction methods/tools to separate the candidate causal mutation ( s ) from other rare mutations in these studies . In this paper , we propose the use of a statistical model to combine prediction scores from multiple methods and to estimate the chance of a rare mutation being Mendelian disease-causing ( or pathogenic ) . We found that our model using all or some individual prediction methods consistently outperforms other prediction methods examined and could exclude more than 55% of rare non-pathogenic mutations in an individual genome . Unfortunately , no method was able to discriminate between autosomal dominant and autosomal recessive disease mutations . In addition , based on the predictions of our model , we estimated that a person can carry ∼22 pathogenic derived alleles at least , which if present at the same position in the genome may lead to Mendelian diseases . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"genome",
"sequencing",
"genomics",
"genetic",
"mutation",
"genetics",
"biology",
"human",
"genetics",
"genetics",
"of",
"disease",
"genetics",
"and",
"genomics"
]
| 2013 | Predicting Mendelian Disease-Causing Non-Synonymous Single Nucleotide Variants in Exome Sequencing Studies |
Isogenic cells sensing identical external signals can take markedly different decisions . Such decisions often correlate with pre-existing cell-to-cell differences in protein levels . When not neglected in signal transduction models , these differences are accounted for in a static manner , by assuming randomly distributed initial protein levels . However , this approach ignores the a priori non-trivial interplay between signal transduction and the source of this cell-to-cell variability: temporal fluctuations of protein levels in individual cells , driven by noisy synthesis and degradation . Thus , modeling protein fluctuations , rather than their consequences on the initial population heterogeneity , would set the quantitative analysis of signal transduction on firmer grounds . Adopting this dynamical view on cell-to-cell differences amounts to recast extrinsic variability into intrinsic noise . Here , we propose a generic approach to merge , in a systematic and principled manner , signal transduction models with stochastic protein turnover models . When applied to an established kinetic model of TRAIL-induced apoptosis , our approach markedly increased model prediction capabilities . One obtains a mechanistic explanation of yet-unexplained observations on fractional killing and non-trivial robust predictions of the temporal evolution of cell resistance to TRAIL in HeLa cells . Our results provide an alternative explanation to survival via induction of survival pathways since no TRAIL-induced regulations are needed and suggest that short-lived anti-apoptotic protein Mcl1 exhibit large and rare fluctuations . More generally , our results highlight the importance of accounting for stochastic protein turnover to quantitatively understand signal transduction over extended durations , and imply that fluctuations of short-lived proteins deserve particular attention .
TNF-Related Apoptosis Inducing-Ligand ( TRAIL ) is a promising therapeutic agent against cancer because it induces apoptosis specifically in tumor cells [1]–[3] . This motivated dozens of clinical trials based on TRAIL-related therapies . However , efficiency was usually limited [4] . Most of the molecular events leading from TRAIL exposure to cell death are known [3] . After TRAIL binding to death receptors , initiator caspases are activated , which in turn promote effector caspases activation either directly or via a mitochondrial pathway ( Fig . S1 ) . In most cells , Mitochondrial Outer Membrane Permeabilization ( MOMP ) is required to efficiently activate effector caspases . Several kinetic models have been proposed to describe a part or all of those biochemical reactions [5]–[12] . Not all cells of an isogenic population die after TRAIL treatment , even at saturating ligand doses . This fractional killing property is widely shared among cell lines and is critical for therapeutical applications [13] . In addition , surviving cells were shown to be transiently resistant to a second TRAIL treatment . This reversible resistance property was observed in various cell lines and could also have important implications for therapy [14] , [15] . Fractional killing is generally thought to result from cross talks between the apoptosis pathway and survival pathways [16] . Indeed , several studies reported that TRAIL increases the production of anti-apoptotic proteins , via the activation of survival pathways [17]–[19] . While fractional killing illustrates cell-to-cell variability in the decision between life and death , variability is also observed among cells that die: they commit to death after a highly variable delay from one another [13] , [20] . This variability cannot be explained by differences in TRAIL-induced gene regulation: it is also observed when cells are co-treated with cycloheximide ( CHX ) , an efficient inhibitor of protein synthesis [13] . Rather , it was proposed to originate mostly from pre-existing differences in the levels of proteins composing the apoptosis pathway . Indeed , recently divided sister cells die almost synchronously [13] , [20] , as expected if protein content is equally shared between daughters and if noise in signaling reactions play a marginal role . This explanation was supported by modeling: when taken as initial conditions of a deterministic , kinetic model describing signaling reactions , differences in protein levels are sufficient to explain observed variability in death times [13] . Another observation indicated that this cell-to-cell variability in protein levels is not frozen but result from a dynamical equilibrium driven by fluctuations in individual cells: death synchrony between sister cells was weaker as the duration between division and treatment increased [13] , [20] . To quantitatively assess this effect , the previously mentioned modeling approach , where only the consequences of protein fluctuations ( cell-to-cell variability at a fixed time ) are accounted , is inadequate . Instead , protein fluctuations themselves should be modeled . The need to account for protein fluctuations is even more stringent when considering observations after treatment with TRAIL alone . In that case , protein synthesis is not blocked and thus can impact the decision between life and death . Even if TRAIL does not change protein synthesis ( via induction of survival pathways ) , significant differences with TRAIL and CHX treatments are expected , as the constitutively noisy protein synthesis will interact with signaling reactions . Thus , before complexifying the model to account for eventual regulations via survival pathways induction , it is critical to assess how much can be explained when protein synthesis is not altered by TRAIL signaling . Here , we investigate this question by enabling a kinetic model of TRAIL-induced apoptosis with stochastic protein turnover for all proteins , following a generic and principled approach . To our knowledge , this is the first attempt to systematically include gene expression noise in a signal transduction model . It enriches the model with a fundamental property as the dynamics of cell-to-cell variability is represented , allowing disentangling the effects of constitutive protein fluctuations , signaling protein-protein reactions and potentially induced changes in protein synthesis .
Protein synthesis and degradation are subjected to noise , resulting in fluctuations of protein concentrations in individual cells and in cell-to-cell variability at the population level [21] , [22] . Such variability could have consequences on signal transduction: aside of conventional epigenetic differences [23] , unequal access to ligand molecules or simply noise in signaling reactions , it often contributes importantly to heterogeneous behavior within an isogenic population [13] , [24] . One approach to account for those differences is to incorporate protein level variability as random initial conditions of an ODE model describing the signaling reactions ( “extrinsic noise approach” ) [13] , [25] . However , variability is imposed at time zero and then behavior is deterministic: it is thus not appropriate to study transduction on long time scales , during which protein levels dynamically fluctuate [26] . A more natural manner to account for protein level variability is to represent their stochastic synthesis and degradation ( “intrinsic noise approach” ) . Although several studies did account for cell-to-cell differences in protein levels in an extrinsic , static manner via random initial conditions [13] , [27] , [28] , and many models of signal transduction considered the effect of noise in protein-protein reactions [29] , [30] or in the expression of signal transduction target genes [31]–[33] , no kinetic model of signal transduction pathways considering systematically noise in protein synthesis and degradation has been developed so far . Here by systematically we mean for all the proteins acting in the pathway . We propose a modeling approach to account for gene expression noise within kinetic models of signal transduction pathways . Following Singh et al . [34] , we model protein turnover with stochastic processes describing mRNA level fluctuations , and deterministic processes for protein translation and degradation ( random telegraph model , see also [35]–[41] ) . These processes are integrated into a kinetic model of protein-protein reactions ( Fig . 1 ) . While the rates of such stochastic protein turnover models ( Fig . 2A ) are rarely directly measurable , their value can be constrained by using experimentally measurable data and analytical results ( Fig . S2 ) . Recently , significant progress has been made on both experimental and theoretical sides to enable this inference approach [34]–[41] . Importantly , we found that for typical protein and mRNA half-lives , a large set of promoter rate combinations lead to similar fluctuations at the protein level , as characterized by protein level variability ( coefficient of variation ) and mixing time ( half-autocorrelation time ) . Interestingly , the obtained mixing time is around 40 hours ( Fig . 2B ) , in the middle of the range of experimentally estimated values for twenty endogenous proteins in human cells [26] . Thus , standard stochastic protein turnover models ( Fig . 2C ) can provide a good approximation of protein fluctuations for most proteins . Only short-lived proteins necessitate particular attention ( Fig . S3 ) . This finding is a cornerstone of our approach . We applied this approach to TRAIL-induced apoptosis , using EARM kinetic model [5] , [13] to describe protein-protein reactions taking place between TRAIL exposure and cell death commitment . We equipped all native proteins with a default model of stochastic protein turnover , with the exception of proteins with fast turnover , here Flip and Mcl1 [42]–[44] . Details of model construction are given in Text S1 ( see dedicated section , Fig . S4 and Tables S1-3 ) . Importantly , all parameters have been constrained based on experimental data and analytical results , with the exception of four parameters ( “ON” and “OFF” promoter switching rates , mRNA and protein half-lives for Flip and Mcl1 ) . Note that because of their similar protein and mRNA half-life , we use the same couple of promoter switching rates for both proteins in our exploration of parameter space . This drastically limited the number of introduced degrees of freedom and made it possible to systematically explore realistic ranges for remaining parameters . To study the influence of stochastic protein turnover on fractional killing and reversible resistance , we sought to confront our model with existing quantitative data about TRAIL-induced apoptosis in HeLa cells . Those experiments , described in detail later , can be classified into two groups based on the type of information they contain: 1 ) quantification of the variability in cell fate , 2 ) characterization of the transient memory in cell state . While previous approaches using ODE models with distributions for initial protein levels ( capturing a static description of cell-to-cell variability ) [13] , [25] are potentially able to reproduce the first type of data , a dynamic view on cell-to-cell variability as proposed in our model is needed to account for both types of data . We adopted the following strategy: first , search for models able to reproduce observations on cell fate variability; and second test whether valid models can robustly predict observed behaviors where transient memory matters . Using live-cell microscopy , Spencer et al . [13] investigated the fate of hundreds of cells after exposure to TRAIL and CHX ( 10 ng/mL and 2 . 5 µg/mL , Fig . 3A ) . All cells undergo MOMP with a highly variable delay ( from 2 to 8 hours , Fig . 3B ) . To study cell fate inheritance , the authors also recorded 20 hours before treatment to identify sister cells ( Fig . 4A ) . They were found to have highly correlated MOMP times ( correlation coefficient close to 1 for recently divided cells , about 0 . 5 for older sisters - Fig . 4B , black curve ) . Here , the MOMP time distribution provides a quantification of the cell fate variability , while MOMP time correlations between sister cells also give information on the transient memory in cell state . Within our framework , in-silico reproduction of those experiments is straightforward ( Figs . 3D and 4C ) , enabling us to investigate possible origins of transient cell fate inheritance . We first asked if the observed cell fate variability could be reproduced . In the model , it is only determined by protein levels at treatment time ( behavior is deterministic as synthesis is assumed to be fully blocked by CHX and noise in signaling reactions is neglected ) , and differences between sister cells are only caused by protein synthesis noise occurring between division and treatment ( in agreement with the fact that recently divided sisters died almost synchronously , we assumed an equal repartition of protein content at division ) . We found that excellent agreement with observed MOMP time variability can be obtained ( Fig . 3E ) . Further analysis revealed that such agreement requires Flip and Mcl1 protein half-life to be short and to fall within a narrow range ( between 0 . 3 and 0 . 6 hours , Fig . S5-C ) . This model prediction is consistent with previous measurements in HeLa cells ( 30 and 40 minutes for Flip short isoform and Mcl1 respectively , [42] , [43] ) . In contrast , Flip and Mcl1 mRNA half-life and promoter switching rates are not strongly constrained , probably because their influence on cell fate is limited by the rapid protein level decrease caused by synthesis blockade . We then asked whether our extended model also capture transient cell fate inheritance ( Fig . 4B ) . It is the case: fitted models accurately predict the MOMP time correlation between sister cells ( Fig . 4D - black curve , Fig . S5-C ) . Of note , assuming standard promoter switching rates for Flip and Mcl1 ( but accounting for their short mRNA and protein half-life - this parameterization will later be referred as the “non-fitted” model ) already provides a good agreement for both MOMP time distribution and MOMP time correlation between sister cells ( Fig . S6 ) . This non-trivial result shows that the speed at which the sensitivity to TRAIL and CHX fluctuates in single cells is well captured and thus suggests that our generic approach permits to describe fluctuations of protein levels with sufficient accuracy . Spencer and colleagues repeated this experiment but treated cells with TRAIL alone ( 250 ng/mL ) . In this condition , an important fraction of cells died fast ( MOMP in ∼2 hours ) but 40% were still alive after 8 hours ( Fig . 3C ) , illustrating the fractional killing property . Also , cell fate inheritance between sister cells was markedly changed: only young sister cells that underwent MOMP rapidly were importantly correlated ( Fig . 4B , grey curve ) . We asked whether the observed cell fate variability , including fractional killing , could be reproduced in-silico . Within our modeling assumptions , absence of co-treatment with CHX makes a fundamental difference: as synthesis continues , the effect of gene expression noise during TRAIL-induced apoptosis could be investigated , and comparison with the TRAIL and CHX condition is insightful . Strikingly , we found that quantitative agreement for both MOMP time distribution and surviving fraction could be obtained ( Fig . 3F ) . Robustness analysis showed that rates of the Flip and Mcl1 stochastic protein turnover model , and particularly promoter switching rates , are in this case strongly constrained . Interestingly , MOMP time distribution and surviving fraction constrain those values differently ( Fig . 5A–B ) , resulting in an narrow ranges for their values: agreement for both observations together is obtained only when promoter switching rates are both low ( Fig . 5C ) . Such low switching rates lead to large , rare fluctuations of protein levels ( Fig . 5D ) . Those atypical fluctuations phenotypes are expected to leave a signature at the population level: the shape of the protein level distribution would be bimodal rather than resembling a lognormal distribution ( Fig . S7 ) . This property is thus a model prediction . Those fluctuations are likely to impact how the fate of sister cells diverge with time . Thus , we asked whether the model could also account for the observed fast loss of cell fate inheritance . Remarkably , the fitted models accurately and robustly predict MOMP time correlations between sister cells ( Figs . 4D and 5E ) . As mentioned earlier , the same couple of promoter switching rates was used for Flip and Mcl1 during exploration , but further analysis showed that assuming low promoter switching rates for Mcl1 alone was sufficient to obtain quantitative agreement for MOMP distributions surviving fractions , and that sister cells MOMP time correlations were still correctly predicted ( Fig . S8 ) . Thus , comparison with transient cell fate inheritance data supports that large , rare fluctuations of Mcl1 could be responsible for the observed cell fate variability . Recently , reversible resistance was observed among various cell lines [14] . Cell populations were submitted to two consecutive TRAIL treatments . The duration between treatments was varied from 1 day to 1 week ( Fig . 6A ) . One-day survivors were significantly more resistant than the initial population , but such resistance was significantly decreased or even lost in one-week survivors . Thus , cells surviving a first TRAIL treatment are transiently resistant . Remarkably , in-silico reproduction of those ( Fig . 6B ) showed that our model predicts the presence of reversible resistance ( Fig . 6C ) : one-day survivors exhibit a dose-dependent increase of resistance to a second TRAIL treatment , which disappears after 3 to 5 days . This is surprising since our model does not include induced regulation mediated by survival pathways . Moreover , the presence of reversible resistance is a robust property of the model as it is also obtained when assuming standard promoter switching rates for Mcl1 and Flip ( “non-fitted” model , Fig . S6 ) . However , agreement with related experimental data [14] , [15] is only qualitative . While we cannot exclude that model parameterizations allowing a quantitative agreement exist , it might be needed to include additional mechanisms such as survival pathways induction to explain the observed sustained resistance gain after one week when treating cells with a high TRAIL dose [14] . What are the mechanisms behind cell escape to TRAIL-induced apoptosis , either on the short-term ( fractional killing ) or the long-term ( reversible resistance ) ? Using the fact that in-silico , all protein , mRNA levels and gene activity states can be monitored in single cells , we investigated those questions at the molecular level . To study the influence of pre-existing differences on cell fate , we compared at the time of stimulation the sub-population of ‘future survivors’ with the whole population ( Fig . 7A–B ) . Future survivors strongly stood out by their Mcl1 protein level and gene activity state ( Fig . 7B ) . Flip also appeared to play an important role in determining cell decision , and smaller but significant effect was also seen for Bid , Bax , Bcl2 and XIAP . Although it is a good predictor of cell fate , initial Mcl1 gene activity status does not completely determine survival: neither all Mcl1 “ON” cells survived nor all Mcl1 “OFF” cells died . Thus , pre-existing differences in protein levels and promoter activities are major determinant of cell fate but stochastic events in gene expression occurring during signal transduction also play a role . While timing of death for cells treated with TRAIL and CHX appeared to be multi-factorial [13] , our results suggest that cell survival is predominantly determined by Mcl1 ( Fig . 7B ) . This important role of Mcl1 is robustly predicted . Indeed , it also holds for the “non-fitted” model , which assume standard promoter switching rates for all proteins , including Mcl1 and Flip ( Fig . S6 ) . To investigate the determinants of reversible resistance , we tracked the temporal evolution of protein levels in surviving cells ( Fig . 7 A&C ) . The protein level composition of one day survivors contrasts with the protein content observed in future survivors: almost all protein levels differ importantly from the naïve population composition , while that was the case only for Mcl1 and Flip in future survivors . This is expected as all proteins are partly activated during signal transduction , leading to a higher degradation ( active forms have a shorter half-life ) . Therefore , the distinction between the causes of cell survival and the consequences of partial apoptosis induction cannot be easily resolved by the sole observation of protein levels in survivors . When signaling stops , recovery of protein levels is expected to follow exponential kinetics governed by the turnover rate ( Fig . 7C , inset – see the death receptor ( R ) , pro-caspase 8 , Bar and Bid ) . Deviation from such kinetics indicates either the persistence of signaling reactions that continue to consume proteins ( it is the case for Apaf , pro-caspase 9 and XIAP , as further analysis confirmed ) or is a consequence of important selection . Indeed , while Mcl1 and Flip should recover normal levels in a few hours in absence of selection because of their high turnover rate , Mcl1 levels ( but not Flip levels ) are still strongly higher than in naive cells one day after TRAIL treatment , consistently with previous observations on the relative selection strength that operated on them . Together , those results indicate that recovery phenotypes in surviving cells result from a complex interplay of three distinct effects: selection during apoptosis , transcriptional noise and protein turnover as a driving force tending to reset protein levels to their initial , pre-stimulus distribution , and long-term residual signaling activity . This explains why it is difficult to understand the recovery process and justifies the use of modeling to disentangle the various contributions .
In this study , we merged those two approaches by integrating such stochastic models of gene expression within an existing kinetic model of TRAIL-induced apoptosis [5] in a systematic and principled manner . Doing so provides advantages compared to previous approaches to account for cell-to-cell variability in protein levels [13] , [25] . First , variability is not considered as an “input” parameter but arises naturally from stochastic fluctuations . The dynamics of this variability is thus intrinsically represented within the system , allowing investigating the effects of transient memory in protein levels . Second , the influence of protein synthesis noise during TRAIL-induced apoptosis could also be investigated . Importantly , we followed a parsimonious parameterization strategy , motivated by the fact that fluctuations of long-lived proteins are rather insensitive to the precise kinetics of transcriptional bursting , enabling us to equip most proteins ( long-lived proteins ) with reasonably accurate fluctuation models even in absence of gene expression data for each and every promoter . The sister cells experiment for which cells were treated with TRAIL and CHX provided ideal data to validate our modeling approach: in that case , behavior is mostly deterministic as soon as treatment starts and only fluctuations occurring before treatment are responsible for death time variability and de-correlation between sister cells . Moreover , gene regulation via survival pathways induction is ineffective as protein synthesis is blocked . Because our model was able to quantitatively reproduce the MOMP time distribution and then accurately predicted sister cells correlation , our modeling approach appears as a promising tool to investigate the effect of protein fluctuations on signal transduction , despite the limitations inherent to its simplicity ( for example , stochastic gene expression events were assumed to be independent between proteins , neglecting the fact that levels of different proteins can be partially correlated [25] , [26] , possibly because of common transcription factors or coordinated chromatin-state transitions ) . Of note , good agreement was readily obtained when assuming for Flip and Mcl1 standard promoter switching rates , but short protein and mRNA half-lives - in agreement with available knowledge ( “non-fitted” model , Fig . S6 ) . Finally , the transposition of fluctuation timescales from individual proteins into ‘TRAIL sensitivity states’ is not trivial: while in our model , stable proteins levels are mixed in about 40 hours , cells were switching between 'fast dying' and 'slow dying' phenotypes more rapidly ( about 10–15 hours ) . As combinatorial effects are at play , mechanistic models of protein-proteins reactions are needed to link protein-level timescales with more high-level phenotypic transitions [45] . Several studies reported that TRAIL can induce survival pathways [17]–[19] . How such induced changes affect signal transduction and eventually stop apoptotic signaling remains unclear . On the other hand , the contribution of constitutive protein synthesis noise , which is responsible for pre-existing differences between cells , has not been evaluated . Although it does not exclude the existence of other mechanisms , an important result of our study is that fractional killing can be obtained without assuming any TRAIL-induced regulation . Alternatively , we find that because of its fast turnover , constitutive expression of the Mcl1 protein has the potential to rescue cells from TRAIL apoptotic signaling . In this context , solely accounting for protein fluctuations within the TRAIL apoptosis pathway predicts the fractional killing property ( Figs . 3 and S6 ) . While our results challenge current opinion on the role of survival pathways in TRAIL-induced apoptosis , they are consistent with observations made on wild type HeLa cells that neither blocking NF-κB response nor inhibiting the Akt pathway do significantly change the surviving cell fraction after TRAIL treatment [46] , [47] . The pivotal role for Mcl1 in TRAIL-induced apoptosis predicted by our model is consistent with the recent finding that Mcl1 silencing by shRNA in HeLa cells completely sensitize cells to TRAIL [48] . While moderate fluctuations of Mcl1 levels were sufficient to obtain fractional killing , a quantitative agreement with the Spencer et al . [13] single-cell data ( MOMP time distribution and surviving fraction ) required large and rare Mcl1 fluctuations , caused by rare switches between long periods of gene activity or inactivity . Interestingly , in that case , the observed rapid loss of MOMP time correlation between sister cells quantitatively emerged from model simulations . Flip is often mentioned as a key factor in cell resistance to TRAIL [49] , but in our model Flip has less impact on cell survival than Mcl1 . Consistently , Lemke et al . [48] silencing experiments demonstrated a dominant role for Mcl1 and a synergy with Flip . However , our model might under-estimate the role of Flip: the representation of DISC-related events in EARM is simple and thus does not account for recent biological findings , including the stoichiometry between its components [50] , [51] . Improving how DISC assembly is modeled might thus be needed to elucidate the precise role of Flip in fractional killing and reversible resistance , especially for cell lines that express higher Flip amounts than HeLa . A second significant result reported here is that our model predicts the phenomenon of reversible resistance , showing that constitutively noisy protein synthesis , protein-protein interactions and protein degradation are by themselves sufficient to explain a dose-dependent , significant increase of resistance in recent survivors and its gradual loss within 3–5 days . This result is consistent with the observation that Nf-κB blockade does not change resistance acquisition after TRAIL treatment [14] ( in MCF10A cells; HeLa cells have not been tested ) . In-silico analysis at the molecular level revealed that reversible resistance as predicted by the model was shaped by a complex interplay between 1 ) selection based on protein levels and transcriptional activity , 2 ) protein turnover and 3 ) residual signaling activity . As opposed to the death process , which involves a sharp and complete activation of effector caspases , our results suggest that recovery in cells that did not commit to death is a slow and complex process . While one should not conclude from our results that parallel activation of survival pathways by TRAIL plays no role in reversible resistance , our results show that the sole contribution of protein level fluctuations occurring within the extrinsic apoptosis pathway can partly lead to reversible resistance . Thus , protein fluctuations should be accounted for to gain quantitative insights into reversible resistance . While here we focused on TRAIL-induced apoptosis , our modeling approach is generic and can be applied to other signal transduction pathways . Our results showed that even in absence of induced gene regulation , gene expression noise interacts with signaling dynamics on a non-trivial manner . Thus , even in contexts where the influence of induced gene-regulation is indisputable , its sound quantification probably requires to investigate first the role of constitutive gene expression noise . Only then models could be enriched parsimoniously with well-characterized regulatory links until all observations are successfully explained . Significant advances to allow such detailed characterization of gene regulation occurred recently [31] , [52] , [53] . Following such approaches could significantly extend the reach of models of signal transduction towards accurate , single-cell level description of populations submitted to varying signaling contexts over multiple cell generations .
Denoting and the two states of the promoter , the number of mRNAs and the protein level , the stochastic protein turnover model comprises the following reactions: ( Gene activity switches ) . ( Transcription and mRNA degradation ) . ( Translation and protein degradation ) . We interpret the four first reactions as stochastic reactions and the translation and protein degradation reactions as deterministic reactions . Thus , the model state is given by a Boolean , an integer and a continuous variable . The statistical properties of this system can be investigated analytically by writing the corresponding chemical master equation , which describes the temporal evolution of the state joint probability distribution . While the obtained equations cannot be solved analytically , it can be used to derive moments of the underlying distribution by applying generating functions techniques , as was done by Paszek [39] . We next give the analytical expressions of several moments of interest: While those expressions characterize the steady-state distribution of the model , and thus can be compared to snapshot measurements of a cell population , they do not provide information about how fast single cells move with time within those distributions . Such information can be obtained by computing the autocorrelation function . Its expression and derivation are given in Supplementary Results ( Text S1 ) . We used the EARM model described in Spencer et al . [13] to represent protein-protein reactions . No kinetic rates were changed , except the rate of pC6 cleavage by C3 , set to zero to remove the feedback loop . Degradation rates of non-native forms were modified according to biological knowledge ( Table S3 ) . The effect of those changes is discussed in SI Appendix . Rates of the standard stochastic protein turnover models and rates of Flip&Mcl1 models were determined following an algorithm constructed to incorporate constraints based on biological knowledge ( Fig . S4 ) . Their values are given in Table S1-2 . Notably , for comparability and consistency with the findings of Spencer et al . [13] , we kept their mean protein values , even if for a few proteins new data were available , as for Mcl1 for example [54] , provided that the order of magnitude is the same . The simulation procedure is detailed in SI Appendix . Briefly , mother cell states ( promoter activity , mRNA and protein levels ) were chosen by Monte-Carlo sampling and the two sister cells were constructed by duplication of the mother cell state . Promoter activity and mRNA fluctuations were simulated using an implementation of the Gillespie algorithm . For the ODEs governing evolution of all protein levels , the Semi-Implicit Extrapolation method was used . MOMP was considered to have occurred when half of mitochondrial Smac has been released . For all simulations , at least 104 pairs of sister cells were simulated . Quantitative comparison of simulation results with related experimental data ( Fig . 5 ) was achieved by computing a measure of model-data agreement described in Supplementary Methods ( Text S1 ) . For repeated TRAIL simulations , a naïve population of 104 cells was obtained as in the sister cells experiment . Cells divided after a gaussian cell cycle duration ( mean 27 hours , standard deviation 3 hours ) . Death was detected via cPARP levels as in Gaudet et al . [25] . | TRAIL induces apoptosis selectively in cancer cells and is currently tested in clinics . Having a mechanistic understanding of TRAIL resistance could help to limit its apparition . Several observations suggested that protein level fluctuations play an important role in TRAIL resistance and its acquisition . However , quantitative , systems-level approaches to investigate their role in cellular decision-making processes are lacking . We propose a generic and principled approach to extend signal transduction models with protein fluctuation models for all proteins in the pathway . The key aspect is to use standard protein fluctuation models for long-lived proteins . We show that its application to TRAIL-induced apoptosis provide a quantitative , mechanistic explanation to previously published but yet unexplained critical observations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"biochemistry",
"biochemical",
"simulations",
"cell",
"death",
"signal",
"transduction",
"mathematics",
"cell",
"biology",
"apoptosis",
"gene",
"expression",
"genetics",
"biology",
"and",
"life",
"sciences",
"cell",
"processes",
"physical",
"sciences",
"computational",
"biology",
"stochastic",
"processes",
"probability",
"theory"
]
| 2014 | Modeling Dynamics of Cell-to-Cell Variability in TRAIL-Induced Apoptosis Explains Fractional Killing and Predicts Reversible Resistance |
Loci identified in genome-wide association studies ( GWAS ) of cardio-metabolic traits account for a small proportion of the traits' heritability . To date , most association studies have not considered parent-of-origin effects ( POEs ) . Here we report investigation of POEs on adiposity and glycemic traits in young adults . The Jerusalem Perinatal Family Follow-Up Study ( JPS ) , comprising 1250 young adults and their mothers was used for discovery . Focusing on 18 genes identified by previous GWAS as associated with cardio-metabolic traits , we used linear regression to examine the associations of maternally- and paternally-derived offspring minor alleles with body mass index ( BMI ) , waist circumference ( WC ) , fasting glucose and insulin . We replicated and meta-analyzed JPS findings in individuals of European ancestry aged ≤50 belonging to pedigrees from the Framingham Heart Study , Family Heart Study and Erasmus Rucphen Family study ( total N≅4800 ) . We considered p<2 . 7x10-4 statistically significant to account for multiple testing . We identified a common coding variant in the 4th exon of APOB ( rs1367117 ) with a significant maternally-derived effect on BMI ( β = 0 . 8; 95%CI:0 . 4 , 1 . 1; p = 3 . 1x10-5 ) and WC ( β = 2 . 7; 95%CI:1 . 7 , 3 . 7; p = 2 . 1x10-7 ) . The corresponding paternally-derived effects were non-significant ( p>0 . 6 ) . Suggestive maternally-derived associations of rs1367117 were observed with fasting glucose ( β = 0 . 9; 95%CI:0 . 3 , 1 . 5; p = 4 . 0x10-3 ) and insulin ( ln-transformed , β = 0 . 06; 95%CI:0 . 03 , 0 . 1; p = 7 . 4x10-4 ) . Bioinformatic annotation for rs1367117 revealed a variety of regulatory functions in this region in liver and adipose tissues and a 50% methylation pattern in liver only , consistent with allelic-specific methylation , which may indicate tissue-specific POE . Our findings demonstrate a maternal-specific association between a common APOB variant and adiposity , an association that was not previously detected in GWAS . These results provide evidence for the role of regulatory mechanisms , POEs specifically , in adiposity . In addition this study highlights the benefit of utilizing family studies for deciphering the genetic architecture of complex traits .
Genome-wide association studies ( GWAS ) have identified multiple loci associated with cardio-metabolic risk ( CMR ) phenotypes and traits , such as adiposity and glycemic traits ( e . g [1–4] ) . GWAS are conducted , very largely , in unrelated individuals . In general , the magnitudes of these genetic associations are small to modest , and the loci account for only a small proportion of the heritability of these traits [5–8] . To explain the “missing” heritability , various research strategies for follow-up studies have been proposed , including focusing on gene-environment interactions , on rare variants with moderate effects and on parent-of-origin effects ( POEs ) [7 , 9] . POEs are non-Mendelian transmittable genetic effects on phenotypes , where the phenotype in offspring depends on whether transmission originated from the mother or father . In mammals , POEs can be caused by genomic imprinting , maternal genetic effects on the intrauterine environment , or maternally inherited mitochondrial genes [10] . The most obvious mechanism underlying POEs is genomic imprinting; imprinted genes show parental-specific monoallelic or partial expression , dictated by the parental origin of the chromosome [11 , 12] . In the past decade , linkage studies have demonstrated POEs on both binary and continuous CMR outcomes [13–17] . Yet , despite evidence for POEs from linkage studies , most association studies have overlooked their potential influence on CMR traits , and the extent and magnitude of POEs on CMR traits remain largely unknown . One important exception , from Iceland , demonstrated that inclusion of the parental source of alleles significantly strengthened previously-reported associations with type 2 diabetes [18] . Two subsequent studies in Icelanders detected POEs on age at menarche and thyroid-stimulating hormone levels , traits that are associated with CMR [19 , 20] . To examine POEs on adiposity and glycemic traits in young adults , we leveraged existing genotype data and CMR trait measurements on mother-offspring pairs from a longitudinal study nested within a historical birth cohort , together with GWAS and bioinformatics databases . We replicated our findings using extended pedigrees from three family studies . Our hypothesis was that , in genes where variants are known to be associated with CMR traits , knowledge of the parental source of genetic variants would reveal POEs on body mass index ( BMI ) , waist circumference ( WC ) , and fasting glucose and insulin levels .
In the discovery stage conducted in JPS , SNP rs1367117 , a coding SNP located within the APOB gene , demonstrated significant POEs on adiposity traits ( Table 1 and S5 Table ) . Specifically , significant associations were demonstrated when the minor allele was inherited from the mother for offspring BMI ( beta = 1 . 44 , p-value = 6 . 5x10-3 ) and WC ( beta = 3 . 82 , p-value = 2 . 7x10-3 ) . The corresponding paternally-derived effects were non-significant ( p>0 . 6 ) . As expected in the setting of POEs , in JPS associations seen with the maternally-derived minor allele were approximately twice as strong as those that did not consider the parental source of the allele ( effects sizes were 0 . 7 ( p-value = . 03 ) and 1 . 5 ( p-value = . 05 ) for BMI and WC respectively ) ( Fig 1 and S5 Table ) . In fact , the differences in simple heritability estimates in JPS derived from mother-offspring correlations [21 , 22] based on models with either adjustment for POE or additive genetic effects of the APOB variant suggested that POE of the APOB SNP explains roughly 0 . 30% and 0 . 58% of BMI and WC heritability , respectively . No changes in heritability estimates for BMI or WC were observed for the additive genetic effect of this variant . We also conducted sensitivity analyses in which we excluded heterozygous mother-offspring pairs ( 191 of 1235 pairs ) instead of using probability-based estimated dosage in these pairs . These analyses resulted in similar or larger estimates for APOB rs1367117 POE on BMI ( maternal beta = 1 . 46; p = . 048 ) , WC ( maternal beta = 5 . 26; p = . 003 ) and fasting insulin ( natural log transformed , maternal beta = . 212; p = . 004 ) compared to those using estimated dosage ( Table 1 ) suggesting that our probability-based estimated dosage approach is likely conservative . Additionally , we used simulations to assess the likelihood of false positive findings . Permutation-based p-values for the maternal effect on BMI ( p-value = . 0079 ) and on WC ( p-value = . 0048 ) were very close to the observed maternal p-values ( p-value = . 0065 and . 0027 for BMI and WC , respectively ) . Detailed information on simulations in JPS ( and in FHS ) can be found in S1 Text ( permutation testing ) . Testing for POEs for APOB SNP rs1367117 in the replication and meta-analysis stage , also showed a significant maternally-derived effect in FHS and FamHS separately , but not in ERF , and in all four studies combined on both BMI ( combined beta = 0 . 75 , p-value = 3 . 1x10-5 ) and WC ( combined beta = 2 . 66 , p-value = 2 . 1x10-7 ) ( Table 1 ) . The effect sizes for the maternal-specific associations in JPS and FamHS were essentially identical , whereas the corresponding effects in FHS were somewhat smaller . The associations of the paternally-derived allele with those traits were not significant ( F-test p-values for difference between maternally- and paternally-derived effects were 5 . 0x10-4 and 1 . 0x10-3 for BMI and WC respectively ) . Maternally-derived associations of this APOB SNP with borderline significance were also seen with fasting glucose and insulin levels in the combined results ( combined p-value = 4 . 0x10-3 and 7 . 4x10-4 for glucose and insulin respectively ) , whereas no associations were observed with the paternally-derived allele ( Table 1 ) . Because APOB is a major component of lipoprotein particles and an important contributor to the atherosclerotic process [23 , 24] , we examined whether there is also evidence for POEs of APOB rs1367117 on other CMR traits , including LDL-C , HDL-C , total cholesterol ( TC ) , triglycerides ( TG ) and systolic and diastolic BP ( S7 Table ) . A maternally-derived association , and not paternally-derived , was observed for SBP in the combined analysis ( combined beta = 1 . 62 , p-value = 1 . 6x10-4; F-test p-value for difference = 3 . 1x10-3 ) . Significant maternal and paternal associations , of similar magnitudes , were shown with both LDL-C ( maternal beta = 4 . 20 , p-value = 2 . 8x10-6; paternal beta = 3 . 54 , p-value = 5 . 0x10-5 ) and TC ( maternal beta = 4 . 67 , p-value = 3 . 6x10-6; paternal beta = 3 . 67 , p-value = 1 . 7x10-4 ) ( F-test p-values for difference between maternal and paternal effects were non-significant ) , reflecting offspring genotype effect , irrespective of the parental source ( additive genetic effect is discussed below ) . To further explore the observed parental-specific associations of APOB with CMR traits , we examined the influence of offspring BMI on these associations . Further adjustment for BMI resulted in attenuation of the observed maternally-derived effects of APOB rs1367117 on waist , SBP , glucose and insulin to the null ( Table 2 ) . In contrast , adjustment for BMI had minimal effect on the maternal and paternal associations with both LDL-C and TC ( S7 Table ) . We have additionally examined the associations of offspring APOB rs1367117 with cardio-metabolic traits , overlooking the parental source of the minor allele , in the four studies separately and combined ( Table 3 ) . The combined data pointed to a substantial additive genotype effect on lipid traits , mainly on LDL-C and TC , which is in agreement with the aforementioned POE results showing both maternal and paternal effects on LDL-C and TC ( S7 Table ) . On the other hand , genotype effects on adiposity traits ( BMI and WC ) were relatively minor both in size and significance , emphasizing the maternal-specific nature of the APOB rs1367117-adiposity association . In an attempt to fine-map the POE observed for APOB rs1367117 on adiposity , we used HapMap tagger [25] and SNAP Proxy Search function [26] to select SNPs that either tag the APOB gene region ( i . e . tag SNPs using r2 thershold≥0 . 8 within the interval spanning the gene itself and 100kb on each side ) or that are in moderate linkage disequilibrium ( LD ) with the SNP of interest ( i . e . r2 values between 0 . 4–0 . 7 ) . Forty two SNPs were identified and we then examined POEs of these selected SNPs on BMI in the 3 studies where genome-wide data was available ( i . e . FHS , FamHS and ERF ) and meta-analyzed the results ( S8 and S9 Tables ) . To illustrate the results , we used a regional plot developed by Saxena et al . [27] ( Fig 2 ) . The plot presents the combined p-values of the maternally-derived effects on BMI , showing rather higher significance for the SNPs that are in higher LD with APOB rs1367117 . As expected , the degree of significance of the maternal-specific associations dropped with the decrease in LD and increase in distance from the SNP of interest . Bioinformatic annotation was undertaken for rs1367117 , located within the 4th exon of apoB , using the Epigenome Browser ( http://epigenomegateway . wustl . edu/; [28 , 29] ) . The region is depicted in Fig 3 . The gene is expressed most abundantly in liver , and far less in small intestine and least in adipose . In liver , there is a broad region with regulatory activity beginning upstream of the promoter and extending past the 4th exon with enhancer activity around exon 4 in both liver and adipose . Therefore , this region displays a variety of regulatory functions . Notably , the DNA methylation in this region spanning exon 4 shows a 50% methylation pattern in liver , consistent with a parent-of-origin effect or allelic-specific methylation; however , the same is not true in adipose and small intestine which show nearly 100% methylation , which might indicate tissue-specific imprinting [30] .
This study tested the hypothesis that including the parental origin of the allele will reveal parent-of-origin genetic effects on adiposity and glycemic traits among young adults . We have shown that a common coding variant in APOB gene has a significant maternally-derived POE on CMR traits , and adiposity in particular . Our findings for APOB gene provide support for the potential importance of POEs on CMR traits . Several findings from linkage studies provide evidence for loci demonstrating POEs on CMR traits and phenotypes [13–17] . Yet despite this evidence , few genetic association studies to date have examined POEs on CMR outcomes . In a study conducted in Iceland , GWAS data were used together with genealogy and long-range phasing to examine POEs on incident Type 2 Diabetes ( T2D ) [18] . Importantly , the Icelandic study demonstrated that the inclusion of the parental source of offspring alleles strengthened the evidence for previously reported associations of T2D with SNPs within imprinted intervals and genes , such as KCNQ1 , and revealed novel associations with T2D . Subsequently , POEs were also detected in this population of Icelanders on age at menarche and thyroid-stimulating hormone levels [19 , 20] , which are related to CMR traits . Other studies have shown that genetic variations in several known imprinted genes , such as IGF2 , INS and GNAS , have been associated with CMR traits , including adult obesity ( BMI ) and T2D [31–36] . GWAS also have identified POEs on body composition in animal models [37 , 38] and on high blood pressure [39] in humans . A novel approach for detecting POEs using genome-wide genotype data of unrelated individuals was recently introduced [40] . Using this method 6 SNPs were identified as having POE on BMI , 2 of which , near genes SLC2A10 and KCNK9 , were replicated in five family studies . POE was not identified for APOB by that study and we were not able to examine SLC2A10 and KCNK9 in our study as genetic data for these genes were not available in JPS . Multiple candidate gene studies ( e . g . reviewed in [41–45] ) and more recently GWAS [8 , 27 , 46–59] identified genetic variants in APOB gene related to lipid and lipoprotein levels , and to coronary artery disease and myocardial infarction . Specifically , APOB rs1367117 , identified here , was shown to be associated with levels of LDL-C , TC , APOB and non-HDL in GWAS [52 , 54] and candidate gene studies [60–62] and recently , an epistatic interaction on BP levels was observed between APOB rs1367117 and VCAM1 rs1041163 [63] . SNP rs1367117 is a nonsynonymous missense variant , located in exon 4 of APOB gene , causing a Thr→Ile amino acid change . A study using samples from both Finnish and Mexican subjects has shown that rs1367117 is associated with serum apoB levels , and that another variant identified in GWAS ( rs7575840 ) , which is in high LD with rs1367117 ( r2 = 0 . 93 in CEU HapMap sample ) , is associated with expression levels of APOB gene in adipose tissue [61] . Obesity is strongly associated with dyslipidemia , which may account for the associated increased risk of atherosclerosis and coronary disease . Although the precise mechanism whereby obesity results in dyslipidemia has not been established , there is some evidence to support that visceral obesity is related to dysregulation of both apoB isoforms [64 , 65] . Only several studies have shown that genetic variants in APOB gene are associated with body size and obesity in children [66] and adults [67–70] and with body growth and obesity in chicken [71] . Our results showing a highly significant additive genetic effect of APOB rs1367117 on lipid traits and a highly significant maternal-specific effect on adiposity traits , in fact demonstrate why genetic associations studies overlooking POE conducted to date have typically identified associations of APOB gene variants with lipid traits but not with adiposity . To our knowledge , APOB gene has not been previously reported as demonstrating POEs on CMR or on other traits nor has it been identified as imprinted ( and neither has its neighboring genes located within 500kb upstream and downstream of APOB , e . g . C2orf43 , GDF7 , HS1BP3 , HS1BP3-IT1 , RHOB , LOC645949 ) . Several explanations may account for the lack of previous support for POEs related to APOB . First , both imprinting and maternal genotype effects on the intrauterine environment are recognized as parent-of-origin mechanisms [10 , 72] , thus further work is needed to determine which of the mechanisms may explain the parent-of-origin associations identified in this work . Specifically , human placenta produces and secretes apoB-containing lipoproteins , and placental lipoprotein formation constitutes a pathway of lipid transfer from the mother to the developing fetus [73] . Whether this pathway may affect adiposity and cardio-metabolic health of adult offspring remains to be explored . Second , it is assumed that not all human imprinted genes have been identified [74–76] and gradually more genes are recognized as being imprinted [77] . Lastly , and perhaps more likely , parent-of-origin associations may be apparent in specific tissues , specific stages of development or in specific genes which are not themselves imprinted but rather regulated by imprinted genes . A recent work by Mott et al . supports the latter by showing that non-imprinted genes can generate parent-of-origin effects by interaction with imprinted loci and that the importance of the number of imprinted genes is likely secondary to their interactions [78] . It has also been suggested that imprinted genetic effects on complex traits are context dependent , and that imprinting patterns may not be consistent among traits and environments or between sexes [79] . For example , a highly significant QTL on chromosome 1 , DMetS1b , was shown to be associated with variation in both serum lipid levels and obesity in mice . Interestingly , for cholesterol there was an additive effect in the full population , whereas for free-fatty acid levels high-fat fed females showed maternal expression imprinting and low-fat fed females showed paternal expression imprinting [80] . Of note , tissue and gender specificity of imprinting has been demonstrated very recently in a study in humans using allele-specific expression data [77] . Similarly , our findings in humans have demonstrated that the same APOB variant has an additive genotype effect on lipids , and yet a maternal POE on adiposity . In addition , context-specificity may underlie the lack of POEs of APOB in the ERF study , possibly due to a specific environmental factor in the Dutch population that could not be accounted for in the models . These hypotheses need to be further explored . The major strengths of this study are the use of common genetic variation in mother-offspring pairs from a population-based cohort with quantitative CMR traits measured at age 32 and the opportunity to replicate our findings in extended pedigrees from three large family studies . Notably , the family design also minimizes confounding due to population stratification [81 , 82] . Additionally , the use of candidate genes identified in GWAS takes advantage of available scientific knowledge as well as reduces the problem of multiple comparisons . There are also several limitations to our study . First , genetic data on fathers were unavailable in JPS resulting in the use of probability-based imputation of the parental source of minor alleles in heterozygous mother-offspring pairs . However , as we have shown in sensitivity analyses , this approach is likely conservative and therefore preferable compared to excluding heterozygous mother-offspring pairs . Furthermore , replication in other studies where genetic data in extended families were used minimizes this source of bias . Second , the APOB variant indentified here has not been previously reported in large GWAS consortia as related with adiposity . Although this is likely the result of GWAS typically ignoring the parental source of alleles due to lack of family data , there is a possibility that our POE results reflect false-positive findings . To address this , we have conducted simulations applying two different approaches , one for a mother-offspring design ( using JPS ) and the other for an extended pedigrees design ( using FHS ) . Based on these simulations , maternal permutation p-values were similar to our observed p-values providing further support for the significance of the reported maternal APOB effect . Lastly , environmental exposures around the time of birth may contribute to the context-specificity of POEs . These data were unavailable for most of the participating studies and therefore their effects could not be assessed in this work . In summary , we have demonstrated that taking into account the parental origin of offspring alleles compared to examining the offspring genotype as a whole may enhance our understanding of genetic associations with CMR traits . Our results highlight the potential contribution of POEs to uncovering complex relations between genetic variants and common traits , and motivate further research in this area , including assessment of the impact of POEs on common traits and investigation of the mechanisms underlying these associations .
Blood was collected in tubes containing EDTA . DNA was extracted from white blood cells using standard salting-out extraction procedures . Common tag SNPs from candidate genes in molecular pathways potentially related to both birth weight and CMR phenotypes were previously genotyped in mother-offspring pairs using Illumina 1536 SNPs panel ( BeadArray technology with a GoldenGate custom panel , Illumina , San Diego , CA , USA [87 , 88] ) . Of the 1536 SNPs genotyped , 1384 ( 90 . 1% ) provided genotype clusters with call rates>95% . Additional quality control measures for genotypic data included: 1 ) testing Hardy-Weinberg Equilibrium; SNPs with chi-square p-value < 0 . 05/1384 were excluded; and 2 ) observing Mendelian inheritance inconsistency between mother’s and offspring genotypes; SNPs with more than 0 . 5% inconsistency were excluded . As a result , an additional 10 SNPs were removed and a total of 1374 SNPs , representing 180 genes , were available . To identify POEs on adiposity and glycemic traits we selected SNPs within genes where previous GWAS in Caucasian populations had revealed variants associated with BMI , glucose and insulin levels , lipids and lipoproteins levels , obesity and type 2 diabetes . This is similar to the approach applied in the Icelandic study in the sense that published GWAS findings were used for targeting variants of interest for POE analysis [18] . Specifically , using the US National Institutes of Health Office of Population Studies catalogue of published genome-wide association studies ( www . genome . gov/gwastudies—accessed July 1 , 2012 ) [89] , out of the 180 genes genotyped previously using the Illumina custom panel in JPS mother-offspring pairs , we identified 18 genes for which significant associations ( p-value<5X10-8 ) with CMR traits were reported in GWAS ( Table 4 and S1 and S2 Tables ) . In these 18 genes 182 tag SNPs were available in JPS ( Table 4 and S3 Table ) . The following adiposity and glycemic traits were measured in offspring at age 32: body mass index ( BMI , calculated by dividing weight ( kg ) by squared height ( m2 ) ) ; waist circumference ( WC , mean of two consecutive measurements at the midpoint between the lower ribs and iliac crest in the midaxillary line ( cm ) ) ; fasting glucose ( mg/dL ) and fasting insulin ( mean of two repeated measures ( mU/mL ) , natural log-transformed ) . In a follow-up analysis limited to replicated findings , we additionally examined lipid and blood pressure traits: low-density lipoprotein cholesterol ( LDL-C ( mmol/L ) ) ; high-density lipoprotein cholesterol ( HDL-C ( mmol/L ) ) , triglycerides ( TG ( mmol/L ) , natural log-transformed ) , systolic and diastolic blood pressure ( SBP and DBP , mean of three consecutive measures ( mmHg ) ) . All cardio-metabolic outcomes were treated as continuous variables ( distributions presented in S4 Table ) . Analyses of JPS data were carried out using Stata 12 . 0 ( StataCorp , College Station TX ) . To assess the separate contribution of the maternally- and paternally-inherited variants to offspring CMR phenotypes , we first determined the parental origin of minor alleles , based on mother-offspring pairs , since genetic data on fathers were not available in JPS . For every given SNP we constructed two variables; maternally-derived minor allele ( M-D ) and paternally-derived minor allele ( P-D ) indicators . These indicators count the number of minor alleles inherited from mother and/or father; each count is a zero or one ( Table 4 ) . When offspring and mother were both heterozygous at a given SNP , the source of the minor allele was ambiguous . But under random mating and transmission equilibrium , the probability that the minor allele was derived from the mother is 1-MAF and from the father is MAF , where MAF = minor allele frequency . Therefore , similarly to use of estimated dosage for uncertain genotypes , for heterozygous mothers-offspring pairs , we used the estimated dosage of M-D and P-D indicators , i . e . 1-MAF and MAF , respectively ( Table 5 and Fig 4 ) . The percentage of heterozygous mother-offspring for each of the 182 selected SNPs ranged between 5% and 27% ( S3 Table ) . When offspring and mother are both heterozygous at a given SNP , the source of the minor allele is ambiguous . For any given SNP with allele frequencies of p and q for the common ( A ) and minor ( a ) alleles respectively , the expected paternal genotype frequencies assuming Hardy-Weinberg Equilibrium are f ( AA ) = p2 , f ( Aa ) = 2pq and f ( aa ) = q2 . Under random mating and transmission equilibrium , the probability that the minor allele was derived from the mother is 1-MAF and from the father is MAF , where MAF = minor allele frequency . Therefore , for heterozygous mothers-offspring pairs , we used the estimated dosage of M-D and P-D indicators , i . e . 1-MAF and MAF . We used linear regression to examine each SNP-trait association . We first assessed the associations of offspring genotype with trait , using an additive genetic model . Then the association of the parental origin of the minor allele on the trait was assessed by including the M-D and P-D indicators together in a single model . The following mean models were used: E[Y]=β0+βGGO+γTZ ( 1 ) E[Y]=β0+βMDGMD+βPDGPD+γTZ ( 2 ) Where Y denotes trait , GO = offspring genotype , GMD = indicator of maternally-derived minor allele , GPD = indicator of paternally-derived allele , and Z = other covariates . Existence of POE in model 2 was assessed via a test of the following null hypotheses: ( 1 ) βMD = 0; ( 2 ) βPD = 0; ( 3 ) βMD = βPD . The significance threshold for rejection of the null hypotheses was set at a within-gene corrected Bonferroni p-value<0 . 05 for the separate effects of the maternally- or paternally derived indicators ( i . e . hypotheses 1 or 2 ) and nominal p-values ( i . e . p-values<0 . 05 ) for the F-test examining the differences between these effects ( i . e . hypothesis 3 ) . Findings that met these criteria for any one of the examined outcomes were moved forward for replication and meta-analysis . All models were adjusted for offspring sex and ethnicity . Following an approach suggested by Thomas and Witte to correct for population stratification in mixed-ethnicity families [90] , ethnicity of offspring was classified based on country of origin of all four grandparents , using nine major ethnicity strata ( Israel , Morocco , Other North Africa , Iran , Iraq , Kurdistan , Yemen , Other Asia and the Balkans and Ashkenazi ) . Rather than allocating offspring to a single ethnicity , we constructed a covariate for each stratum giving the proportion of grandparents derived from each of the nine ethnic groups ( ranging from 0 to 1 , reflecting none or all four grandparents originating from the specific ethnic group , respectively ) and then included these covariates as adjustment variables in a multiple regression , excluding one stratum ( Ashkenazi ) to eliminate complete multicollinearity . To account for the stratified sampling , we used Stata’s ‘pweight’ option to weight estimates by individuals’ inverse probability of being sampled [91] . Findings exceeding the aforementioned threshold of significance in JPS were followed up in three additional family studies with extended pedigrees: 1 ) Framingham Heart Study ( FHS ) ( max N = 2225 probands ) ; 2 ) Family Heart Study ( FamHS ) ( max N = 773 probands ) ; and 3 ) Erasmus Rucphen Family ( ERF ) study ( max N = 631 probands ) . Probands were restricted to individuals of European decent , aged≤50 , with available genetic and CMR data and genetic data in at least one parent . Probands on medication to treat diabetes , lipid-lowering medication or BP-lowering medication were excluded from the corresponding analyses . Detailed description of each of the studies , including study sample , genotyping techniques and CMR trait measurements and distributions , is provided in S1 Text and S4 Table . Distributions of CMR traits were comparable between the studies , with the exception of fasting insulin in FHS . Yet substantial differences in fasting insulin distributions across studies are commonly observed [92 , 93] . We used the quantitative transmission disequilibrium test ( QTDT ) software [94] to test for the association of maternally and paternally inherited minor alleles using information from the extended families within each study , adjusting for sex and age , using a model similar to the one used in JPS . Specifically , qtdt -at -ot was used to test for POE ( maternal effect = paternal effect ) , and qtdt -at -om and qtdt -at -op were used to test maternal and paternal effects , respectively ( modified to adjust for other parent contribution ) . QTDT software uses a linear mixed effect model to account for familial correlations . We did not use the QTDT statistic that is robust to population stratification because , based on available genome-wide data , population stratification did not affect the traits evaluated in most studies . In limited instances where there was some suggestive evidence for the presence of population stratification , we corrected for it using principal components analysis within the relevant study [95] . Additionally , z-based meta-analysis combining results from all 4 studies ( i . e . JPS , FHS , FamHS and ERF ) was conducted , as a joint analysis approach was shown to be more efficient than a two-stage approach for genetic association studies [96] . A Bonferroni correction was applied and SNPs were declared statistically significant if their p-values were below 0 . 05 divided by the total number of SNPs initially tested ( N = 182 ) , i . e . p-value<2 . 75x10-4 . We also applied an inverse variance weighted meta-analysis approach to obtain a pooled estimate of effect sizes ( betas ) across studies . Since the QTDT approach does not provide estimates of standard errors ( SEs ) , SEs were estimated by converting p-values into a z-statistic and setting: SE = beta/z . The JPS study was approved by the Institutional Review Board of the Hadassah-Hebrew University Medical Center ( approval #10–01 . 04 . 05 ) and the University of Washington Human Subject Review Committee ( approval #31032 ) . FHS was approved by Boston University Medical Center Institutional Review Board ( approvals #H–32132 , #H–26671 and #H–28859 ) , FamHS was approved by the Institutional Review Board of the Washington University School of Medicine in St . Louis ( approval #201403014 ) and ERF study was approved by the Medical Ethics Committee of Erasmus MC , Rotterdam ( approval #MEC 213 . 575/2002/114 ) . All participants provided written informed consent for genetic studies . | To date , genetic variants identified in large-scale genetic studies using recent technical and methodological advances explain only a small proportion of the genetic basis of obesity , diabetes and other cardiovascular risk factors . These studies were typically conducted in samples of unrelated individuals . Here we utilize a family-based approach to identify genetic variants associated with obesity-related traits . Specifically , we examined the separate contribution of maternally- vs . paternally-inherited common genetic variants to these traits . By examining 1250 young adults and their mothers from Jerusalem , we show that a specific genetic variant , rs1367117 , located in the APOB gene on chromosome 2 is related to body mass index and waist circumference when inherited from mother and not from father . This maternal effect is not restricted to Jerusalemites , but is also seen in a large sample of individuals of European descent from independent family studies worldwide . Our findings provide support of the role of complex genetic mechanisms in obesity , and highlight the benefit of utilizing family studies for uncovering genetic pathways underlying common risk factors and diseases . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| []
| 2015 | Parent-of-Origin Effects of the APOB Gene on Adiposity in Young Adults |
Dengue virus ( DENV ) infection causes dengue fever , dengue hemorrhagic fever and dengue shock syndrome . It is estimated that a third of the world’s population is at risk for infection , with an estimated 390 million infections annually . Dengue virus serotype 2 ( DENV2 ) causes severe epidemics , and the leading tetravalent dengue vaccine has lower efficacy against DENV2 compared to the other 3 serotypes . In natural DENV2 infections , strongly neutralizing type-specific antibodies provide protection against subsequent DENV2 infection . While the epitopes of some human DENV2 type-specific antibodies have been mapped , it is not known if these are representative of the polyclonal antibody response . Using structure-guided immunogen design and reverse genetics , we generated a panel of recombinant viruses containing amino acid alterations and epitope transplants between different serotypes . Using this panel of recombinant viruses in binding , competition , and neutralization assays , we have finely mapped the epitopes of three human DENV2 type-specific monoclonal antibodies , finding shared and distinct epitope regions . Additionally , we used these recombinant viruses and polyclonal sera to dissect the epitope-specific responses following primary DENV2 natural infection and monovalent vaccination . Our results demonstrate that antibodies raised following DENV2 infection or vaccination circulate as separate populations that neutralize by occupying domain III and domain I quaternary epitopes . The fraction of neutralizing antibodies directed to different epitopes differs between individuals . The identification of these epitopes could potentially be harnessed to evaluate epitope-specific antibody responses as correlates of protective immunity , potentially improving vaccine design .
Dengue virus ( DENV ) is a single stranded positive sense RNA virus that is transmitted by the Aedes mosquito [1] . There are four distinct DENV serotypes ( DENV1-4 ) , and infection results in a range of symptoms , from fever and rash to the more serious dengue hemorrhagic fever and dengue shock syndrome . Over a third of the world’s population is at risk for infection , and there are an estimated 390 million infections yearly [1] . A primary infection with DENV results in the induction of serotype cross-neutralizing antibodies which can provide temporary serotype cross-protective immunity that is not maintained [2] . Over the course of the following year , these cross-reactive neutralizing antibodies wane , leaving individuals susceptible to infection by the remaining three heterologous serotypes [3] . Serotype-specific neutralizing antibodies are maintained in circulation for decades following exposure and may play a critical role in providing subsequent protection against the infecting serotype [2 , 4 , 5] . While antibodies are known to play a key role in protection against DENV reinfection [6] , it has also been shown that CD8+ T-cells [7 , 8] , CD4+ T-cells [9] , and other mechanisms of cellular immunity are important for protection [10 , 11] . The leading DENV vaccines are tetravalent formulations , designed to elicit independent , hopefully protective , neutralizing antibodies against all four serotypes simultaneously [12] . Phase 3 efficacy trials in Asia and Latin America showed that the recently licensed tetravalent vaccine , Dengvaxia , had variable efficacy depending on immune status prior to vaccination and the serotype of infection [13 , 14] . In mixed populations of susceptibles and DENV-immunes , Dengvaxia was 50–80% efficacious against DENV1 , DENV3 and DENV4 , but only 35–42% against DENV2 [13 , 14] . Vaccine efficacy was significantly lower in those persons seronegative to DENV compared to individuals who were DENV seropositive at the time of vaccination [13 , 14] . Moreover , younger vaccinated individuals were hospitalized for DENV more frequently than their unvaccinated counterparts , suggesting that poor immunogenicity in naïve subjects might place individuals at a greater risk of developing severe disease as antibody levels decline over time [15–18] . Indeed , based on long-term follow up data , Dengvaxia is no longer recommended for use in DENV-naive individuals [17] . The Dengvaxia clinical trials have revealed that even individuals with detectable neutralizing antibodies to a particular serotype experienced vaccine break-through infections indicating the mere presence of antibodies that neutralize infection in cell culture assays is not sufficient for protection [19] . Therefore , in addition to the level of neutralizing antibodies to each serotype , it is critical to define other properties of human antibodies potentially responsible for durable , protective immunity . Importantly , while Dengvaxia contains the structural proteins from DENV , the non-structural proteins are from yellow fever virus . It is therefore possible that sufficient T-cell immunity towards DENV epitopes was not achieved [16 , 20] . DENV vaccines that contain DENV non-structural proteins might generate a more robust T-cell response , and therefore more closely resemble a natural DENV infection , which results in a protective immune response [21] . The DENV envelope glycoprotein ( E ) ectodomain , which is comprised of three domains ( EDI , EDII and EDIII ) is the major target of neutralizing antibodies [22] . Two E monomers form a dimer in a head-to-tail arrangement , three dimers form a raft , and thirty rafts ( 180 monomers ) cover the entire surface of the virus [23] . Our group has previously characterized components of the antibody response to DENV1 , DENV2 , DENV3 and DENV4 infection by mapping the epitopes of strongly neutralizing human monoclonal antibodies ( hMAbs ) [24 , 25] . Importantly , many of these strongly neutralizing hMAbs target quaternary structure epitopes that form as the envelope glycoprotein is assembled on the virus surface [24–31] . In addition , we have demonstrated that we can transplant these quaternary epitopes between DENV serotypes and maintain their biological functions [25 , 30 , 32 , 33] . While determining the properties of individual hMAbs is valuable , complex polyclonal antibody response governs protection against subsequent infection . Importantly , the epitope of a single DENV2 serotype-specific hMAb , 2D22 , was shown by our group to be targeted by a large fraction of DENV2 neutralizing antibodies in many , but not all individuals after recovery from DENV2 infections , highlighting the potential role of this epitope in protective immunity [25] . Despite this , there are additional DENV2 hMAbs that use other epitopes within E , suggesting that there are potentially multiple neutralizing antibody epitopes for each serotype . The goals of this study are to identify novel neutralizing epitopes in DENV2 , to develop robust diagnostic reagents for evaluating epitope specific responses with recombinant DENVs ( Fig 1 ) and to evaluate the role of these novel and previously described epitopes as targets of polyclonal serum antibodies induced by natural infections and DENV vaccines .
To characterize the epitopes of DENV2 human monoclonal antibodies ( hMAbs ) , we used a panel of three DENV2-specific , strongly neutralizing hMAbs ( S1 Table ) . The hMAbs were isolated from two donors infected in geographically distinct locations with different DENV2 genotypes [34] . The three hMAbs , 3F9 , 2D22 and 1L12 , bound to whole DENV2 virus ( Fig 2A ) . Recently , it has been reported that human antibodies that strongly neutralize DENVs bind to quaternary structure epitopes displayed on E homo-dimers or higher order surface structures required for virion assembly [24 , 25 , 29 , 35] . Consonant with previously published results , DENV2 hMAb 2D22 did not bind rE or rEDIII , confirming the quaternary epitope specificity ( Fig 2B and 2C ) . HMAb 1L12 was similar and did not bind to rE or rEDIII ( Fig 2B and 2C ) . In contrast , hMAb 3F9 weakly bound to rE ( Fig 2B ) . Because hMAb 3F9 bound well to DENV2 virions and weakly to rE , it is likely that the epitope is dependent on E protein assembly into virions for optimal display . In a blockade of binding assay , 2D22 interfered with 1L12 for binding to DENV2 , suggesting that they recognize proximal or overlapping epitopes on the viral envelope ( Fig 2D ) . In contrast , 3F9 only partially blocked the binding of 2D22 ( Fig 2D ) indicating the two hMAbs recognize distinct epitopes on the viral envelope . The cryo-EM structure of hMAb 2D22 Fab in complex with DENV2 has been solved [28] and the footprint of the antibody spans EDIII and EDII of two E molecules forming each homo-dimer . Although 2D22 did not bind rEDIII ( Fig 2C ) , the antibody binds and neutralizes a DENV4 virus containing the entire EDIII from DENV2 ( rDENV4/2-EDIII ) ( Fig 3 ) [25 , 28] . Introducing a single point mutation into this virus ( rDENV4/2-EDIII R323G ) , previously identified as a 2D22 escape mutation [24] , ( Fig 3A ) , resulted in a loss of binding and neutralization ( Fig 3B and 3C ) , confirming 2D22 uses the transplanted EDIII region . HMAb 1L12 , which was isolated from a different donor , showed nearly identical phenotypes , where it gained binding and neutralization to rDENV4/2-EDIII , indicating that it uses EDIII as part of its complex quaternary epitope ( Fig 3D and 3E ) . Similarly , the R323G mutation in rDENV4/2-EDIII results in complete loss of 1L12 binding and neutralization ( Fig 3D and 3E ) . In addition to binding highly conserved residues in EDII , cryo-EM studies predict that hMAb 2D22 interacts with eight ( 307 , 309 , 310 , 316 , 318 , 362 , 363 , 364 ) surface-exposed amino acids in DENV2 EDIII [28] , five ( 307 , 309 , 316 , 362 , 364 ) of which differ between DENV2 and DENV4 . To refine the map coordinates of 2D22 and 1L12 epitopes , we generated a new EDIII recombinant virus in which these five amino acids in DENV4 were replaced with those from DENV2 ( rDENV4/2-EDIII 5aa ) ( Fig 4A ) . 2D22 was able to partially bind and neutralize this virus at high concentrations of antibody ( Fig 4B and 4C ) . Because the gain in function is only partial , these data suggest that the epitope requires other critical residues in EDIII for maximal binding and neutralization . In contrast , hMAb 1L12 did not bind or neutralize rDENV4 –EDIII 5aa ( Fig 4D and 4E ) , suggesting that its epitope overlaps with 2D22 but engages a different set of residues on EDIII . Competition assays with 2D22 indicated that 3F9 binds to an epitope that has minimal if any overlap with 2D22 or IL12 ( Fig 2D ) . To map the epitope of hMAb 3F9 , we evaluated its binding to a panel of chimeric recombinant DENVs ( rDENVs ) with alterations in specific domains ( Fig 5A ) . Our group has previously shown that strongly-neutralizing hMAbs for DENV1 and DENV3 use the EDI/II hinge region in their epitope [24 , 26 , 27 , 32] . hMAb 3F9 bound to a DENV2 virus that had the EDI/II hinge residues replaced with those from DENV4 ( rDENV2/4-EDI/II ) , suggesting it does not use this region in its epitope ( Fig 5B ) . Conversely , hMAb 3F9 lost most binding to and neutralization of a DENV2 with 11 of its EDI residues replaced with those from DENV4 ( rDENV2/4-EDI ) , suggesting 3F9 uses an epitope that contains the replaced residues located in EDI ( Fig 5B and 5C ) . To further characterize the 3F9 epitope , we tested its binding to and neutralization of a DENV4 virus that contained 22 surface-exposed EDI residues from DENV2 ( rDENV4/2-EDI ) ( Fig 5D ) . 3F9 bound to and neutralized the EDI transplant virus , confirming EDI as the main target of this hMAb ( Fig 5E and 5F ) , however gain of binding was not complete , suggesting there are other residues that are required for maximal binding . Our data underscore the importance of cryo-EM analyses to help elucidate the complete 3F9 binding epitope . In summary , these studies define the location of epitopes recognized by DENV2 type-specific neutralizing hMAbs 2D22 , 1L19 and 3F9 . Both 2D22 and 1L19 bind to proximal but distinct quaternary epitopes centered on EDIII . hMAb 3F9 , on the other hand , binds to an epitope on EDI of the E protein . To determine if epitopes defined using hMAbs were targets of polyclonal serum neutralizing antibodies , we first performed competition ( blockade of binding ) assays with human immune sera and hMAbs . Convalescent immune sera from primary DENV2 cases effectively blocked the binding of 2D22 to its epitope ( Fig 6A ) . Under identical conditions of treatment , DENV1 or DENV3 immune sera did not block 2D22 from binding , confirming that primary DENV2 infection elicited a 2D22-like serotype-specific antibody response ( Fig 6A ) . The same DENV2 immune sera also blocked 3F9 from binding to its epitope , whereas control DENV1 and DENV3 sera did not ( Fig 6B ) . Remarkably , the ratio of antibodies targeting the two epitopes appeared to differ across individuals . Two individuals ( DT001 and DT158 ) were more effective at blocking 2D22 binding than 3F9 binding , whereas DT134 and DT155 were more effective at blocking 3F9 than 2D22 ( Fig 6A and 6B ) , suggesting these individuals had different ratios of antibodies targeting each epitope . DENV2 monovalent vaccine sera also blocked 2D22 and 3F9 binding to their respective epitopes ( Fig 6C and 6D ) , indicating that natural infection and monovalent DENV2 vaccine-elicited antibodies target both these epitopes . Interestingly , no DENV2 sera samples were able to completely inhibit either 2D22 or 3F9 from binding to their epitopes , suggesting a limit in the amount these blocking antibodies are present in the sera . Next , we performed studies to determine if 2D22 and 3F9 epitopes were targets of DENV2 neutralizing serum antibodies . Epitope exchanged recombinant viruses not only provide an approach to map hMAbs , but they can also be used to quantify epitope-specific neutralizing antibodies in immune sera . To measure the amount of neutralizing antibodies targeting 2D22 and 3F9 epitopes , we evaluated the ability of polyclonal DENV2 immune sera ( 10 samples ) or vaccine sera ( 9 samples ) to neutralize rDENV4/2-EDIII ( Fig 3A ) and rDENV4/2-EDI ( Fig 5D ) viruses . Consistent with previous results [25] , a large fraction of DENV2 neutralizing antibodies tracked with DENV2 EDIII displayed on the rDENV4/2-EDIII virus ( Fig 7A ) . Interestingly , most individuals also had neutralizing antibodies that tracked with the DENV2 EDI epitope displayed on the rDENV4/2-EDI virus ( Fig 7A ) . In some individuals ( e . g . DT155 ) there are similar levels of neutralizing antibodies that target both epitopes , whereas in other individuals ( e . g . DT128 ) few if any neutralizing antibodies target the 3F9 EDI epitope ( Fig 7A ) . In individuals that received a monovalent DENV2 vaccine , the majority of their neutralizing antibodies target EDIII with a much smaller fraction of the response targeting the EDI epitope ( Fig 7B ) . Overall , there is higher tracking of DENV2 specific responses with both the EDIII and EDI epitopes ( 80% and 54% respectively ) in the natural infection sera , as compared with the vaccine sera ( 69% and 30% respectively ) , suggesting vaccination elicits a slightly different antibody response ( Table 1 ) .
People infected with DENVs develop robust and durable antibody responses that contribute to protection against re-infection against the homologous serotype; however , rare instances of re-infection with the same serotype do occur [36] . Antibodies that neutralize DENVs in cell-culture assays have been considered to be surrogates of protective immunity in vivo . However , this assumption has been challenged by recent results from DENV vaccine trials . Most notably , people who received a tetravalent live attenuated DENV vaccine and developed neutralizing antibodies experienced DENV2 breakthrough infections [15] . Breakthrough infections were also documented with the other serotypes despite the presence of neutralizing antibodies [15] . This landmark vaccine trial has established that the presence of cell-culture neutralizing antibodies identified using FRNT assays , is not predictive of protection . Indeed , breakthrough DENV infections of vaccinated seronegative children underscore the urgency to understand the essential mechanisms of immune protection in DENV . Moving forward , we need to define key epitopes on DENVs targeted by neutralizing and potentially protective antibodies and develop assays to measure both the level and the molecular specificity of neutralizing antibodies . In this study , we used a panel of hMAbs , human DENV polyclonal immune sera , and recombinant DENVs ( Fig 1 ) to map the location of epitopes recognized by DENV2 neutralizing antibodies . First , we used three DENV2 type-specific and strongly neutralizing hMAbs to map epitopes . hMAbs 2D22 and 1L12 isolated from different people had similar properties and recognized overlapping quaternary epitopes centered on EDIII . Recently Fibriansah et . al . determined the cryo-EM structure of 2D22 bound to DENV2 and demonstrated that the footprint of the 2D22 spanned EDIII and EDII of two E proteins forming a single homo-dimer [28] . Our data indicating that 2D22 recognizes an EDIII centered quaternary epitope are entirely consistent with the footprint determined by Fibriansah et . al . We suspect that 1L12 also binds a similar but not identical epitope because of subtle differences in the binding of 2D22 and 1L12 noted in this study . These findings highlight the importance of cryo-EM studies with IL12 , which would provide a more comprehensive view of this larger DENV2 antigenic site . Nevertheless , our observation that two individuals infected with different DENV2 genotypes produced type-specific neutralizing hMAbs targeting a similar region suggests that EDIII is a dominant target of DENV2 neutralizing antibodies . The DENV1 , 3 and 4 type-specific , neutralizing hMAbs identified to date do not map to the regions defined by 2D22 and 1L12 indicating that major targets to type-specific neutralizing Abs can differ between serotypes . However , several DENV serotype cross-neutralizing hMAbs that bind across the E homo-dimer have been described recently [29] . While these E dimer-dependent epitope ( EDE ) hMAbs partially overlap with the 2D22 epitope , they recognize patches that are highly conserved between serotypes unlike 2D22 . HMAb 3F9 and 1L12 , which were isolated from the same person , have distinct epitopes , consistent with bivalent recognition of the EDIII and EDI DENV epitopes in most DENV polyclonal immune sera . The 3F9 epitopes is centered on EDI at a site that overlaps with known DENV1 and DENV4 neutralizing hMAbs [26 , 37 , 38] . Therefore , unlike 2D22 , the region recognized by 3F9 is targeted by type-specific neutralizing antibodies to other serotypes as well . Our previous work demonstrated that a majority of the polyclonal antibody response following DENV2 infection and vaccination appeared to be directed to a quaternary EDIII epitope [25] . In some individuals however , neutralization titers did not track as strongly with this epitope , suggesting that two or more neutralizing epitopes are targeted disproportionately after primary DENV2 infections . We propose that the EDI epitope defined by the hMAb 3F9 represents a second major neutralizing epitope on DENV2 . Most individuals with naturally acquired DENV2 infections contained antibodies targeting both epitopes however some individuals targeted only one epitope , or had a skewed response . Similar results were observed in DENV2 vaccinated individuals , where there were antibodies targeting each epitope , however the overall response is dominant to the EDIII epitope . Overall , there was a higher response of antibodies tracking with the EDIII than the EDI epitope in both the natural infection and vaccinated sera ( Table 1 ) . Interestingly , some individuals had complete neutralizing antibody responses tracking with both epitopes , suggesting that they potentially generated redundant populations of antibodies . Generating populations of antibodies directed to different regions on E could be an important component of an effective antibody response . Viruses can mutate to escape antibody pressure , but simultaneously escaping antibody pressure to multiple sites on E would be more challenging [39 , 40] . As some individuals appear to mount preferential responses to one site or the other after natural infection or vaccination , it is possible that strains with natural variation within one of these epitopes may allow for repeat or breakthrough DENV2 infections . Without a clear understanding of what constitutes a protective DENV antibody response to each serotype , it is challenging to evaluate current DENV vaccines . By defining the epitopes targeted by DENV2 hMAbs and polyclonal sera , we hope to determine if there are antibody based correlates of protection and use these to evaluate current vaccines in the pipeline , and inform the design of next-generation vaccines . Using recombinant DENVs that contain both gain of function and loss of function epitopes , we can rapidly map in high-throughput assays the epitopes of large panels of hMAbs , prioritizing targets for crystallographic studies and downstream analyses .
Recombinant viruses were constructed using a four-cDNA cloning strategy . The DENV genome was divided into four fragments , and subcloned into separate cDNA plasmids with unique type IIS restriction endonuclease cleavage sites at the 5’ and 3’ ends of each fragment . A T7 promoter was introduced into the 5’ end of the A fragment . Plasmid DNA was grown in Escherichia coli cells , digested with the corresponding enzymes , gel purified , ligated together with T4 DNA ligase and transcribed with T7 polymerase to generate infectious genome-length capped viral RNA transcripts . RNA was electroporated into C6/36 cells , cell culture supernatant containing virus was harvested and passaged onto C6/36 cells to generate a passage one virus stock . C6/36 cells ( ATCC CRL-1660 ) were grown in Gibco minimal essential medium ( MEM ) at 32°C . Vero-81 cells ( ATCC CCL-81 ) were maintained in Dulbecco’s modified Eagle’s medium ( DMEM ) at 37°C . Media were supplemented with fetal bovine serum ( FBS ) ( 10% for Vero-81 and 5% for C6/36 ) which was lowered to 2% after infection . C6/36 media were supplemented with nonessential amino acids . All media were additionally supplemented with 100U/ml penicillin , 100μg/ml streptomycin and 0 . 25μg/ml Amphotericin B . All cells were incubated in 5% CO2 . Human dengue immune sera used in this study were obtained from a previously described Dengue Traveler collection at University of North Carolina , and were all primary DENV2 natural infections [24 , 25 , 30] . Vaccine sera were obtained from individuals who received a live-attenuated monovalent DENV2 vaccine as developed by the US National Institutes of Health ( NIH ) and were provided by Anna Durbin and Stephen Whitehead . All human sera samples were obtained under Institutional Review Board approval and were anonymized . One day prior to inoculation , 24-well cell culture plates were seeded with 5x104 Vero-81 cells . Virus stocks were serially diluted 10-fold then added to cells ( after growth media was removed ) for one hour at 37°C . After incubation , cells were overlaid with 1% methylcellulose in OptiMEM I ( Gibco ) supplemented with 2% FBS , nonessential amino acids and 100U/ml penicillin , 100μg/ml streptomycin and 0 . 25μg/ml Amphotericin B , and incubated at 32°C . After four days incubation , overlay was removed , cells were washed with phosphate-buffered saline ( PBS ) and fixed in 80% methanol . Cells were blocked in 5% non-fat dried milk ( blocking buffer ) then incubated with anti-prM MAb 2H2 and anti-E MAb 4G2 diluted in blocking buffer . Cells were washed with PBS , then incubated with horseradish peroxidase ( HRP ) -conjugated goat anti-mouse antibody ( Sigma ) diluted in blocking buffer . Plates were washed and foci were developed using TrueBlue HRP substrate ( KPL ) . For whole DENV ELISA , plates were coated with 100ng/well mouse MAb 4G2 and 2H2 overnight at 4°C . Plates were washed with Tris-buffered saline with 0 . 05% Tween ( TBST ) and blocked in 3% non-fat dried milk in TBST ( blocking buffer ) , and equal quantities of virus ( as previously titrated by ELISA using cross-reactive polyclonal DENV immune sera ) were added and incubated for 1 hour . For rE and rEDIII ELISA , plates were directly coated with protein and incubated . Plates were washed and primary human MAbs were diluted in blocking buffer and added to plate for 1 hour . Plates were washed and alkaline phosphate ( AP ) -conjugated secondary antibodies were added for 1 hour . Plates were washed , developed using p-nitrophenyl phosphate substrate and color changes were quantified by spectrophotometry . Assays were developed until OD values were within linear range of the assay , therefore absolute OD values may vary between graphs . All binding assays are based on two experiments performed in duplicate . Plates were coated with antibody , blocked , and virus was captured as described above . DENV polyclonal immune sera were depleted of cross-reactive antibodies as described previously [30] . Briefly , sera were incubated with beads coated with purified DENV4 antigen , then beads were pelleted to removed cross-reactive antibodies bound to bead:antigen complexes . Cross-reactive depleted sera were then diluted 1:10 in blocking buffer and incubated for 1 hour . Plates were washed and alkaline phosphate ( AP ) -conjugated 2D22 ( 100ng/well ) or 3F9 ( 50ng/well ) were added for 1 hour . Plates were developed as described above . Percent blockade was calculated as follows = ( 100-[OD of sample/OD of negative control]*100 ) . Blockade of binding assays are based on two experiments performed in duplicate . For the focus reduction neutralization test ( FRNT ) , hMAbs were diluted 4-fold and mixed with ~45 focus-forming units ( FFU ) of virus , and incubated for 1 hour at 37°C . After incubation , virus:hMAb mixture was added to Vero-81 cells for 1 hour at 37°C or C6/36 cells for 1 hour at 32°C , then overlay was added and cells were incubated and fixed and stained as described above . Foci were counted and FRNT50 titers were calculated as the concentration of antibody or sera dilution factor required to neutralize 50% of the virus . Neutralization assays are based on two ( HMAbs ) or one ( immune sera ) experiments performed in triplicate . | Dengue viruses ( DENV ) are flaviviruses transmitted by mosquitos . There are approximately 390 million DENV infections every year , making dengue virus a major global public health concern . While there is a recently licensed DENV vaccine , it has low efficacy against preventing DENV2 infections . Individuals that are naturally infected with DENV2 generate neutralizing antibodies that can be protective against reinfection with DENV2 . By studying three of these neutralizing antibodies , we found that they bind to two different locations on the surface of the virus . Additionally we found that most individuals that were naturally infected with DENV2 , have antibodies circulating in their blood that target both of these regions . People who were vaccinated against DENV2 also make antibodies targeting both of these sites , suggesting they might also be protected against DENV2 infection . These studies reveal that human antibodies against DENV2 target the same two regions across multiple individuals . Additionally , for a DENV2 vaccine to be protective , it may be important to elicit antibodies directed to these regions as well . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"blood",
"serum",
"dengue",
"virus",
"medicine",
"and",
"health",
"sciences",
"chemical",
"characterization",
"immune",
"physiology",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"body",
"fluids",
"pathogens",
"immunology",
"microbiology",
"viruses",
"vaccines",
"preventive",
"medicine",
"rna",
"viruses",
"infectious",
"disease",
"control",
"antibodies",
"immunologic",
"techniques",
"vaccination",
"and",
"immunization",
"antibody",
"response",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"immunoassays",
"binding",
"analysis",
"immune",
"response",
"biochemistry",
"blood",
"anatomy",
"flaviviruses",
"physiology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"vaccine",
"development",
"immune",
"serum",
"organisms"
]
| 2018 | Human dengue virus serotype 2 neutralizing antibodies target two distinct quaternary epitopes |
Despite elimination efforts , the number of Mycobacterium leprae ( M . leprae ) infected individuals who develop leprosy , is still substantial . Solid evidence exists that individuals living in close proximity to patients are at increased risk to develop leprosy . Early diagnosis of leprosy in endemic areas requires field-friendly tests that identify individuals at risk of developing the disease before clinical manifestation . Such assays will simultaneously contribute to reduction of current diagnostic delay as well as transmission . Antibody ( Ab ) levels directed against the M . leprae-specific phenolic glycolipid I ( PGL-I ) represents a surrogate marker for bacterial load . However , it is insufficiently defined whether anti-PGL-I antibodies can be utilized as prognostic biomarkers for disease in contacts . Particularly , in Bangladesh , where paucibacillary ( PB ) patients form the majority of leprosy cases , anti-PGL-I serology is an inadequate method for leprosy screening in contacts as a directive for prophylactic treatment . Between 2002 and 2009 , fingerstick blood from leprosy patients’ contacts without clinical signs of disease from a field-trial in Bangladesh was collected on filter paper at three time points covering six years of follow-up per person . Analysis of anti-PGL-I Ab levels for 25 contacts who developed leprosy during follow-up and 199 contacts who were not diagnosed with leprosy , was performed by ELISA after elution of bloodspots from filter paper . Anti-PGL-I Ab levels at intake did not significantly differ between contacts who developed leprosy during the study and those who remained free of disease . Moreover , anti-PGL-I serology was not prognostic in this population as no significant correlation was identified between anti-PGL-I Ab levels at intake and the onset of leprosy . In this highly endemic population in Bangladesh , no association was observed between anti-PGL-I Ab levels and onset of disease , urging the need for an extended , more specific biomarker signature for early detection of leprosy in this area . ClinicalTrials . gov ISRCTN61223447
Leprosy is an infectious disease caused by Mycobacterium leprae ( M . leprae ) , which causes damage to the skin and peripheral nerves[1] . The highest numbers of new leprosy cases are detected in India ( 127 , 326 in 2015 ) , Brazil ( 26 , 395 in 2015 ) and Indonesia ( 17 , 202 in 2015 ) [2] . Bangladesh also has highly endemic areas , with a number of new cases of above 3 , 000 per year[2] . Although leprosy prevalence has decreased tremendously along with the widespread availability of multidrug therapy ( MDT ) in endemic areas , detection of new cases worldwide has shown only a modest decline in the last five years , and has stabilized in some countries[3] . In Bangladesh , the number of new cases was 3 , 976 in 2015 , compared to 3 , 848 new cases in 2010[2] . Indirect evidence indicates that worldwide millions of unreported cases linger undetected as a gradual result of a decline in leprosy control activities after the disease was declared eliminated[1] . The continued transmission is probably largely due to M . leprae infected individuals , carrying substantial numbers of bacteria but ( yet ) lacking clinical symptoms . Thus , early detection and subsequent ( prophylactic ) treatment of asymptomatically infected individuals as well as subclinical disease is essential to reduce transmission . Diagnosis of leprosy is still largely dependent on clinical signs and symptoms and detection of acid fast bacteria . However , user friendly lateral flow assays provide new possibilities for rapid diagnosis of leprosy patients in early stages of the disease or of M . leprae infected individuals without any symptoms[4 , 5] . Such assays are likely to contribute to reduction of current diagnostic delay in endemic areas and also aid classification of leprosy disease , allowing appropriate treatment . Currently , there is no specific and sensitive test available that can detect asymptomatic M . leprae infection or predict progression to clinical disease[6] . In view of the long incubation time of leprosy ( typically 3–5 years ) as well as its low incidence , identification of predictive biomarkers requires longitudinal monitoring of M . leprae-specific immunity in those at risk of developing disease . Therefore , investments in large-scale longitudinal follow-up studies , allowing intra-individual comparison of immune profiles in leprosy patients’ contacts , is essential to evaluate which markers correlate with progression to disease and may be used as predictive biomarkers . M . leprae phenolic glycolipid I ( PGL-I ) is an extensively studied antigen on the outer surface of the mycobacterium[7] . The existence of high levels of IgM antibodies to PGL-I[5–7] , has allowed the development of several tests that were investigated broadly for diagnostic purposes[7–10] . Although useful for identifying multibacillary ( MB ) leprosy patients , anti-PGL-I antibody ( Ab ) titers have little value in detecting PB leprosy patients , since the latter develop cellular rather than humoral immunity and therefore often lack antibodies to PGL-I5 . In a previously conducted cluster randomized controlled trial , designated the COLEP study , the effect of single dose rifampicin versus placebo in preventing leprosy in close contacts of newly diagnosed leprosy patients was studied between 2002 and 2009 in a leprosy endemic area in the Northwest of Bangladesh[11 , 12] . To investigate whether anti-PGL-I Ab seropositivity can be used as a predictive biomarker for progression to leprosy in contacts , the current study compared anti-PGL-I Ab levels of the prospective cohort at intake and at three time points covering six years of follow-up per contact .
Contacts of leprosy patients were voluntarily recruited as part of the COLEP study ( a cluster randomized controlled trial ) in 2002 and 2003 in the districts Rangpur and Nilphamari in the northwest of Bangladesh , which is a leprosy endemic area[11 , 12] . Eligible participants ( patients and contacts ) were informed verbally about the study and invited to participate . Written consent was obtained from all participants at recruitment or from the parent or guardian of under 18s . Contacts were followed prospectively from 2002/2003 to 2008/2009 for the development of leprosy . Blood samples were collected by spotting on Whatman filter paper ( Sigma ) and subsequently stored at -80°C . Blood samples were collected at 4 time points: recruitment into the study , follow-up 1 ( FU1; two years after intake ) , follow-up 2 ( FU2; four years after intake ) and follow-up 3 ( FU3; six years after intake ) [12] . Leprosy was diagnosed when at least one of the following signs was present: one or more skin lesions with sensory loss , thickened peripheral nerves , or a positive skin smear result for acid-fast bacilli . Patients with negative smear results and no more than five skin lesions were classified as PB leprosy , and those with a positive smear or more than five skin lesions as MB leprosy[12] . Clinical and demographic data was collected in the COLEP study database[11] . A random sample was taken from 28 , 092 contacts of leprosy patients recruited within the COLEP study[11] . A total of 239 contacts developed leprosy within the six years of follow-up . 25 contacts were included into this sub-study who were diagnosed with leprosy at either FU1 , FU2 or FU3 and for whom filter papers of at least three different time points were available . Out of the contacts who did not develop leprosy , 199 were randomly included using the RAND formula ( Excel 2010 ) , aiming for an equal ratio of three age groups ( 0–14 , 15–29 , and 30+ years ) . The COLEP study represents a unprecedented field trial for leprosy , because it includes valuable longitudinal analysis of contacts and thus is uniquely suited to identify the predictive value of biomarkers . However , the COLEP study did not collect blood samples from contacts as the only samples collected was blood on filter paper . Therefore , this limited biomarker analysis to anti-PGL-I Ab only . In this part of the country , the new case detection rate of leprosy was 3 . 21 per 10 , 000 in 2002 ( DBLM Annual Report 2002 ) . In these cases leprosy was diagnosed by active and passive case detection . In 2002 and 2003 random samples from the general population were taken to calculate the prevalence of previously undiagnosed leprosy ( PPUL ) . In the contact group of the COLEP study , the PPUL rate was 73/10 , 000 , compared to 15 . 1/10 , 000 in the samples taken from the general population . These cases were found by active door-to-door screening[13] . Disaccharide epitope ( 3 , 6-di-O-methyl-β-D-glucopyranosyl ( 1→4 ) 2 , 3-di-O-methylrhamnopyranoside ) of M . leprae specific native PGL-I glycolipid was synthesized and coupled to human serum albumin ( synthetic PGL-I; designated ND-O-HSA ) . This was generated with support from the NIH/NIAID Leprosy Contract N01-AI-25469 and obtained through the Biodefense and Emerging Infections Research Resources Repository ( http://www . beiresources . org/TBVTRMResearchMaterials/tabid/1431/Default . aspx ) . Antibodies ( IgM , IgG , IgA ) against M . leprae PGL-I were detected as described previously[5 , 14 , 15] . ND-O-HSA was coated onto high-affinity polystyrene Immulon 4HBX 96-well Nunc ELISA plates ( Thermo Scientific , Rochester , NY ) using 500 ng per well in 50 μl of 0 . 1M sodium carbonate/bicarbonate pH 9 . 6 ( i . e . coating buffer ) at 4°C overnight . Unbound antigen was removed by washing six times with PBS ( phosphate buffered saline ) with 0 , 05% Tween 20 ( wash buffer ) . The wells were blocked with PBS containing 1% BSA ( bovine serum albumin ) ( Roche Diagnostics , Germany ) for 1 hour at room temperature ( RT ) . Bloodspots were punched from filter papers . Three punches ( 2 mm each ) per individual were added to 100 μl PBST ( PBS/0 , 1% Tween20 ) and incubated at 4°C in 24 wells plates . After overnight incubation , 50 μl PBST/NRS ( PBST + 10% normal rabbit serum ) was added to each well and the plates were shaken gently for 1 hour at RT . The eluate was added to the ELISA plates ( 50 μl/ well ) and incubated for 2 hours at RT . After incubating with the primary antibody , the wells were washed six times with PBS with 0 . 05% Tween 20 ( wash buffer ) , followed by the addition of 50 μl of a 1:8 , 000 dilution of the secondary antibody anti-human ( Dako P0212 ) for two hours . Following washing the wells with the wash buffer six times , 50 μl of p-nitrophenylphosphate substrate ( Kirkegaard and Perry Labs , Gaithersburg , MD ) was added . Antibodies ( IgM , IgG , IgA ) against M . leprae PGL-I were detected as previously described[14] . Absorbance was determined at a wavelength of 450 nm . Samples with an optical density at 450 nm ( OD450 ) , after correction for background OD above 0 . 150 , were considered positive . This threshold was determined by a threefold multiplication of an average EC value . As quality control , anti-PGL-I IgM levels were determined for 10 Dutch leprosy patients by ELISA using serum as well as blood spots on filter paper: Although IgM levels were higher for 9 individuals in sera , all seropositive individuals were also positive using blood spots and OD450 values correlated well ( R2 = 0 , 80 ) . Multivariable logistic regression was used to calculate adjusted odds ratios for the level of ant-PGL-I Ab levels at intake , and corrected for age and sex . A p-value ≤ 0 . 05 was used as a cut-off for statistical significance . To investigate the association of changes in anti-PLG-I Ab from baseline to the time of development of leprosy , generalized linear mixed models were used . The dependent variable was the development of leprosy at a time point and the differences in anti-PLG-I Ab levels from baseline were included as independent variables . To adjust for the correlation between intra-individual measurements we included a random intercept for each subject . The difference between the anti-PGL-I Ab levels between contacts of MB or PB index patients was calculated using a t-test comparing averages . All analyses were performed in R version 3 . 2 . 0 ( R , Vienna , Austria; https://www . R-project . org ) .
From the 28 , 092 contacts of leprosy patients recruited within the COLEP study[11] , 239 contacts developed leprosy within the six years of follow-up . For 25 contacts who were diagnosed with leprosy during follow-up and 199 contacts who remained free of leprosy , good quality filter paper was available for at least three different time points during follow-up . Characteristics of the study populations are shown in Table 1 and Table 2 . Of the 25 contacts who developed leprosy , 10 contacts developed leprosy at 2 years after intake ( FU1 ) , 7 contacts at 4 years after intake ( FU2 ) and 8 contacts at 6 years after intake ( FU3 ) . Four contacts ( 16% ) developed MB leprosy and 21 ( 84% ) developed PB leprosy . This is the same proportion of MB versus PB as in the total group of new leprosy cases diagnosed within the COLEP study[12] 4 years after intake ( 24 MB contacts versus 126 PB contacts; 16% versus 84% ) . The group was evenly distributed for sex ( M/F = 1 . 17:1 ) and age categories . For 10 contacts the index patient had MB leprosy , whereas for 15 contacts this was PB leprosy . The anti-PGL-I Ab levels at intake were compared between the two groups of contacts ( Fig 1 ) . In the group of contacts who developed leprosy , the average anti-PGL-I Ab titer at intake was 0 . 11 , and varied between zero and 0 . 424 . 6 of these 25 ( 24% ) contacts who developed leprosy had a positive anti-PGL-I Ab level of >0 . 150 at intake . In the group who did not develop leprosy , the average anti-PGL-I Ab titer was 0 . 10 and varied between zero and 1 . 275 . 35 out of 199 ( 17 . 6% ) contacts who did not develop leprosy had a positive anti-PGL-I Ab level of >0 . 15 at intake . No significant association was observed for the anti-PGL-I Ab levels at baseline ( OR: 1 . 01 ( 0 . 78 , 1 . 31 ) , 95% CI p = 0 . 94 ) between the two groups . To further analyze the longitudinal pattern of PGL-I serology in contacts , the anti-PGL-I Ab levels are depicted at different follow-up times , comparing the titers of contacts developing leprosy ( Fig 2A ) to the titers of contacts without leprosy ( Fig 2B ) . The difference between anti-PGL-I Ab level at diagnosis was compared to the anti-PGL-I Ab level at intake . This difference was minus 0 . 047 , indicating that the level of anti-PGL-I Ab titer was lower at time of diagnosis compared to time of intake . Next we calculated the difference between anti-PGL-I Ab titer at various time points of follow-up to the anti-PGL-I Ab level at intake of the contacts who did not develop leprosy using a generalized linear mixed model analysis . Thus , for all contacts who did not develop leprosy , we compared the anti-PGL-I Ab level at FU1 to intake , the level of FU2 to intake and the level at FU3 to intake . If a contact developed leprosy at FU2 ( or FU3 ) , we also included the difference between anti-PGL-I Ab titer at FU1 ( and FU2 ) and intake into the group of contacts who did not develop leprosy . Differences in anti-PGL-I Ab levels had no significant association with the development of leprosy at either of the three follow-up time points ( OR: 0 . 62 ( 0 . 15 , 2 . 62 ) , p = 0 . 52 ) . Thus changes in anti-PLG-I Ab levels are not predictive of disease progression in contacts of new leprosy patients in Bangladesh . Since MB patients harbor a higher quantity of bacteria than PB , we separately considered the longitudinal pattern of the anti-PGL-I Ab levels in the four contacts who developed MB leprosy ( Fig 2C ) . The mean OD450 at the time of diagnosis for both MB/BT and PB patients was below threshold for positive ( < 0 . 15 ) . Also , no increase in anti-PGL-I Ab levels was observed at the moment of leprosy diagnosis; actually , anti-PGL-I Ab levels were often even lower at diagnosis time compared to intake . The findings indicate that not only for newly diagnosed PB , but also for MB patients , anti-PGL-I Ab levels do not represent a practical tool for prediction of leprosy .
In view of the dichotomy of the leprosy spectrum in terms of immunity against M . leprae , current research is focused on identification of predictive biomarker profiles associated with early stage leprosy , consisting of multiple cellular and humoral ( disease-specific ) biomarkers . Early diagnosis of leprosy and subsequent appropriate multidrug therapy ( MDT ) will not only decrease severe nerve damage and subsequent lifelong handicaps , but also significantly contribute to further decrease of M . leprae transmission . This study shows that measurement of anti-PGL-I Abs alone is not sufficient to predict the development of clinical leprosy amongst household contacts of newly diagnosed leprosy cases in ( highly ) endemic area such as Bangladesh . Because of the high number of PB patients in Bangladesh , using anti-PGL-I titers as a screening test to discriminate which contacts to treat , may lead us to miss a lot of potential new cases . Ethical clearance was obtained from the Ethical Review Committee of the Bangladesh Medical Research Council in Dhaka ( ref . no . BMRC/ERC/2001-2004/799 ) . All subjects were informed verbally in their own language ( Bengali ) about the study when they were invited to participate . Written consent was received from each adult , while a parent or guardian had to sign the consent form for children who participated in the study . | Leprosy is an infectious disease caused by the bacterium Mycobacterium leprae , which causes skin and nerve damage . Despite worldwide efforts to eliminate leprosy , the number of infected individuals who develop leprosy , is still substantial . Household contacts of new leprosy patients are especially at risk . Early diagnosis of leprosy is key in preventing lifelong handicaps as well as transmission . This requires field-friendly tests that identify individuals at risk of developing the disease before they develop clinical symptoms so that they can receive ( prophylactic ) treatment . Measuring antibody levels directed against the M . leprae-specific phenolic glycolipid I ( PGL-I ) provides an indication of the bacterial load . To identify whether anti-PGL-I Ab levels correlate with the development of leprosy in contacts of newly diagnosed leprosy cases , we analyzed these levels in 25 contacts who developed leprosy during 6 years of follow-up and 199 contacts who were not diagnosed with leprosy at 3 time points in 6 years . This study showed that anti-PGL-I Ab levels at intake did not significantly differ between contacts who developed leprosy during the study and those who remained free of disease . Therefore , anti-PGL-I Ab levels alone do not represent a practical tool for prediction of which household contacts will develop leprosy in an endemic area such as Bangladesh , with high levels of patients with paucibacillary leprosy . This stresses the need for a diagnostic test composed of a biomarker signature consisting of multiple biomarkers . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"mycobacterium",
"leprae",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"laboratory",
"equipment",
"engineering",
"and",
"technology",
"tropical",
"diseases",
"geographical",
"locations",
"biomarkers",
"bacterial",
"diseases",
"neglected",
"tropical",
"diseases",
"filter",
"paper",
"immunologic",
"techniques",
"bacteria",
"bangladesh",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"serology",
"immunoassays",
"actinobacteria",
"people",
"and",
"places",
"biochemistry",
"diagnostic",
"medicine",
"asia",
"equipment",
"leprosy",
"biology",
"and",
"life",
"sciences",
"organisms"
]
| 2017 | Longitudinal assessment of anti-PGL-I serology in contacts of leprosy patients in Bangladesh |
Timely resolution of inflammation is critical for the restoration of homeostasis in injured or infected tissue . Chronic inflammation is often characterized by a persistent increase in the concentrations of inflammatory cells and molecular mediators , whose distinct amount and timing characteristics offer an opportunity to identify effective therapeutic regulatory targets . Here , we used our recently developed computational model of local inflammation to identify potential targets for molecular interventions and to investigate the effects of individual and combined inhibition of such targets . This was accomplished via the development and application of computational strategies involving the simulation and analysis of thousands of inflammatory scenarios . We found that modulation of macrophage influx and efflux is an effective potential strategy to regulate the amount of inflammatory cells and molecular mediators in both normal and chronic inflammatory scenarios . We identified three molecular mediators − tumor necrosis factor-α ( TNF-α ) , transforming growth factor-β ( TGF-β ) , and the chemokine CXCL8 − as potential molecular targets whose individual or combined inhibition may robustly regulate both the amount and timing properties of the kinetic trajectories for neutrophils and macrophages in chronic inflammation . Modulation of macrophage flux , as well as of the abundance of TNF-α , TGF-β , and CXCL8 , may improve the resolution of chronic inflammation .
Prolonged inflammation is a recognized contributor to a multitude of pathological conditions , including cardiovascular , metabolic , and neurodegenerative diseases , as well as chronic injuries [1 , 2] . Timely resolution of inflammation is essential for tissue homeostasis . Inflammation resolution , previously believed to be a passive , self-regulatory process , is now known to be actively modulated by several different classes of endogenous molecular mediators , such as anti-inflammatory cytokines and growth factors [interleukin-10 ( IL-10 ) and transforming growth factor-β ( TGF-β ) ] , oxygenated lipid mediators ( lipoxins , resolvins , protectins , and maresins ) , and protease inhibitors [2–5] . Ongoing research efforts , including pharmacological animal model research and clinical trials , are focused on novel pro-resolution therapies for a variety of inflammatory conditions [6 , 7] . There is a clear , documented need for new approaches to develop resolution-centric therapeutic interventions [2] to supplement or replace the currently used anti-inflammatory therapies , which are only modestly effective [8–10] . The concentrations of molecular and cellular components of the inflammatory process are typically characterized by single-peak temporal trajectories reflecting a distinct period of activation followed by resolution [2 , 5 , 11 , 12] ( Fig 1 ) . The quantitative properties of these trajectories vary as a result of differences in inflammatory conditions and scenarios . Recently , four quantitative indices [namely , peak height ( Ψmax ) , activation time ( Tact ) , resolution interval ( Ri ) , and resolution plateau ( Rp ) ] ( Fig 1 ) were introduced as informative measures to analyze the quantitative patterns characterizing temporal inflammatory trajectories [1 , 13] . For a given molecular species or cell type , the Ψmax is defined as the maximum value of the corresponding temporal trajectory . Tact is the time after inflammation initiation required for the temporal trajectory to reach the Ψmax level . Ri is the time difference between Tact and the time it takes to reach 50% of Ψmax . Rp is the trajectory level at the end of the considered time period , expressed as a percentage of the Ψmax value; it therefore reflects residual inflammation . Based on the aspects of the temporal trajectories that they represent , the indices can be divided into amount indices ( Ψmax and Rp ) and timing indices ( Tact and Ri ) [14 , 15] . Additionally , the area under the curve ( AUC ) of a kinetic trajectory has been recently introduced as a metric to quantitatively assess cumulative tissue damage under various inflammatory scenarios [16] . Experimental studies have demonstrated the utility of the inflammation indices in distinguishing between the time courses for normal and pathological inflammation , as well as in establishing the quantitative efficacy of external interventions [1 , 3 , 17] . For example , in a mouse peritonitis model , it was shown that the macrophage Ψmax , Tact , Ri , and the neutrophil Tact were noticeably higher ( specifically , higher by 2-fold , 3-fold , 2-fold , and 12-fold , respectively ) in the chronic inflammatory scenario compared to the acute inflammatory scenario [17] . In the same study , several drugs ( specifically , ibuprofen , resolvin E1 , a prostaglandin D2 receptor 1 agonist , dexamethasone , rolipram , and azithromycin ) were evaluated for their inflammation resolution efficacy by quantifying their regulation of these inflammation indices in the chronic inflammation scenario . Another peritonitis mouse model study investigated the efficacy of drug-filled nanoparticles in inflammation resolution . The drug efficacy was established based on its ability to reduce the neutrophil Ψmax and Tact by ~1 . 5-fold [3] . Yet , systematic experimental characterization of the patterns and mutual relationships for the inflammation indices is challenging due to the immense diversity of possible inflammatory scenarios . This diversity hampers our understanding of the global control of the indices by specific molecular mechanisms , and our understanding of the modulation of the indices by therapeutic interventions . Computational modeling offers a possibility to complement experimental investigations and address this complex problem using an integrated approach . Computational modeling can be used to “screen” thousands of inflammatory scenarios in a systematic way and thereby guide the generation of focused , mechanistic , experimentally testable hypotheses [11 , 16 , 18–21] . Previous works have demonstrated the utility of mathematical models in the study of inflammation in specific disease scenarios and in the identification of crucial inflammatory mechanisms [11 , 12 , 16 , 20–32] . For example , computational models have been used to predict the efficacy of cytokine regulation therapies in chronic inflammatory diseases [24 , 27 , 33] . However , quantitative characterization of the inflammatory response of different cell types and molecular mediators in terms of indices , as well as a rigorous analysis of their mechanistic determinants , have not been carried out in previous modeling efforts . In the present study , we used our recently developed quantitative kinetic model of acute and chronic inflammation in wounds to generate hypotheses regarding mechanistic control of the inflammation indices [11] . Our model could capture the behavior of inflammatory cell type and molecular mediator kinetics in acute and chronic inflammation initiated by both infection and injury [17 , 34 , 35] . In our simulations , the chronic inflammatory scenarios were characterized by higher concentrations ( Ψmax ) and delayed resolution timing ( Ri ) for key inflammatory components in comparison with acute inflammatory scenarios . Using this model , we identified essential inflammation-driving mechanisms ( specifically , macrophage influx and efflux rates ) and informative indicators ( i . e . , IL-6 , TGF-β , and PDGF ) of chronic inflammation . Our findings regarding the mechanistic regulation of inflammation by macrophage fluxes are supported by experimental studies where wound macrophage levels ( regulated by macrophage fluxes ) controlled the timing of wound healing [36–39] and the quality of wound scarring [40] . Furthermore , our modeling predictions characterizing IL-6 as an informative indicator of pathological inflammation are consistent with a recent clinical wound study [41] . Here , we use this model to elucidate the functional relationships between specific molecular/cellular processes and inflammation indices during normal and pathological inflammation . The goal of our study was to identify specific mechanistic determinants that can be targeted to modulate the index values during abnormal ( delayed ) inflammation and drive them toward a desired outcome . We used two complementary analyses ( namely , sensitivity and correlation analysis ) and identified such targetable mechanisms for regulating the inflammatory indices in the model . Furthermore , we wanted to test the effectiveness of cytokine inhibition as an intervention strategy to regulate the inflammation indices . For this purpose , we extended the model to represent cytokine inhibition kinetics for three cytokines ( chosen based on our predictions regarding targetable mechanisms ) . We used this extended model to study the efficacy of individual/combined cytokine inhibition in the regulation of the inflammation indices for neutrophils and macrophages . Our modeling results indicate that , for the majority of the model output variables representing inflammatory cell types and molecular mediators , the amount indices Ψmax and Rp were robustly regulated by the macrophage influx and efflux rate , respectively . In contrast , for the timing indices ( i . e . , Tact and Ri ) , such a robust functional dependence on single model parameters was not detected in the sensitivity analysis . Yet , the timing indices for the inflammatory components were strongly correlated with the platelet degradation rate . Moreover , strong correlations between the timing indices and certain mechanistic processes existed , but only under specific inflammatory situations representing chronic inflammation . Our inflammatory mediator inhibition modeling suggested that during an abnormal ( delayed ) inflammatory response , TNF-α and TGF-β inhibition strongly shifted the macrophage Ψmax and Rp indices toward inflammation resolution , whereas CXCL8 inhibition regulated the neutrophil Ri and could nearly restore this index to its normal ( i . e . , acute-inflammation ) value . Notably , combined inhibition of TNF-α and CXCL8 resulted in improved restoration of normal neutrophil dynamics compared with the inhibition of these two targets acting independently . Comparisons with available experimental data provided validation for our TNF-α inhibitor modeling results .
To investigate the regulation of the inflammation indices via changes in the model parameters , we performed a local sensitivity analysis for the model in 10 , 000 inflammatory scenarios ( simulations ) , as described in the Materials and Methods Section . We regarded the regulation of an inflammation index ( computed for a given model output variable ) by a specific parameter as robust , if the sensitivity sij ( Eq 1 , defined in the Materials and Methods Section ) of the index with respect to this parameter was the highest ( or second highest , third highest , etc . ) across all parameters in the majority of the 10 , 000 simulations . In the figures and tables , for the sake of brevity , we identify each of the model’s 69 parameters using its assigned number in the model preceded by the prefix “P# . ” We use these designations when referring to the figures and tables in the text and , in addition , provide descriptive names for the parameters ( see Table 1 for a full list of the model parameters ) . The amount indices , Ψmax and Rp , for the majority of the model output variables were robustly regulated by the parameters representing the rate of macrophage influx ( P#7 ) into the inflamed area and the rate of their efflux ( P#10 ) , respectively . Table 2 shows the first , second , and third most influential ( based on the ranking of the corresponding sensitivities ) model parameters for the inflammation indices for six specific model output variables . These variables ( namely , TNF-α , IL-1β , IL-6 , IL-10 , total neutrophils , and total macrophages ) are the ones that typically demonstrate abnormal characteristics during pathological inflammation [17 , 34 , 36 , 42] . By definition , Ψmax and Rp characterize the inflammation intensity at its peak and during its resolution ( Fig 1 ) . Therefore , these results are consistent with , and complement , our earlier findings identifying macrophage influx ( P#7 ) and efflux ( P#10 ) rates as the main regulators of kinetic trajectories during the initial and final phases of inflammation , respectively [11] . In addition to the well-pronounced regulation by macrophage flux rates , our local sensitivity analysis identified the TNF-α degradation rate ( P#19 ) and the neutrophil influx rate ( P#3 ) as robust regulators of the Ψmax for the TNF-α and the neutrophil ( Ntot ) model output variables , respectively ( Table 2 ) . In contrast to the robust regulation of the amount indices , no model variables had their timing indices ( i . e . , Tact and Ri ) robustly regulated by a model parameter . Indeed , the largest sensitivity values for Tact and Ri corresponded to different parameters depending on the specific simulation ( i . e . , on the specific model parameter set ) . For all of the16 variables , the model parameter effecting the strongest Tact regulation in the largest fraction of the 10 , 000 inflammatory scenarios was the platelet degradation rate ( P#1 ) ( Table 2 ) . Yet , this largest fraction ( which we call the “robustness fraction” ) was too small ( ~18–46% ) to consider this regulation robust . For the Ri , the parameter effecting the strongest regulation of this index in the largest fraction of scenarios was macrophage efflux rate ( P#10 ) for 10 out of the 16 model output variables . However , the corresponding robustness fraction for it was even smaller ( ~15–35% ) . These results suggest that robust control of the timing indices may require a sophisticated strategy involving simultaneous modulation of multiple inflammatory mechanisms . To gain additional insights into the regulation of the inflammation indices , we calculated the correlation coefficients ( CCs ) and the associated p-values between the inflammation indices of each model output variable and each model parameter ( see Materials and Methods ) . The CC values ranged between −1 to +1 reflecting a negative or a positive index-parameter correlation , respectively . The signs of the CCs for the four indices are shown in S2 Fig . Absolute index–parameter correlations above 0 . 5 were considered strong [43 , 44] . Among the correlations identified as strong , only correlations with p ≤ 0 . 05 were considered statistically significant . For the amount indices , we detected a strong positive correlation between inflammation peak height ( i . e . , Ψmax ) and macrophage influx rate ( P#7 ) for 10 model output variables , including macrophages , IL-6 , and IL-10 ( Fig 2a and 2b ) . Furthermore , we found a strong positive correlation between the Ψmax for neutrophils ( Ntot ) and neutrophil influx rate ( P#3 ) ( Fig 2a and 2b ) , which was consistent with our sensitivity analysis results ( Table 2 ) . This result suggests that the inflammation indices for the neutrophil inflammatory response trajectory may depend on the degradation rates of CXCL8 ( P#31 ) and TGF-β ( P#13 ) , which are the two main neutrophil chemoattractants in our model [11] . For the other amount index , i . e . , Rp , only one parameter [namely , macrophage efflux rate ( P#10 ) ] showed a strong ( negative ) correlation with a large number ( namely , 13 ) of model output variables ( Fig 2c and 2d ) . These findings were consistent with our sensitivity analysis results regarding the Ψmax and Rp regulation by macrophage influx and efflux rates ( Table 2 ) . For the timing indices , the correlation analysis highlighted a strong negative correlation between TGF-β degradation rate ( P#13 ) and the Tact of four model variables , including neutrophils ( Ntot ) and IL-6 ( Fig 3a and 3b ) . Moreover , we detected a strong negative correlation between the platelet degradation rate ( P#1 ) and inflammation activation time ( i . e . , the Tact index ) for 9 model variables ( Fig 3a and 3b ) . However , this relationship simply demonstrates the connection between Tact and the strength and persistence of inflammation-initiating stimuli , reflected in our model by the presence of platelets in the wound . Additionally , the Ri for the majority of the model variables , including macrophages and IL-6 ( results not shown ) , exhibited a strong negative correlation with macrophage efflux rate ( P#10 ) , which could be expected based on our sensitivity analysis results ( Table 2 ) . Yet , our correlation analysis did not detect any other strong correlations for the timing indices . The limited number of detected strong correlations prompted us to hypothesize that a larger number of strong correlations between the timing indices and the model parameters could be detected in a smaller set of simulations reflecting specific conditions , such as chronic inflammation . To test this , we divided the 10 , 000 simulations into two subsets , “acute” and “chronic” ( see Materials and Methods ) . Then , for only the “chronic” subset , we performed the correlation analysis between each of the timing indices ( i . e . , Tact and Ri ) for the 16 model output variables and the 69 model parameters . As hypothesized , in this analysis many model parameters emerged as strongly correlated with Tact ( 18 parameters ) and Ri ( 9 parameters ) , for different model variables ( Fig 3c and 3d , respectively ) . In subsequent analyses , we specifically focused on TGF-β and CXCL8 degradation rates ( P#13 and P#31 , respectively ) , because they were strongly correlated with several key model outputs and those correlations were statistically significant . In summary , the correlation analysis provided us with candidate mechanisms for directly regulating both of the amount indices ( via neutrophil influx rate and the macrophage flux rates ) and the timing indices ( via the TGF-β and CXCL8 degradation rates ) of the model output variables . The goal of model parameter randomization in our 10 , 000 simulations was to account for possible biological variability in inflammation scenarios [11] . While wider variation ranges for the parameters may allow for a fuller representation of this variability , such random , simultaneous , uncorrelated variations may also introduce excess noise that could mask the biologically relevant patterns we want to detect . An informative analysis should therefore utilize parameter ranges sufficiently wide to represent variation and yet narrow enough to define the vicinity of our carefully chosen default parameter set , which had been derived directly from experimental data and represented the expected “typical” injury-triggered inflammation scenario [11] . To assess the impact of large parameter deviations , we performed the sensitivity and correlation analysis for an additional set of 40 , 000 simulations . In these simulations , the parameters were randomly and uniformly sampled from a 9-fold range ( i . e . , 3-fold down and 3-fold up ) around the default parameter values . In the sensitivity analysis , the introduction of these larger parameter deviations reduced by 5–30% ( results not shown ) the robustness of the Ψmax regulation by macrophage influx rate for the six different outputs shown in Table 2 . Similarly , the robustness of the Rp regulation by macrophage efflux rate was reduced by 1–14% ( results not shown ) in comparison with the results shown in Table 2 . However , in the correlation analysis , the major trends [i . e . , the strong correlation between the Ψmax and macrophage influx rate and the strong correlation between the Rp and macrophage efflux rate] were preserved , while other strong correlations ( Figs 2 and 3 ) were not detected in the case of large parameter deviations ( S3 Fig ) . Thus , it appears that some of the detected biological patterns are particularly strong in the fewfold vicinity of our default parameter set , which reflects their dependence on specific type of inflammatory scenario . The results of our sensitivity and correlation analyses ( Table 2 and Figs 2 and 3 ) suggested that specific inflammation indices of the model variables can be considerably impacted by the following five model parameters: macrophage influx and efflux rates and the TNF-α , CXCL8 , and TGF-β degradation rates . We therefore wanted to investigate whether modulation of these parameters during chronic inflammation could result in ( at least , partial ) restoration of the normal ( i . e . , acute-inflammation ) values of the respective inflammation indices . We addressed this question for all of the five parameters except macrophage influx rate , because we used its modulation to induce chronic inflammation in the model ( see caption for Fig 4 ) . Using the same protocol as implemented in our previous modeling study ( see Figure 5 in [11] ) , we simulated a chronic inflammatory scenario ( Fig 4a and 4b , red line ) . Here , we specifically focused on the kinetic trajectories for neutrophils and macrophages , because their kinetic behavior is well studied and is known to be disrupted during delayed ( or chronic ) inflammation [34 , 45] . To investigate the regulation of neutrophil and macrophage inflammatory trajectories , we repeated the chronic inflammation simulations upon modifying the values of the parameters mentioned above , as follows . In the chronic inflammation simulation for neutrophil regulation , we implemented two distinct parameter modification strategies . In one strategy , we increased the CXCL8 degradation rate , which caused the Ri for the neutrophils to decrease ( Fig 4a , solid black line and green values in the table ) compared to the chronic scenario with no modification ( Fig 4a , red line ) . In the other strategy , we increased the TGF-β degradation rate , which caused both Tact and Ψmax for the neutrophils to decrease ( Fig 4a , dashed black line , green values in the table ) compared to the chronic scenario with no modification ( Fig 4a , red line ) . Similarly , for macrophage regulation , we introduced two distinct parameter modification strategies into the chronic inflammation simulation . In the first strategy , we increased the TNF-α degradation rate , which caused the Ψmax for the macrophage variable to decrease ( Fig 4b , solid black line and green values in the table ) compared to the chronic scenario with no modification ( Fig 4b , red line ) . In the second strategy , we increased the macrophage efflux rate , which caused the Tact , Ri , and Rp for the macrophage variable to decrease ( Fig 4b , dashed black line and green values in the table ) compared to the chronic scenario with no modification ( Fig 4b , red line ) . The observed decrease in the inflammation index values suggested that the considered parameter modifications can partially restore acute inflammatory kinetics , i . e . , can bring the inflammation index values for a chronic inflammatory scenario closer to those for an acute inflammatory scenario . The analyses described above were supplemented with modulation of other parameters . This was done to illustrate that not all model parameters exhibiting strong correlations with certain model variables could effectively regulate their indices . For example , the IL-10 production rate parameter ( P#26 ) was strongly and negatively correlated with the macrophage Tact ( Fig 3c ) , and the parameter for TNF-α inhibition by TGF-β ( P#55 ) had a strong and negative correlation with the neutrophil ( Ntot ) Ri ( Fig 3d ) . However , increasing those two parameters did not result in a significant modulation of the respective neutrophil or macrophage indices ( Fig 4a and 4b , dotted lines ) . The model parameters that effected weak inflammatory index regulation or exhibited low-confidence correlations with the inflammatory indices were not selected for further analysis . Because the TGF-β , CXCL8 , and TNF-α degradation rates represent inherent biochemical properties of the respective molecular mediators , they cannot be easily changed in vivo . We therefore wanted to test whether a mechanistically distinct process , such as cytokine inhibition , could be used to obtain functionally similar outcomes . For this purpose , we extended our model [11] to include the kinetics of inhibitors for TGF-β , CXCL8 , and TNF-α and performed inhibition kinetics simulations for chronic inflammatory scenarios ( Figs 5 and 6 and S1 ) . We derived the values for the association ( kon ) and dissociation ( koff ) rate constants ( Eq 2 , defined in the Materials and Methods Section ) for each mediator inhibitor from literature data ( S1 Table ) . In our simulations , CXCL8 inhibition primarily regulated inflammation timing ( specifically , the Ri index for the neutrophil variable; Figs 5a and 6a , black lines ) , whereas TNF-α and TGF-β inhibition primarily regulated inflammation intensity ( specifically , the Ψmax and Rp indices for both neutrophil and macrophage variables ) ( Figs 5b and 5e and 6b and 6e; S1a and S1d Figs , black lines ) . For each inhibitor , we performed simulations for three different inhibitor concentrations . We used inhibitor concentrations of 10 nM , 100 nM , and 500 nM for both the TNF-α and CXCL8 inhibitors . For the TGF-β inhibitor , we used the concentrations of 1 nM , 20 nM , and 200 nM . We found that the inhibitors for different targets are most effective within concentration ranges that can be vastly different ( Figs 5 and S1 ) . Indeed , the CXCL8 and TNF-α inhibitors were most effective in restoring ( at least , partially ) their target indices when the inhibitors were added at concentrations ≥100 nM ( Fig 5a , 5b and 5e ) , while TGF-β inhibition was effective for inhibitor concentrations in the range ~1–10 nM ( S1 Fig ) . These results attest to the efficacy of the simple strategy involving the inhibitor addition at time 0 . To validate our modeling predictions regarding TNF-α inhibition , we compared our simulations with the experimental data generated for H1N1 virus-induced lung inflammation regulated by the TNF-α inhibitor etanercept [46] . These data characterized three scenarios: 1 ) control ( i . e . , no infection and , therefore , no inflammation ) , 2 ) inflammation with infection without added TNF-α inhibitor , and 3 ) inflammation with infection and with added 200 nM TNF-α inhibitor . From these data , we calculated the ratios of the total neutrophil and macrophage concentrations for the inflammation scenario without the TNF-α inhibitor to the corresponding concentrations for the inflammation scenario with added TNF-α inhibitor . We compared these ratios with the respective ratios calculated from our model simulation of injury-induced chronic inflammation . There was a reasonable agreement between the simulation-derived and experiment-derived ratios ( Table 3 ) . Note that , in our analysis , we did not attempt to model the nonzero neutrophil and macrophage concentrations detected for the experimental control scenarios , because in our model , which represents extravascular space , the concentrations of neutrophils and macrophages in the absence of inflammation are zero . In contrast , the control experiments measured the cell concentrations in entire lung lobes , which included cells present in the vasculature , resulting in nonzero neutrophil and macrophage concentrations even in the absence of inflammation . We chose this particular experimental study for the validation because TNF-α is a key pro-inflammatory cytokine that is secreted by inflammatory cells in most inflammatory scenarios , and whose expression and regulatory functions are largely independent of the specific inflammation-inducing stimuli ( e . g . , viral load , LPS [47] , and wounding [39] ) . Thus , despite the difference in the inflammation-inducing stimulus between the experimental study and our modeling study ( viral loading in lungs vs . injury , respectively ) , the reasonable agreement of the neutrophil and macrophage kinetics between the two studies suggests that our model adequately captured TNF-α inhibition . To test whether the timing of inhibitor administration can impact the index restoration outcomes , we performed simulations in which the inhibitors were added at three different time points ( i . e . , 24 , 48 , and 72 h ) during chronic inflammation . We chose 24 h as the earliest intervention time point because neutrophils are the first blood leukocytes to arrive at the inflammation site , and 24 h approximately corresponds to the neutrophil peak time for acute inflammatory response [35 , 48] . Macrophages peak at 48 h , which motivated our choice of the second intervention time . For all the simulations , the added CXCL8 and TNF-α inhibitor concentrations equaled 200 nM ( selected based on the observed effective range >100 nM , Fig 5a , 5b and 5e ) . In the case of CXCL8 , the inhibitor added at 24 h was more effective in reducing the Ri for the neutrophil variable than the inhibitor added at other time points ( Fig 6a , dotted black line ) . For TNF-α , the inhibitor added at 24 h provided a more complete restoration of the Ψmax for both neutrophils and macrophages to its acute-inflammation value than the TNF-α inhibitor added at later time points ( Fig 6b and 6e , dotted black lines ) . A nearly identical effect on the neutrophil Ri and macrophage Ψmax was observed when the CXCL8 and TNF-α inhibitors , respectively , were added at 48 h ( Fig 6a and 6e , dashed black lines nearly overlapped the dotted black lines ) . Mediator addition at 72 h showed the least degree of restoration in the neutrophil Ri and macrophage Ψmax among the three inhibitor addition times analyzed ( Fig 6a and 6e , black solid lines ) . However , this observed effect of mediator inhibition on the inflammation indices was not monotonic . Specifically , when the CXCL8 and TNF-α inhibitors were added at 36 h ( results not shown ) , the degree of restoration in the neutrophil Ri and macrophage Ψmax values , respectively , was less than that when the mediator inhibitors were added at the 48 h time point . Experimental data from cytokine inhibition studies [49 , 50] support the possibility of this type of non-monotonic behavior , which may be due to the complex nonlinear functional dependencies at work in the system . In summary , the CXCL8 and TNF-α inhibitors were characterized by similar optimal-efficiency concentration ranges and preferred administration timing regimens , but distinct preferentially regulated inflammation indices . Among the three time points considered , mediator inhibition was most effective when introduced during peak neutrophil response ( i . e . , at 24 h ) . For TGF-β , however , the inhibition outcomes were more complicated ( see S1 Fig and S1 Text ) . Because the neutrophil is the primary microbicidal cell type that produces powerful cytotoxic molecules that can potentially destroy the surrounding healthy tissue , ( at least partial ) restoration of the “normal” ( i . e . , acute-inflammation ) timing and intensity of the neutrophil surge is essential for the resolution of chronic inflammation [51] . Based on the preferential regulatory action of individual CXCL8 and TNF-α inhibitors on the total neutrophil Ri and Ψmax , respectively ( Fig 5a and 5b ) , we hypothesized that simultaneous inhibition of these two mediators in chronic inflammation might induce simultaneous restoration of the normal values for both of these indices . We tested this hypothesis by simulating the effect of adding both CXCL8 and TNF-α inhibitors at time 0 . Each of the two inhibitors was added at a concentration of 200 nM . As hypothesized , the combined inhibition resulted in simultaneous restoration of both the Ri and Ψmax of the total neutrophil trajectory in a chronic inflammation simulation ( Fig 5c , solid black line ) . Similarly to the action of the CXCL8 inhibitor ( Fig 6a ) , the combined inhibition of CXCL8 and TNF-α was most effective at 24 h ( Fig 6c and 6f , dotted black lines ) . These results suggest that “cocktails” of therapeutic agents with different targets may provide strategies of improved efficacy for simultaneous control of both timing and intensity of inflammation after wounding .
Timely resolution of inflammation following an injury , infection , or disease is essential for the maintenance of healthy tissue [2 , 52] . Despite ongoing research and development efforts , current anti-inflammatory therapies are only modestly effective and have significant negative side effects [9 , 53 , 54] . This work was motivated by the need to identify molecular mechanisms that could serve as targets for intervention strategies intended to modulate chronic inflammatory responses . Chronic inflammation may have distinct phenotypic manifestations depending on the inflammatory condition , e . g . , increased apoptotic neutrophil levels in diabetic ulcers [36] , or increased levels and prolonged presence of classically activated macrophages in ischemic wounds [45] , or prolonged oscillations in the levels of inflammatory cells and cytokines [55] . Here , we focused on chronic inflammation characterized by heightened levels and/or delayed resolution timing for kinetic trajectories describing accumulation and depletion of the inflammatory cell types and molecular mediators in comparison with acute inflammation [2] . In our simulations ( both acute and chronic ) , the kinetic trajectories of all the inflammatory components return to their baseline levels following a peak triggered by inflammation initiation . We used our computational model of injury-initiated local inflammatory response [11] to identify the mechanistic factors regulating the four inflammation indices , i . e . , Ψmax , Rp , Tact , and Ri ( Fig 1 ) , which were recently introduced as quantitative markers of the levels and timing of the inflammation time course [1 , 3 , 13 , 17] . Our sensitivity analysis elucidated the effects of parameter changes on the inflammation indices . For the majority of the model output variables , the amount indices ( i . e . , Ψmax and Rp ) were robustly regulated by macrophage influx ( P#7 ) and efflux ( P#10 ) rates , respectively , in the 10 , 000 performed simulations ( Table 2 ) . Because macrophages are major cytokine-producing cells at the site of inflammation [39] , identification of the macrophage flux rates as critical inflammation intensity control mechanisms is , perhaps , not surprising . However , the kinetics of each cytokine are additionally affected by a large number of factors , such as its individual production and degradation rates , feedback parameters , cell phenotype conversion rates , etc . In view of this diversity , identification of the macrophage flux rates as the only robust modulators for almost all model variables is an unexpected result . Interestingly , these results are supported by experimental studies involving direct macrophage manipulation [36 , 37 , 39 , 40] . For example , elevated macrophage abundance was shown to cause wound fibrosis [40] and delayed wound healing [38] . Moreover , reducing the levels of functional macrophages in wounds can severely delay the wound healing time in mouse models [36 , 37 , 39] . The macrophage flux rates , being strong regulators of the intensity inflammation indices , ( Table 2 and Fig 2 ) , are in fact the primary targets of certain recently emerged pro-resolution molecular mediators , such as lipoxins and resolvins [4] . By perturbing the recruitment mechanisms of neutrophils and macrophages in our model ( Fig 4 ) , we were able to reproduce the experimentally observed effects of exogenous delivery of resolvins and lipoxins in models of murine peritonitis ( Fig 4 , black lines and green values in the tables ) , e . g . , a reduction in the total neutrophil Ψmax and Tact ( see Table II in [1] ) , and a reduction in the total macrophage Ψmax ( see Figure 5 in [17] ) . While these experiments addressed only a handful of specific inflammatory scenarios , the robustness of the inflammation index regulation by the macrophage flux parameters , detected in our sensitivity analysis ( Table 2 ) , attests to the general nature of this regulation type . It should be noted , however , that the macrophage influx rate parameter ( P#7 ) in our model is a lumped parameter representing macrophage chemotaxis stimulated by any combination of four distinct macrophage chemoattractants ( specifically , TGF-β , PDGF , TNF-α , and MIP-1α ) ( see Table 1 in [11] ) . Likewise , the neutrophil influx rate parameter ( P#3 ) , which robustly regulates the total neutrophil Ψmax ( Table 2 ) , characterizes the chemotaxis of active neutrophils stimulated by any combination of TGF-β and CXCL8 . Thus , the sensitivity analysis alone [which identified P#7 and P#3 as influential parameters ( Table 2 ) ] was not sufficient for obtaining deeper insights into the specific mechanistic factors regulating the macrophage and neutrophil influx rates . This was accomplished in our parameter/output correlation analysis . The correlation analysis explored the effects of simultaneous , random parameter variations within specified limits . The analysis identified a number of functional associations between model parameters and inflammation indices ( Figs 2 and 3 ) . Consistently with the sensitivity analysis results , macrophage influx and efflux rates were strongly correlated with Ψmax and Rp of model outputs ( Fig 2 ) . However , in contrast to the sensitivity analysis results , where the timing indices ( i . e . , Tact and Ri ) did not display robust regulation by any of the kinetic model parameters across the 10 , 000 simulations , a strong negative correlation was observed between the platelet degradation rate and the Tact of the majority of output variables ( Fig 3a and 3b ) . This correlation may simply reflect that the presence of platelets in the wound area is the only inflammation-initiating stimulus in our model . While other mediator-releasing cells ( e . g . , endothelial cells , resident macrophages , mast cells , etc . [56] ) may contribute to the inflammatory response , this simplifying modeling assumption is in accord with the prominent roles played by platelets in initiating wound healing [57 , 58] . Of the many platelet-secreted molecular mediators facilitating inflammation initiation [57] , we only modeled the most potent neutrophil and macrophage chemoattractant , i . e . , TGF-β [59 , 60] . The results of this choice are evident from the strong correlations between the TGF-β degradation rate and key inflammatory components , such as total neutrophils , pro-inflammatory macrophages , and IL-6 ( Fig 3b ) . Overall , Ri was the least sensitive of all the inflammation indices and had a limited response to the parameter variations applied , which was consistent with its experimentally detected insensitivity [1] . Notably , for a subset of simulated scenarios reflecting chronic inflammation , certain parameters ( e . g . , TGF-β and CXCL8 degradation rates ) exhibited a strong correlation with the Tact and Ri values for many model output variables , including key outputs such as total neutrophils , TNF-α , and IL-6 ( Fig 3c and 3d ) . Thus , our correlation analysis identified relationships between specific inflammation indices and the specific molecular components regulating the macrophage ( TGF-β , Fig 3c ) and neutrophil ( TGF-β and CXCL8 , Fig 3c and 3d ) influx . We extended our model to describe inflammatory mediator inhibition kinetics in order to mechanistically represent the regulatory effects of inhibiting TNF-α , CXCL8 , and TGF-β ( Figs 5 and 6; S1 Fig , black lines ) . Mediator inhibition could provide a pharmacologically feasible strategy to mimic the functional effects of the modulation of the neutrophil and macrophage flux mechanisms , as well as of the modulation of TNF-α , CXCL8 , and TGF-β degradation rates , which we identified as robust regulators of specific inflammatory indices . In our simulations , individual and combined mediator inhibition demonstrated the potential for significantly improved restoration of acute-scenario kinetics for specific cell types in certain injury-initiated chronic inflammatory scenarios ( Figs 5 and 6 ) . Interestingly , the molecular mediators TGF-β and CXCL8 , which were identified as strong regulators of the timing indices Tact and Ri ( Fig 3 ) , successfully regulated the amount indices Ψmax and Rp of neutrophils and macrophages ( Figs 5 and S1 ) . This suggests that CXCL8 and TGF-β may be the primary contributors to the neutrophil and macrophage abundance in the wound , respectively . Yet , the detected neutrophil Ψmax regulation by TNF-α inhibition ( Fig 5 ) was surprising because , unlike CXCL8 , TNF-α is not a chemoattractant for neutrophils . TNF-α has been reported to be pro-apoptotic to neutrophils in specific concentration ranges [61] . It is conceivable that the non-linear feedback effects present in our model allowed us to capture this indirect effect of the neutrophil Ψmax regulation by TNF-α . The detected difference in regulation robustness between the timing and amount inflammation indices appears to be an intrinsic property of the inflammation regulation system . Interestingly , a regulation dichotomy between timing and amount properties of the output kinetics has been detected for other systems in molecular biology . For example , in bacterial signal transduction systems , the amount characteristics of the response curves are primarily defined by the numerical values of the systems’ kinetic parameters , whereas the response timing properties are determined largely by the systems’ architecture and are less sensitive to parameter variation [14] . Therefore , a lack of robust patterns characterizing the control of response timing via kinetic parameter modulation might be a common property of biological control systems . Traditional anti-inflammatory therapies , such as nonsteroidal anti-inflammatory drugs targeting the production of prostaglandins [10] , have focused on the inhibition of pro-inflammatory pathways . More recently , administration of pro-resolution molecular agents , such as lipoxins and resolvins , has shown promise in promoting inflammation resolution and is now a topic of active research [4 , 62] . Direct inhibition of inflammatory mediators is another therapeutic strategy that has seen moderate success for specific inflammatory conditions . For example , TNF-α inhibition and IL-1β inhibition have been successful as therapies for rheumatoid arthritis ( RA ) [8 , 9] and chronic inflammatory pathologies associated with cancer [63] , respectively . Previously published reports suggest that single mediator inhibition strategies are useful predominantly in chronic inflammation scenarios driven by specific mediators and/or when only one inflammation system component is affected , e . g . , in the neutrophil overload condition [36] . Our results are consistent with these findings and suggest that the observed therapeutic effects can be explained by preferential regulation of a given inflammation index by a specific mediator . Indeed , TGF-β and TNF-α inhibition regulate the total neutrophil and total macrophage Tact and Ψmax ( Figs 4a and 4b and 5b and 5e and 6b and 6e and S1a and S1d ) , while CXCL8 inhibition generally regulates the total neutrophil Ri ( Figs 5a and 6a ) . The generality of our approach suggests that TNF-α inhibition could potentially be used ( alone or in combination with other mediator inhibitors ) for a variety of inflammatory conditions characterized by elevated neutrophil and macrophage levels , such as traumatic injuries [64 , 65] . TGF-β regulation has largely been studied for its role in fibrosis during the later stages of wound healing [66 , 67] . Our results show that TGF-β inhibition might have a role in promoting the resolution of the crucial inflammatory stage of wound healing . In fact , recent clinical trials have begun evaluating the efficacy of TGF-β inhibition for various inflammatory pathologies , such as cancer [68] . CXCL8 inhibition has until now been speculated to have a role in inflammatory airway diseases , such as chronic obstructive pulmonary disease [69] and severe asthma [53] . Our results indicate a potential role for CXCL8 inhibition in timely neutrophil resolution during injury-initiated inflammation that warrants further experimental investigation . For pathological situations affecting several components of the inflammation process , the use of combination mediator therapy has been recently proposed and is being tested in clinical trials for rheumatoid arthritis [54] . Encouragingly , our simulations showed that a combined intervention targeting both TNF-α and CXCL8 in a chronic inflammatory scenario reduced both Ψmax and Ri for the total neutrophil concentration and thereby restored the near-“normal” ( in comparison with the acute inflammatory scenario ) total neutrophil kinetic trajectory ( Figs 5c and 6c ) . Overall , our modeling results suggest that individual and combined mediator inhibition approaches may be applicable during trauma-induced chronic inflammation , in addition to the inflammatory conditions for which they are currently indicated or being tested . Our model may be used to test the efficacy of different intervention strategies , involving combined administration of inhibitors for TNF-α , TGF-β , and CXCL8 , and to possibly determine the intervention timing . The use of robust control methodologies with our model might facilitate the design of improved parameter selection strategies to simulate chronic inflammation and to optimize drug administration regimens [70] . The main limitations of our approach arise from the computational nature of our study and from the assumptions made during the computational model development [11] . First , in our sensitivity analysis , we only examined the effects of parameter variations in the vicinity of the default parameter set . Larger ( e . g . , several‐fold ) parameter changes , which could represent a higher degree of dysregulation or severity of inflammation , were not examined . Nevertheless , local sensitivity analysis is an established research methodology that allowed us to generate model results consistent with experimental data ( Table 2 ) [11 , 19 , 71–74] . While this methodology is not designed to investigate simultaneous variations in several parameters , simultaneous parameter variation was a part of our correlation analysis strategy . Moreover , the effects of simultaneous parameter variations can be understood using global sensitivity analysis , as was recently done for a model of acute inflammation [16] . Second , our model contained a simplified representation of the interaction between inhibitors and their targets . Specifically , we did not include the individual degradation/removal kinetics for the inhibitors , which reflects the assumption that the inhibitors are present at the inflammation site at nearly constant total concentrations . While this assumption may be only an approximation to the in vivo situation , it allowed us to focus our inhibition analysis specifically on the inhibitor-target interactions ( rather than potentially complex inhibitor pharmacokinetics ) . Besides , an increase in the number of modeled components and their interactions would have caused an increase in the number of unknowns , thereby introducing additional uncertainty , which we wanted to avoid at this stage of the inhibitor modeling . Finally , our model did not explicitly represent the action of the recently emerged lipid mediators of inflammation resolution , such as lipoxins and resolvins [1 , 17] . Yet , our modeling approach allowed us to implicitly represent their effects by modifying their suggested target mechanisms ( see Figure 8 and Table II in [1] ) . With the recognition of inflammation as the key contributor to several pathologies , new approaches are needed to identify mechanistic regulators of pathological inflammation . A promising methodology is based on the analysis of the inflammation indices that quantitatively characterize the shapes of inflammation trajectories . Our study shows the applicability of systems biology approaches to identify mechanistic regulators of the inflammation indices . Moreover , computational models can provide a non-invasive and cost-effective framework for testing the efficacy of potential therapeutic strategy aimed to improve inflammation resolution . Computational analyses employing robust control strategies can guide the development of focused , hypotheses-driven experimental and clinical studies by reducing the ambiguity in the timing and strength of therapeutic interventions .
To simulate the inflammatory response , we used our recently developed quantitative model of acute and chronic local inflammation in a wound [11] . The model reflects initiation of inflammation by the platelets present at the site of injury; the platelets release TGF-β , whose gradient attracts inflammatory cells that interact and release soluble inflammatory mediators . The model’s variables describe the kinetics of 5 types of inflammatory cells ( namely , active and apoptotic neutrophils , pro- and anti-inflammatory macrophages , and platelets ) and 11 molecular mediators [namely , TNF-α , interleukin ( IL ) -1β , IL-6 , IL-12 , IL-10 , CXCL8 , TGF-β , platelet derived growth factor ( PDGF ) , macrophage inflammatory protein-1α and 2 ( MIP-1α , MIP-2 ) , and interferon gamma-induced protein 10 ( IP-10 ) ] . We chose to model these components because they are widely regarded as essential cell types and molecular mediators involved in an innate immune response to injury [48 , 75] . In our analysis , we did not include inflammatory molecular mediators produced by cells of adaptive immunity ( e . g . , IL-4 and IFN-γ ) that are reported to affect innate immune components , such as macrophages [40] . For modeling tractability , we reduced the wide spectrum of possible wound macrophage phenotypes [34 , 40] to just two: pro-inflammatory ( similar to the “classically activated , ” or M1 , phenotype induced in vitro by IFN-γ and bacterial LPS ) and anti-inflammatory ( similar to the “alternatively activated , ” or M2 , phenotype induced in vitro by IL-4 and IL-13 ) phenotypes . The essential mechanisms ( i . e . , chemotaxis and phenotype conversion of inflammatory cells , cellular apoptosis , and molecular mediator production/degradation and positive/negative feedback effects ) that govern the kinetics of the inflammatory response are represented via 69 model parameters ( Table 1 ) . The model describes extracellular signaling between different cell types and cytokines . Intracellular processes , such as transcription , translation , and export of proteins , are not represented mechanistically and are instead described implicitly via the model’s rate parameters . Our model reflects experimentally observed acute and chronic inflammatory response kinetics initiated by injury or infection ( see Figures 3 and 5 in [11] ) . While the kinetic trajectories of the model’s output variables in our simulations had qualitatively similar shapes during both acute and chronic inflammation , the chronic inflammatory scenarios were characterized by higher concentrations ( Ψmax ) and delayed resolution timing ( Ri ) for neutrophils , macrophages , and pro-inflammatory mediators ( e . g . , TNF-α , IL-1β , and IL-6 ) in comparison with acute inflammatory scenarios , which is consistent with experimental reports of chronic inflammation [17 , 34 , 36 , 45] . The model is a coupled system of 15 ordinary differential equations and one delay differential equation ( DDE ) . The DDE in the model is used to describe the chemotaxis of pro-inflammatory macrophages . Indeed , there exists a ~12 h delay between the arrival of the macrophage precursors ( i . e . , monocytes ) at the wound site and their differentiation into pro-inflammatory macrophages [12] , which is accounted for in the model by using the DDE . Each of our simulations reflected a 20-day period after inflammation initiation . We performed all computations in the software suite MATLAB R2012a ( MathWorks , Natick , MA ) and solved the model equations using the MATLAB solver DDE23 with default tolerance levels . The MATLAB files used to simulate the results reported in this article and a document providing details on how run the code are provided as S1 Code and S1 Text , respectively . For each model output variable representing a molecular species or cell type ( with two exceptions ) , we calculated the four inflammation indices defined in the Introduction ( Fig 1 ) . We did not calculate the indices for the model variables representing the platelet and TGF-β concentrations , because their kinetic trajectories do not have the same characteristic single-peak shape as the trajectories for other model variables . We additionally considered , and computed the inflammation indices for , two variables representing the total concentrations of neutrophils ( Ntot ) and macrophages ( Mtot ) . We computed these variables directly from the model output variables representing two distinct neutrophil and two distinct macrophage phenotypes . We thus analyzed 16 model variables in total . In the article text , we sometimes refer to the total neutrophil and total macrophage model variables simply as neutrophils and macrophages , respectively . First , we calculated temporal trajectories and the inflammation indices for the model’s default parameter set , which represents an acute inflammatory scenario [11] . Then , applying a previously described Latin hypercube sampling approach [71] , we generated 10 , 000 random 69-parameter sets , in which each individual parameter was sampled independently from an interval permitting up to twofold deviations ( up or down ) from the parameter’s default value . We used the MATLAB function LHSDESIGN to perform this sampling . The parameter sets were intended to represent the natural variability in the inflammatory scenarios occurring under different circumstances or in different individuals . In the existing literature , there is no consensus regarding the criteria for selecting the parameter randomization sample size to ensure sufficient coverage of the possible kinetic scenarios . It is recommended that this sample size be increased until no significant further changes in the main analysis results are detected [76] . Following this strategy , we arrived at the parameter randomization sample size equal to 10 , 000 parameter sets . For the considered model variables , we calculated the inflammation indices from the temporal trajectories generated for each of these 10 , 000 parameter sets . We refer to these trajectories as 10 , 000 simulations . We calculated logarithmic local sensitivities , sij , for each inflammation index of every model output variable with respect to every model parameter according to the standard definition ( see , e . g . , [11 , 73] ) : sij=∂logXi/∂logpj= ( dXi/Xi ) / ( dpj/pj ) , ( 1 ) where Xi is a given inflammation index ( Tact , Ψmax , Ri , or Rp ) for the model’s ith variable ( of the 16 output variables ) and pj is the model’s jth parameter ( of the model’s 69 parameters ) . To obtain numeric approximations of the derivatives in Eq 1 , each parameter was individually perturbed by ±1% of its value , and the derivative was approximated using the second-order central finite difference formula . We restricted our attention to local sensitivities because local sensitivity analysis is a powerful tool that has been successfully used to identify critical mechanisms and intervention points in a variety of biological systems [11 , 19 , 72–74] . We performed a local sensitivity analysis for the default parameter set , as well as for each of the 10 , 000 simulations with random parameter sets . For each considered parameter set , each model variable , and each inflammation index , we sorted the absolute sensitivity values in descending order to determine the top three most influential parameters for that variable’s index . Using the MATLAB function CORR , we calculated Spearman’s rank correlation coefficients ( CCs ) between each of the 69 model parameters and each of the four inflammation indices , for each of the 16 model variables . For these calculations , we used the model parameter values and the inflammation index values from the 10 , 000 simulations with randomized parameters described above . Moreover , based on the calculated Ri values for the neutrophil and macrophage variables , we introduced a criterion for dividing the 10 , 000 simulations into two subsets . One of the subsets contained simulations representing acute inflammation , and the other one contained simulations representing chronic inflammation . To create these subsets , we first calculated 20 , 000 ratios by dividing the Ri values for neutrophils and macrophages in each of the 10 , 000 simulations by the respective Ri values calculated using the default parameter set . Second , we used a cutoff value ( equal to 2 ) to identify the simulations ( out of 10 , 000 ) for which both the neutrophil and macrophage ratios were above the cutoff . The criterion for choosing the cutoff value was based on experimental studies , in which the Ri was increased by ~2-fold in abnormal inflammatory scenarios [12 , 17] . The simulations separated based on this criterion were regarded as reflecting chronic inflammation . We then re-calculated the CCs between the model parameters and model output timing indices ( i . e . , Tact and Ri ) using only the chronic inflammation simulations . This was done to identify any significant correlations that may exist for a specific chronic inflammation condition and were not detected when all 10 , 000 inflammation scenarios ( simulations ) were considered . A description of the algorithm for the separation of the simulation subsets is provided in S1 Text . We modeled inflammatory mediator inhibition by adding two new differential equations to the original model for each modeled inhibitor . The equations represented the volumetric concentration of the inhibited mediator and its respective mediator-inhibitor complexes . We used mass action kinetics to model the individual reactions between the inhibitors and their target mediators according to the following reaction scheme ( which reflects inflammatory mediator sequestration that prevents the mediator’s participation in normal signaling ) : I + C ←koff→ kon IC ( 2 ) where C denotes a mediator , I represents an inhibitor , and IC denotes the mediator–inhibitor complex . For any given inhibitor , kon and koff denote the values for the association and dissociation rate constants , respectively . Using our model extended in this way , we performed simulations to predict the effects of mediator inhibition at different inhibition concentrations and the addition of inhibitors at different time points after inflammation initiation . In these analyses , the inhibitors were added only in chronic inflammation simulations , and only one mediator was inhibited in each simulation , unless stated otherwise . A detailed description of the computational implementation of inflammatory mediator inhibition is provided in the Supplemental Material ( S1 Text ) . | A recent approach to quantitatively characterize the timing and intensity of the inflammatory response relies on the use of four quantities termed inflammation indices . The values of the inflammation indices may reflect the differences between normal and pathological inflammation , and may be used to gauge the effects of therapeutic interventions aimed to control inflammation . Yet , the specific inflammatory mechanisms that can be targeted to selectively control these indices remain unknown . Here , we developed and applied a computational strategy to identify potential target mechanisms to regulate such indices . We used our recently developed model of local inflammation to simulate thousands of inflammatory scenarios . We then subjected the corresponding inflammation index values to sensitivity and correlation analysis . We found that the inflammation indices may be significantly influenced by the macrophage influx and efflux rates , as well as by the degradation rates of three specific molecular mediators . These results suggested that the indices can be effectively regulated by individual or combined inhibition of those molecular mediators , which we confirmed by computational experiments . Taken together , our results highlight possible targets of therapeutic intervention that can be used to control both the timing and the intensity of the inflammatory response . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | Computational Identification of Mechanistic Factors That Determine the Timing and Intensity of the Inflammatory Response |
Unlike the core structural elements of a protein like regular secondary structure , template based modeling ( TBM ) has difficulty with loop regions due to their variability in sequence and structure as well as the sparse sampling from a limited number of homologous templates . We present a novel , knowledge-based method for loop sampling that leverages homologous torsion angle information to estimate a continuous joint backbone dihedral angle density at each loop position . The φ , ψ distributions are estimated via a Dirichlet process mixture of hidden Markov models ( DPM-HMM ) . Models are quickly generated based on samples from these distributions and were enriched using an end-to-end distance filter . The performance of the DPM-HMM method was evaluated against a diverse test set in a leave-one-out approach . Candidates as low as 0 . 45 Å RMSD and with a worst case of 3 . 66 Å were produced . For the canonical loops like the immunoglobulin complementarity-determining regions ( mean RMSD <2 . 0 Å ) , the DPM-HMM method performs as well or better than the best templates , demonstrating that our automated method recaptures these canonical loops without inclusion of any IgG specific terms or manual intervention . In cases with poor or few good templates ( mean RMSD >7 . 0 Å ) , this sampling method produces a population of loop structures to around 3 . 66 Å for loops up to 17 residues . In a direct test of sampling to the Loopy algorithm , our method demonstrates the ability to sample nearer native structures for both the canonical CDRH1 and non-canonical CDRH3 loops . Lastly , in the realistic test conditions of the CASP9 experiment , successful application of DPM-HMM for 90 loops from 45 TBM targets shows the general applicability of our sampling method in loop modeling problem . These results demonstrate that our DPM-HMM produces an advantage by consistently sampling near native loop structure . The software used in this analysis is available for download at http://www . stat . tamu . edu/~dahl/software/cortorgles/ .
Starting from a known structural homolog , template based modeling ( TBM ) of protein structure provides the most accurate predictions of protein sequences with unknown structure [1] , [2] . However , even with close structural homologs , structurally variable regions ( SVRs ) , commonly referred to as loops , are the worst predicted segments [3] , [4] , [5] . Because loop regions join elements of regular secondary structures and often play an important role in active site composition , ligand binding , and protein-protein interactions , accurate sampling is integral to a useful TBM prediction of protein structure . Structurally , loops often lie on the solvent-exposed surface of proteins , allowing them more conformational flexibility and susceptibility to insertions and deletions . This variability makes loop regions notoriously difficult to align at both the sequence and structural level , which often results in large stretches of gapped positions . As an added level of complexity , the conformational space is usually poorly populated due to the low structural homologs . This variability and sparsity of data pose much of the challenge in modeling with current approaches , and these problems increase with loop length . Typically , loop-modeling methods have adopted one of two general strategies , de novo and knowledge-based loop modeling methods . In de novo loop modeling [4] , 6 , physico-chemical based principles are used to compute the lowest energy conformations for a loop [7] , [8] . In successful applications to short loop modeling , de novo methods include molecular dynamics simulations [9] , simulated annealing [4] , buildup from discretized φ , ψ pairs [10] , [11] , [12] , and ‘random tweak’ [8] , [13] . However , these methods are limited because they require significant computational resources to sample near-native conformations . Alternatively , the loops in some proteins can be classified into structural families or canonical types , as in the antibody hypervariable regions ( complementarity determining regions or CDRs ) [14] , [15] , [16] , [17] , [18] . Such knowledge-based schemes utilize known structures or fragments of structures to efficiently sample loop conformations , [19] , [20] , [21] , [22] , [23] , but are limited to sampling within the knowledge base . Using large databases of supersecondary structures [24] , loops are successively aligned with templates based on parameters such as the stem region geometry , length , and sequence similarity [25] , [26] , [27] . While the strategies in various methods differ in many respects , the fundamental idea is to efficiently sample the available conformational space for loops of the particular length , and then score the samples using various energy functions [7] . The modeling of longer loops up to 13 residues in length has been achieved using exhaustive sampling of φ , ψ space with clustering and energy minimization [28] . In addition , there are approaches that combine the use of loop databases and physical-based algorithms [29] , [30] , [31] as well as methods sampling loop libraries that focus on loop closure [32] , [33] , [34] , [35] . For all methods leveraging information from known structures , sampling is limited to the discrete conformational space represented in the structural library . While providing efficient sampling , this approach poses difficulties in completely representing the structural variability of a loop region . To address these obstacles in sampling of loop models , a novel statistical method has been developed that implements a Dirichlet process mixture of hidden Markov models ( DPM-HMM ) [36] , [37] for continuous density estimation of φ , ψ residue torsion angles in the loop region . This statistical modeling not only retains the advantages of utilizing information from homologous proteins but also provides the continuous sampling of conformational space allowed by physico-chemical methods . From the sparse sampling at each loop position , the DPM-HMM method computes a joint φ , ψ density using statistical inferences from neighboring residues to make probable estimations of a continuous probability . The approach uses the φ , ψ data from homologous loops to model the joint φ , ψ distributions at each loop alignment position . The results are continuous density estimations of each residue's Ramachandran space , which allows sampling from a wider range of φ , ψ values than the discrete possibilities using a loop library [7] , [10] , [12] , [28] , yet the distribution is informed by the homologous loops . A related statistical method with a different formulation called DBN-torus has been concurrently developed by Boomsma et . al . [34] for the modeling of fragments in template-free protein structure prediction . Unlike the specificity of this method for fragment generation , our approach is tailored for TBM and produces nearer native loop samples even when good templates are not available . Moreover , the DPM-HMM method allows fast , knowledge-based sampling of backbone torsion angles focused within probable regions of φ , ψ space [36] , [37] . In this study , the ability of the DPM-HMM approach to sample near-native candidates is demonstrated in the modeling of loops from the following three groups: ( 1 ) canonical and non-canonical hypervariable loops within the heavy chain complementarity-determining regions ( CDRHs ) of immunoglobulins , ( 2 ) the conserved EF loop from the globin fold , and ( 3 ) the loops of CASP9 targets . Examples of these are shown in Figure 1 . Sampling near native loop conformations was tested in leave one out ( LOO ) approach and general applicability of the method is demonstrated with the results for loop modeling of TBM targets from CASP9 experiment . Also , the performance of DPM-HMM method was compared with LoopyMod by using CDRH1 and CDRH3 data sets .
At the heart of our approach is the DPM-HMM density estimation of the backbone φ , ψ angles [36] , [37] , and the method's ability to correctly model the torsion angle space helps to explain our success or failure in modeling particular targets in our LOO tests . Figure 2 shows four examples of Ramachandran plots [38] taken from predictions of targets from the CDRH2 loops . Our method uses the normalized φ , ψ data from template loops [39] as a prior or basis for its density estimations ( see Materials and Methods ) of the probability distributions . As shown by the scattered points in Figure 2 , the backbone φ , ψ angles from the templates provides the raw data that combine with the prior to produce the estimated distributions shown as contour lines in the plots . In a number of cases our statistical estimation of density performed well , as evidenced by the presence of the native , target φ , ψ pair ( shown as a red point in Figure 2 ) being predicted within the highest probability regions . Panels ( a ) and ( c ) in Figure 2 show the native φ , ψ pair within the highest region of estimated density . For Figure 2a , the observed result is expected as these positions hold the anchoring residues for CDRH2 , which are consistently in the β-sheet region of the Ramachandran plot . As can be seen in Figure 2c , certain positions heavily favor the left-handed helical region . This method's success in loop prediction corresponds well with density plots that contain a majority of residues with highest density around the native φ , ψ pair . By contrast , panels ( b ) and ( d ) in Figure 2 show instances in which the native φ , ψ resides in a lower probability region of our density estimates . Figure 2b shows a residue sampling the second highest region of a left-handed helix . In Figure 2d , unlike the majority residues that populate the β-sheet region , the glycine residue at the anchor position after the CDRH2 loop exhibits φ , ψ values in the commonly disallowed lower right quadrant of the Ramachandran plot . Because our density estimation model does not exclude but places a lower probability distribution in this region , these positions in the loop are more of a challenge to our sampling and helps to explain the prediction limit of 3 . 66 Å for poor/sparse input data described below . Instead of a measure based just on Cα atoms , root mean squared deviations ( RMSDs ) were calculated using all of the main-chain heavy atoms between the candidates and the native target to analyze the data and measure the accuracy of the prediction ( see Materials and Methods ) . In addition , we performed global superposition of the loop fragments on the protein structure to calculate the RMSD between the models and the reference structure , which is a departure from the more commonly used local superposition that is independent of the overall protein structure . A local superposition of the loop candidates certainly produces lower average RMSD values to the reference structure as loop fragments often fit well locally to the reference , but the loop might not be the best candidate due to lever arm effects in the take off and landing residues . In contrast , while a global superposition will always yield a higher value for RMSD than local superposition as pointed out by Choi et . al . [40] assessing loop accuracy in the global context of the protein structure properly reproduces modeling conditions , where the native loop or overall structure is not known and loops are placed onto a backbone template . Figure 3 demonstrates the accuracy of global over local alignment in providing a more realistic measure to evaluate loop modeling . In both parts of Figure 3 , the same 97 candidates of a loop for CASP9 target T0617 , whose average Cα distance for C-terminal anchor is below 1 . 0 Å , are either locally ( Figure 3a ) or globally ( Figure 3b ) superposed to the native target crystal structure depicted by a thicker red backbone trace . Comparing Figures 3a with 3b , local superposition of the 97 candidates produces a much smaller spread over a global superposition . The average RMSD proves this observation: 1 . 86 Å for local superposition opposed to 3 . 17 Å for global superposition . However , the most significant difference occurs at the take-off and landing positions . In the local superposition , the variation around the ends is larger , whereas in the global superposition , it is quite small . Comparing the closest candidates by both methods demonstrates the importance of using global superposition . The blue line is the best candidate by local superposition , which has a local RMSD of 0 . 58 Å yet a global RMSD of 1 . 10 Å . So , while this candidate looks to be the best match in Figure 3a , this loop would not be the best fit on the protein structure as shown in Figure 3b . It can be seen that even though the first N-terminal residue ( anchoring residue ) coordinates are shared between the target and all candidates , the overall orientations of the loops are very diverse . The green backbone is the best overall loop candidate found by global superposition at a RMSD of 0 . 77 Å ( Figure 3b ) , which is closer to the red native backbone than the top loop selected by local superposition . This candidate would have been missed in a local superposition with a RMSD of 0 . 65 Å . By not considering the fit of the loop onto the structure , local superposition accuracy is misleading and impractical in TBM as loops need to be evaluated in the context of a complete structure . Therefore , even though the RMSD values are higher for global superpositioning , the comparison stays truer to real prediction situations where the loop is being matched onto the body of model structure . The DPM-HMM method is able to produce consistent results across the various types of loop targets . In our LOO tests modeling the 465 loop data set ( Table 1 ) , the low mean global RMSDs for the best candidates shown in Table 1 demonstrate that our method performs well at sampling near-native loop candidates . To provide more details about the DPM-HMM's performance , sampling accuracy was measured by comparing the global RMSD of the best sampled candidate to that of the best template from the discrete set of template loop structures ( Figure 4 ) . The best template is the one with lowest global RMSD to the target native loop segment . In Figure 4 , the points below the diagonal line indicate loops our method modeled better than the best available template ( best candidate's RMSD is lower than that of the best template ) . The DPM-HMM method performs consistently well for common targets with templates averaging between 3 to 4 Å RMSD and even for the difficult targets with a mean template RMSD above 7 Å . Of all the 465 targets predicted by the DPM-HMM method , the best candidate global RMSDs are in the range from 0 . 45 Å to a top value of 3 . 66 Å , regardless of the loop length , number of templates , and the quality of the templates . So , we can reliably say that our method samples loop conformations at least within 3 . 66 Å to the native . The inset in Figure 4 shows the percentage of better or worse candidates compared to the best template binned by RMSD . In the very close canonical RMSD range of 0–1 Å , 38% candidates were sampled better than the best templates . Moreover , in this regime very close to the native structure where there is a higher probability to produce incorrect structures over the right ones , the DPM-HMM method sampled the remaining 62% in this canonical class not far from the best template . The worst case is with maximum deviation of 0 . 6 Å and mean deviation of 0 . 2 Å from the best template . As shown in the inset to Figure 4 , our sampling percentages from the DPM-HMM method only improve as the difficulty of loop modeling increases . In the RMSD range of 1–2 Å , around 75% of the best candidates improved on the best templates , and of the 25% that did not , the average increase in RMSD was 0 . 3 Å with a worst case of 1 . 3 Å deviation from the best template . In the next bin between 2–3 Å RMSD , 93% or almost all cases produced better candidates . In this range , the 7% of the cases that produced worse candidates averaged 0 . 6 Å RMSD with a maximum at 1 . 4 Å . For the cases with templates above 3 . 0 Å RMSD , which combines some common and all the difficult loop targets , our DPM-HMM method consistently constructs candidates that were better than the closest templates . Overall , about 76% of the loop conformations are sampled more accurately than the best templates available . This consistency of the DPM-HMM method in building improved loop models over these sets of varying difficulty demonstrates its utility and promise . We wanted to investigate how much influence the input data set had on our ability to build near native models . Figure 5 shows the correlation between the input templates' average RMSD and the best predictions for each of the 465 targets in our LOO tests . As a measure of the diversity of the templates , the average RMSD is calculated as the mean value between all the templates used in the DPM-HMM density estimation with each other . A larger average RMSD indicates greater diversity in the input template data . As expected , near native input data produces better model structures . As shown previously , the DPM-HMM has a limit of 3 . 66 Å even with very poor input data with average RMSD values past 7 . 0 Å . Furthermore , the targets were classified into 3 groups according to the number of templates used to produce the DPM-HMM models: ( 1 ) those relying upon less than 10 templates , ( 2 ) those with between 11–30 templates , and ( 3 ) those with greater than 30 templates . For the loops molded with fewer than 10 templates , their best RMSDs are mostly above 2 Å and do not demonstrate a strong dependence on the quality of input data . This suggests that the influence of the prior distribution determines the upper limit of our approach's abilities to sample the native structure . The targets that used between 11–30 templates display the expected correlation of improved candidate production from nearer native sets of templates . For this amount of input data , the DPM-HMM approach increases the probability of sampling near the target structure , which results in RMSDs of most of the best candidates below 2 Å . In our data set , there were only three loop examples that possessed more than 30 templates for input data: CDRH1 12 residue loop with 111 targets , CDRH2 8 residue loop with 87 targets , and EF 13 residue loop with 66 targets . Their average RMSD values are similar and cluster around 2 . 5 Å ( see black filled circles in Figure 5 ) . The large clustering is due to the numerous LOO tests that could be performed in this group . The best RMSD values range from 0 . 5 to 1 . 8 Å with a few exceptions discussed below , and 260 target tests in this class are modeled as below 1 . 8 Å RMSD . Although a large number of templates gives a better chance to model the long loops close to the native structure , the results from this class suggests that there is a saturation limit to the amount of information provided by the input data . The DPM-HMM approach fails to produce a model that is better than the average RMSD of the templates in two special cases ( data points above the unity line in Figure 5 ) . One particular case is a 15 residue EF loop that was modeled using only 2 templates . For this target , the average RMSD of templates is 1 . 46 Å but the RMSD of the best candidate is 2 . 77 Å , which is shown as the grey filled circle above the unity line in Figure 5 . This high RMSD arises primarily from using idealized bond lengths and bond angles to build loop structures from φ , ψ angles that are unable to reproduce that native loops conformation due to irregularities in bond angles , which has been previously discussed in detail [41] . The other loop that was poorly modeled belongs to the CDRH1 segment from the humanized anti-gamma-interferon antibody ( 1b2w [42] ) in the class of greater than 30 templates . The best sampled model has a global RMSD of 2 . 53 Å to the native loop structure ( black dot above the diagonal in Figure 5 ) . This loop possesses a 310 helical conformation in the middle of the CDRH1 , which places it as a distinct outlier in the dataset with over 100 canonical templates . The relationship between the loop length and the sampling efficiency was also investigated . In general , loop-modeling methods are more effective at predicting the shorter loops , where the accuracy decreases as the loop length increases . Figure 6 shows sample of various sizes of loops ranging from 7 to 17 amino acid residues . A linear correlation exists between the loop length and best-sampled loop conformation ( Figure 6a ) . In loop modeling , loops with 11–13 amino acid residues are considered long and prediction accuracies of about 1 . 0–1 . 5 Å for these long loops are considered to be a success [40] . In this study , sampling efficiency for shorter loops ( 7–10 amino acid residues ) was found to be below 0 . 5 Å . For the loops with 11–13 residues , the best candidates' global RMSDs are below 1 . 2 Å . For longer loops with 14–17 amino acids in length , the global RMSD is within the range of 1 . 8–3 . 0 Å , which improves upon the sampling reported by other methods [40] , [43] . The upper bound of sampling efficiency achieved here is about 3 . 66 Å , which encompasses the largest global RMSD for one of the predicted candidates belonging to the longest ( 17 residue long ) of CDRH3 loop category . The best candidates' RMSDs are also plotted against the number of templates in Figure 6b . As expected , the higher number of templates improves upon the sampling . From Figure 6b , the DPM-HMM method requires at least 30 templates in a data set to consistently make a prediction below 1 Å . With less than 30 templates , the dependency is more about how close the input data is to the target loop structure , where some instances are successful and others approach the 3 . 66 Å limit of our method . To further investigate the DPM-HMM method , sampling was analyzed for 90 properly identified loops modeled in 45 TBM targets during our group's CASP9 campaign . Figure 7 shows RMSD of the best candidate as a function of loop length . This dataset represents realistic modeling under real-world conditions where very few templates are available to model the torsion angle space . Also , available templates were of various sizes in loop length for each target loop . Our results demonstrate that sampling efficiency is very good for the loops of smaller lengths ( 3–7 residues ) with best-sampled candidates global RMSD of 0 . 25 Å to the native reference structure . Average global RMSD for best-sampled candidates in this group is about 0 . 89 Å . For medium sized loops ( 8–13 residues ) , best candidate RMSD is 0 . 99 Å and the mean over the group is 1 . 9 Å . For longer loops with more than 16 residues , sampling efficiency escapes the DPM-HMM limit of 3 . 66 Å . These longer loops pose a problem to the DPM-HMM to accurately model the data over so many residues . As can be seen from Figure 7 , the limit is stretched at 20 residues , where the best candidates are greater than 5 Å . Overall , the results demonstrate the general applicability of our method in realistic TBM situation where limited number of templates with variable loop lengths was used for modeling . Figure 8 compares the sampling efficiencies of DPM-HMM and LoopyMod for 2 sets of loops: the canonical CDRH1 and the non-canonical CDRH3 . First , canonical conformations from CDRH1 dataset containing 111 target loops were sampled with LoopyMod and results compared to the DPM-HMM method in the first 2 columns of Figure 8 . The global RMSD of the best candidate by LoopyMod is 1 . 02 Å , which is higher than the 0 . 61 Å by DPM-HMM method . Variance within RMSDs of the best candidates is also lower in DPM-HMM method than in LoopyMod . The DPM-HMM method produces all of its best candidates below 2 . 5 Å global RMSD , whereas LoopyMod has some cases upwards of 4 Å . In this canonical class , the DPM-HMM demonstrates that it performs well . Secondly , we tested the sampling efficiency of our method against LoopyMod for the non-canonical class of loops from CDRH3 . The third and the fourth columns in Figure 8 show comparison of sampling efficiency for CDRH3 by both DPM-HMM and LoopyMod , respectively . As expected , the distribution from both methods is wider as compared to the RMSD distribution for canonical class of loops ( CDRH1 ) . The median global RMSD of the best candidates is lower from the DPM-HMM method . The best models have RMSDs of 0 . 77 Å and 1 . 05 Å by DPM-HMM and LoopyMod , respectively . The tighter distribution of the DPM-HMM for both the canonical and non-canonical class of loops indicates this method's ability to take advantage of the knowledge base in producing near native candidates . ( See Figure S1 in Text S1 for scatter plot of individual data points )
The DPM-HMM method's use of a knowledge base implies that the approach is dependent on quality of the input data . Because longer loops are sampled less accurately and less consistently , loop length needs to be included in the discussion . As Figures 4 and 8 show , the DPM-HMM method performs well with canonical loops , so the discussion will focus on the longer more difficult to predict loops . Loops with lengths of 15 , 16 and 17 are modeled only with 2 to 8 templates . The longest 17 residue loops from CDRH3 was modeled with 8 templates and is considered to be a very difficult loop to model because of the length and the conformational variability ( templates' average global RMSD is 7 . 43±0 . 24 Å ) . The best models for this group show average global RMSD of 3 . 13±0 . 32 Å , which improves upon the closest templates as well as models predicted by other methods for the loops of similar length [40] , [44] . Recently , Choi et . al . [40] reported best models for 15–17 residue loops in the range of 3 . 48–4 . 75 Å . With best candidate global RMSDs between 1 . 12 to 1 . 81 Å ( Figure 4 ) , moderate length loops of 10–13 residues from CDRH3 were sampled closer to the target structure than that of the best templates . Another example of a longer loop used in this study is the 15 residue EF loop from the globin fold . There are only 3 EF loop targets with 15 residues; therefore one target is modeled using only the remaining two loops as input template data ( average RMSD 3 . 11±1 . 45 Å ) . The best predicted candidates were found to deviate from the target structure on average by 2 . 43±0 . 39 Å . While many of the φ , ψ density estimations properly model the backbone torsion angles in high density areas ( see Figures 2a and 2b ) , it requires only a few residues with angles reside in lower density regions of our density estimation ( see Figures 2b and 2d ) to make sampling a close model more difficult . It's also worth mentioning that two of the three globin EF loop 15mers are from crystal structures of the similar proteins with identical sequence in loop regions , however loop conformations varies from these 15mers with global RMSD of 2 . 26 Å . This reflects an extreme case of template based loop modeling with a limited number of templates , and the DPM-HMM method still achieves reasonable sampling efficiency in such difficult cases . The templates' average global RMSD provides an independent measure about the variability within the input knowledge base that can be used in real-world conditions to predict model's sampling performance . Lower values of RMSDs result from similar loop conformations to the other templates and higher values are attributed to the large deviation of loop conformations in the template set . Hence , the wider the range of template average RMSDs , the more diverse the template set is . Yet , even with a large variability in conformational space , our DPM-HMM can sample a diverse φ , ψ distribution , and still produce models better than the best template . Even if templates' average RMSDs are larger than 7 . 0 Å , the best candidates are at a maximum 3 . 66 Å RMSD . One example of improvement is 17 residue CDRH3 loop from monoclonal antibody hGR-2 F6 ( 1dqd [44] ) . The 8 input templates possess an average template RMSD of 7 . 05 Å and the best template has the RMSD of 8 . 30 Å to the target structure ( data point not shown in Figure 4 ) . For this difficult case , the DPM-HMM method produced a best candidate with a 3 . 11 Å RMSD to the native loop . This result demonstrates the ability of the DMP-HMM method to produce consistently good models for even difficult loop modeling examples . A major reason for this ability is that the continuous density estimations do not outright exclude areas of Ramachandran space , but rather bias the more probable regions as informed by the input data from the templates . Unfortunately , the DPM-HMM method has a residue limit of about 20 amino acids as shown by our CASP9 results in Figure 7 , where the method begins to under-sample the density estimations due to computational constraints . Overall , our approach directly addresses the familiar problem of insufficient templates as well as those all too common instances where the native loop uniquely deviates from the prevalent conformation of the templates at certain positions . For these reasons , the DPM-HMM method proves to be a reliable tool for loop modeling , since even with a small number of templates and low structural similarity , the DPM-HMM approach can quickly and thoroughly sample backbone φ , ψ space to identify loop structures near to native structure . In this study , we applied a novel loop modeling method , the Dirichlet process mixture of hidden Markov models or DPM-HMM [36] , [37] for φ , ψ density estimation in loop regions . 465 target loops classified in 17 groups depending on the loop identity and length were modeled . These targets were representatives of the various challenges in loop modeling from the easier canonical loops to non-canonical loops with many residues and insufficient sampling . By estimating a continuous distribution across conformational space , the DPM-HMM method combines the advantages of continuous sampling from physical methods and propensities from knowledge-based methods without compromising modeling speed or being limited to specific conformations found in fragment-based libraries . The best global RMSD of a candidate is as low as 0 . 45 Å for one of the shortest loops ( 7 amino acid residues ) from CDRH2 . For these canonical loops with templates below 1 . 0 Å , the DPM-HMM produces improved models in about 38% of target loops . Also , It is also very encouraging that we can always improve the best templates when best template RMSDs are higher than 3 . 0 Å . For the most difficult case of a long 17 residue loop with sparse input data , the DPM-HMM approach produces models within 3 . 66 Å , which is the limit independent of loop length and quality of input data ( Figures 4 and 5 ) . Our results demonstrate that the DPM-HMM method provides consistent and reliable model sampling across the spectrum of loop modeling up to 20 residues . The modeling accuracy was found to depend on three factors . The first , and most important , is loop length . It is well known that a loop becomes more difficult to model as length becomes longer ( Figures 6a and 7 ) , since more residues exponentially increase the potential conformational space . This quickly reduces the effective sampling that can be done . However , a loop with 17 amino acids was successfully modeled to 2 . 82 Å RMSD of the native with only 8 templates in the input data set . The low number of templates for input data points out the second factor: the number of templates available ( Figure 5 and 6b ) . The sampling efficiency shows negative correlation with the number of templates when less than 30 templates are used ( Figure 6b ) . The sampling efficiency becomes saturated when more than 30 templates are available . The last factor is the quality of the templates , which provides the input data for our density estimations . If near native templates are available , modeled loops are most likely close to the target structure . Even in cases where no good templates are available , the DPM-HMM method can produce improved loop models . Since all of the allowable Ramanchandran space possesses some probability in the density estimation , this approach can sample into underrepresented areas of conformational space and account for novel loop conformations outside of the representation of the knowledge base . To conclude , the DPM-HMM method can be generally applied as an effective and reliable template based loop-modeling algorithm as seen from the results for benchmarking loops from the CASP9 targets .
A dataset of 465 target loops was compiled for this study , as given in Table 1 . For all loops , two anchoring residues on either side were included as anchoring residues . Structural alignments were performed using MUSTANG [45] . Structures of 132 immunoglobulin heavy variable domains at greater than 95% sequence identity were retrieved from the ASTRAL compendium of protein structure [46] . As one of the most common representatives for template-based loop modeling , the three complementarity-determining regions ( CDRs ) from the heavy chain were selected for modeling . According to IMGT numbering scheme [47] , the CDR loop sets were constructed by extracting the residues at following sequence positions: 23–39 for CDRH1 , 56–67 for CDRH2 and 104–118 for CDRH3 . The second loop data set was taken from 92 globin structures which were downloaded from the PDB [48] and structurally aligned , as used previously by Tsai et . al . [36] , [37] . DSSP [49] secondary structure profiles were used to determine the boundaries of the longest loop in the globin fold . This loop connecting helices E and F consists of alignment positions 93–106 in the multiple structural alignments . The third data set consisted of the templates for a CASP9 target , T0617 ( 3nrv ) . The longest loop containing 12 residues was extracted from 21 structurally superposed non-redundant template structures . Table S1 in Text S1 shows PDB identifiers and loop sequence positions for all the data sets used in this study . Table 1 provides the details of the final data sets used and Figure 1 shows structural superposition of all the templates in each dataset . The set is briefly described here . A total of 352 well-defined loops from 132 antibody structures were classified into three major classes as CDRH1 , CDRH2 and CDRH3 . Not all the loop regions are well defined in each PDB , so each set consisted of slightly different numbers of templates . Therefore , 111 protein structures are in the CDRH1 loop set , which is well conserved with 12 residues in each loop structure . CDRH2 contains 130 loop structures and is subdivided in three groups by loop lengths of 7 , 8 and 10 residue loops . CDRH3 is the most diverse dataset , containing a total of 111 loop structures that are grouped by sizes ranging from 8–17 amino acid residues . Next , 92 globin EF loops are grouped into the 3 classes: 12 , 13 , and 15 residue loops . Lastly , 21 target loops were extracted from the template structures to the CASP9 target T0617 and all loops are 12 residues in length . The crystal structure geometry of the backbone atoms ( N , Cα and C ) of the first anchoring residue ( N-terminal ) of the target loop was used to build the models . All the models were built starting from the second residue at N-terminal residue to the last residue at C-terminal . Sharing the first anchoring residue backbone coordinates results in globally superposed loops , so no further superposition is needed . Length of the loops refers to the total number of residues modeled . Although two more residues on both sides of the loop region are included in the sampling , only the second residue from the N-terminus and last two residues at the C-terminus were counted in the total number of residues ( as defined by loop length ) . To show the general applicability of our sampling method outside a specific class or fold in protein family , data was compiled for 90 identified loops modeled during the CASP9 campaign . As this work focuses on loop sampling , only cases were considered where loop regions were identified correctly . For each target protein sequence from 305 putative loops in 45 TBM targets , closely similar templates were identified by a PSI-BLAST [50] search . Template structures were superposed by MUSTANG [45] program as described above . The target sequence was aligned to the multiple templates using the profile alignment function in Muscle [51] . Based on the multiple sequence/structure alignment of target sequence and templates , loop regions in the target sequence were defined . The loop region definition in some of the cases was erroneous depending on the quality and number of templates as well as accuracy of the sequence alignment . Also , loops with no available reference structures were excluded from this analysis . ( See Table S2 in Text S1 for PDB ids of reference structures and positions of loops with their RMSDs ) . For fair comparisons with a common method for loop modeling , the dataset of canonical ( CDRH1 ) and non-canonical ( CDRH3 ) loop conformations were modeled using both DPM-HMM and LoopyMod program [13] , [21] . Although LoopyMod is a complete loop prediction algorithm that includes sampling , scoring and ranking steps , we are interested in only comparing the sampling efficiency of our method to that of LoopyMod . Therefore , scoring and ranking steps in LoopyMod were omitted and all the sampled loop conformations were collected for global RMSD calculations . To simulate a realistic loop-modeling problem , the best template was provided as the input to the LoopyMod and a million conformations were generated . From these , a global backbone RMSDs against the reference crystal structure of the loop was calculated . ( See Figure S1a and S1b in Text S1 for comparison of RMSD of best candidates to the RMSD of best template used by DPM-HMM and LoopyMod methods ) . The joint φ , ψ distribution were estimated using the Dirichlet process mixture of hidden Markov models ( DPM-HMM ) [36] , [37] . Data consists of sequences of angle pairs ( φij , ψij ) , where i = 1 , 2 , 3 , … , n is the index for a particular observed loop and j = 1 , 2 , 3 , … , m is the index for the sequence position within the alignment . The model uses standard Bayesian nonparametrics density estimation techniques to estimate the joint density of all angle pairs across all m positions . Conceptually , it states that the data of loop i across the m positions - ( φi1 , ψi1 ) , ( φi2 , ψi2 ) , … , ( φim , ψim ) - arises from one of many clusters . Each cluster has a unique “centering” backbone angles , whereas members of a given cluster randomly deviate from its cluster center . The cluster to which each loop belongs is uncertain and the method mixes over this uncertainty , providing so-called mixture models . In contrast to many traditional mixture modeling approaches , however , the number of component distributions in our model is theoretically infinite , increasing the flexibility of the model . Naively , one might simply model the “centering” backbone angles of each cluster as being independent , but that would ignore the obvious secondary structure that can readily be inferred from the observed data . Instead , the DPM-HMM considers a hidden Markov model for these “centering” backbone angles . Statistically , this represents a prior distribution on the values of parameters for the bivariate von Mises ( BVM ) sine model . The hidden states consisted of four secondary structure types: coil , helix , strand , and turn . The emission distributions for BVM location ( or “centering” ) parameters were bivariate von Mises sine model mixtures designed to mimic the distributions of torsion angles within each state from the PDB . Conditioning on inferred secondary structure ( i . e . , the hidden state in the Markov chain ) , the location parameters can be very specific . Transition probabilities among the states were also calculated based on observed distributions . Different emission distributions were used at locations containing proline or glycine due to the distinctive properties of these amino acids . ( The emission distributions for scale parameters were identical for all states . ) This informative centering distribution allowed us to leverage information along a sequence to provide informative secondary structure based density estimates even at positions with poor representation in an alignment . Briefly , the formal statistical model can be written as:where p ( ( φ , ψ ) | μ , ν , Ω ) is a bivariate von Mises sine model [52] with mean parameters ( μ , ν ) , μi = ( μi1 , μi2 , μi3 , … , μim ) , νi = ( νi1 , νi2 , νi3 , … , νim ) , Ωi = ( Ωi1 , Ωi2 , Ωi3 , … , Ωim ) , and precision matrix Ω , G is a draw from a Dirichlet process with mass parameter τ and centering distributions H1 for μi , νi , and H2 for Ωi . . H2 is taken to be the product of m identical Wishart distributions with shape parameter α0 and scale matrix β0 , with an expected value of α0/ ( 2β0 ) . The distribution H1 is the hidden Markov model discussed previously , with a state space consisting of four secondary structure classes ( helix , turn , coil , and strand ) each of which is represented by a mixture of between one and five bivariate von Mises sine models . A complete description of this method , including computational details , is provided in [36] . For each density estimate , we ran two Markov chain Monte Carlo ( MCMC ) [53] runs for 11 , 000 iterations with the first 1 , 000 discarded as burn in . Using 1-in-20 thinning , this gave us 1 , 000 draws from the posterior distribution , which forms the basis for our density estimate . For our hyperparameter settings , we took τ = 5 . H2 was the product of m independent Wishart distributions with shape parameter α0 = 2 and a 2×2 scale matrix β0 , which had diagonal elements equal to 0 . 25 and off diagonal elements equal to 0 . Because density estimation is the most computationally intensive portion of our loop-modeling scheme , this approach makes the simplification of only uniquely modeling positions with proline and glycine . For this reason , two loops can produce equivalent posterior distributions if their prolines and glycines appear in the same positions . Additional details on fitting this model and adjustments for gaps in alignment data are provided in previous work [36] . We used the leave one out ( LOO ) approach to model every target in a dataset . For each sampling and prediction run , the target loop is left out: not included as input data for the DPM-HMM density estimation . Remaining loops from the subgroup of the target were then used as templates to model and sample the joint φ , ψ distributions for a target sequence . A set of one million φ , ψ draws from the estimated densities was generated for each of the target loops . For all the one million draws of the torsion angles , all backbone atom models were constructed in Cartesian coordinate space using Self-Normalizing Natural Extension Reference Frame ( SNerf ) algorithm [54] with standard bond length and angle data [52] . In comparison to the initial density estimation , these two steps of making draws from the distribution and building the loop in Cartesian coordinates are relatively fast . About three hours of CPU time are required to sample one million points in φ , ψ space for an average sized target loop ( about 12 amino acid residues ) . Model building from the sampled torsion angle space and filtering using average backbone α carbon ( Cα ) distance takes about 1 . 5 hours of CPU time . The computational expense scales linearly with the number of residues in the loop and the number of models to be built . Three backbone atom coordinates of the first residue at entering N-terminal side of the target loop were used as the anchor for construction of the models in a Cartesian space . So all models in the set are built from the same starting point . To ensure appropriate loop closure , models were refined using a simple distance filter with a 2 . 0 Å cutoff value . This Cα distance filter is very basic using the average distance between the last two Cα atoms of a candidate loop model and those of target loop crystal structure in the loop exit . Since the backbone atom coordinates of first anchoring residue are shared in all the models and the reference structure , this simple filter works well and produces a pool of suitable candidates . Filtered models can be further scored for side chain clashes after grafting on the surface of the whole protein . The DPM-HMM software used in this analysis is available for download at http://www . stat . tamu . edu/~dahl/software/cortorgles/ . | A protein's structure consists of elements of regular secondary structure connected by less regular stretches of loop segments . The irregularity of the loop structure makes loop modeling quite challenging . More accurate sampling of these loop conformations has a direct impact on protein modeling , design , function classification , as well as protein interactions . A method has been developed that extends a more comprehensive knowledge-based approach to producing models of the loop regions of protein structure . Most physical models cannot adequately sample the large conformational space , while the more discrete knowledge based libraries are conformationally limited . To address both of these problems , we introduce a novel statistical method that produces a continuous yet weighted estimation of loop conformational space from a discrete library of structures by using a Dirichlet process mixture of hidden Markov models ( DPM-HMM ) . Applied to loop structure sampling , the results of a number of tests demonstrate that our approach quickly generates large numbers of candidates with near native loop conformations . Most significantly , in the cases where the template sampling is sparse and/or far from native conformations , the DPM-HMM method samples close to the native space and produces a population of accurate loop structures . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"biochemistry",
"mathematics",
"statistics",
"biology",
"computational",
"biology"
]
| 2011 | Near-Native Protein Loop Sampling Using Nonparametric Density Estimation Accommodating Sparcity |
Reassortments and point mutations are two major contributors to diversity of Influenza A virus; however , the link between these two processes is unclear . It has been suggested that reassortments provoke a temporary increase in the rate of amino acid changes as the viral proteins adapt to new genetic environment , but this phenomenon has not been studied systematically . Here , we use a phylogenetic approach to infer the reassortment events between the 8 segments of influenza A H3N2 virus since its emergence in humans in 1968 . We then study the amino acid replacements that occurred in genes encoded in each segment subsequent to reassortments . In five out of eight genes ( NA , M1 , HA , PB1 and NS1 ) , the reassortment events led to a transient increase in the rate of amino acid replacements on the descendant phylogenetic branches . In NA and HA , the replacements following reassortments were enriched with parallel and/or reversing replacements; in contrast , the replacements at sites responsible for differences between antigenic clusters ( in HA ) and at sites under positive selection ( in NA ) were underrepresented among them . Post-reassortment adaptive walks contribute to adaptive evolution in Influenza A: in NA , an average reassortment event causes at least 2 . 1 amino acid replacements in a reassorted gene , with , on average , 0 . 43 amino acid replacements per evolving post-reassortment lineage; and at least ∼9% of all amino acid replacements are provoked by reassortments .
The genome of influenza A virus consists of 8 segments , each represented by an RNA molecule . Coinfection of a cell by viruses of different genotypes occasionally leads to reassortments , i . e . formation of genotypes containing molecules from different sources . Most of the major Influenza A pandemics during the last century were caused by reassortant strains [1] . Indeed , reassortments , especially those creating novel combinations of hemagglutinin ( HA ) and neuraminidase ( NA ) genes , may lead to radical changes in antigenic properties and give rise to viral types that escape the herd immunity . Still , after a reassortment event , the viral segments find themselves in a novel genetic environment , which may lead to disruption of coadaptations that previously existed between them and reduce viral fitness [2] , [3] . Thus , it is likely that only a small proportion of reassortment events lead to creation of novel , successful viral genotypes . Reassortments between different Influenza A subtypes gave rise to the major pandemics of the 1957 , 1968 , 2009 , and possibly 1918 [1] , [4] . Reassortments between strains belonging to a single subtype likely occur much more frequently than inter-subtype reassortments; however , they leave a less pronounced phylogenetic signal , and are therefore harder to study [5] . In theory , reassortments can be detected through the incongruencies between phylogenies of different segments of a viral genome . Indeed , after a reassortment , the segments obtained from the same viral isolate will occupy conflicting phylogenetic positions , due to the differences in their evolutionary histories . In practice , however , detecting reassortments is difficult . The influenza sequence databases are subject to ascertainment biases , with recent sequences being oversampled , and some countries sampled better than others . Reassortment events are prone to be missed when one or both parental strains are not sampled properly , or when they are closely related . Conversely , spurious reassortments may be inferred due to differences in phylogenies caused by phylogenetic noise . Multiple reassortments nested within a single clade compound the difficulties . In most of the early studies , reassortments were inferred via manual detection of incongruencies between phylogenies of different viral segments [6]–[9] . However , this approach is impractical for systematic analyses of large datasets of influenza genomes with complex reassortment histories . Recently , several methods for automatic detection of reassortments have been proposed . These methods can be broadly categorized into two groups . The distance methods [10] , [11] measure , for each viral segment , the degree of similarity between all pairs of viral genomes , and infer reassortments from the differences between the distance matrices obtained from different segments . The phylogenetic methods [4] , [12]–[14] make explicit use of the evolutionary histories of individual segments , comparing their phylogenies and detecting incompatibilities between them . In general , phylogenetic methods are more robust than the distance methods [12] , [14] , particularly in detecting reassortments that became fixed or reached high frequencies within the population . Comparisons of relative frequencies of different progeny coming from coinfecting strains in experiments [3] , [15]–[20] as well as phylogenetic analyses of circulating reassortant strains [10] , [21] demonstrate that reassortments between different segments are not equiprobable . Such differences likely arise , in part , from variance in the extent of epistatic interactions between pairs of genes . For example , a recent analysis of experimental data suggests that polymerase genes ( PB1 and PA ) tend to be inherited together , and that reassortment preferences for HA depend on the subtypes of the parental strains , while NA and matrix protein ( MP ) have no preferences in their reassortments with other segments [19] . Analysis of distributions of coalescent times for different segments suggests , however , that reassortments between HA and NA are particularly frequent [22] . Besides radical shifts in antigenic properties , reassortments are associated with a reduction in genetic diversity of circulating strains , indicative of positive selection favoring the spread of the reassortant strain [22] . Several observations also suggest that , subsequent to the “antigenic shift” , reassortants tend to undergo increased “antigenic drift” , i . e . , elevated rate of amino acid replacements , perhaps due to follow-up coadaptation of genes that find themselves in new genetic environments [5] , [23] . For example , a reassortment associated with host change has led to short-term positive selection in NS gene of swine Influenza A [24] . Still , the phenomenon of coevolution of viral genes subsequent to reassortments has not yet been studied systematically . Here , we purport to close this gap .
To study the within-subtype reassortments in the influenza H3N2 virus , we first used GiRaF [12] to automatically infer the reassortant taxa in the dataset of 1376 complete influenza H3N2 genotypes . GiRaF compares large pools of Monte Carlo-sampled phylogenetic trees constructed for each viral segment separately , inferring the topological incongruencies between them . Assuming that an incongruence between phylogenies of two segments , observed in a high fraction of comparisons , reflects an ancestral reassortment event , GiRaF then predicts the subsets of taxa that are descendant to such reassortments . The number of reassortments predicted by GiRaF in which a particular segment was involved was similar for all segments except M1 and NS1 ( Table 1 ) ; in M1 and NS1 , fewer reassortments were detected , probably because their shorter sequences and/or higher conservation ( Table 1 ) lead to a weaker phylogenetic signal . We saw no preferences for particular pairs of segments to be reassorted ( Table S1 ) . We then mapped the reassortments inferred by GiRaF onto the reconstructed phylogenies of each of the segments involved . We assumed that each reassortment event had happened on the phylogenetic branch leading to the last common ancestor of the reassortant taxa ( reassortment-carrying branch , RCB ) . This way of mapping reassortments was self-evident for monophyletic sets of taxa . However , when the histories of reassortment events are complex ( which is the case for Influenza A; [7] , [9] ) , the subsets of genotypes resulting from a single reassortment predicted by GiRaF are not necessarily monophyletic [12] . Indeed , we found that in our dataset , many of the inferred sets of reassortant taxa were not monophyletic ( Figure 1 ) . To test the robustness of our conclusions , we therefore also tried alternative ways of mapping reassortment events , which also supported our key findings ( see below ) . We asked whether reassortments affect subsequent accumulation of amino acid-changing replacements . We hypothesized that subsequently to RCBs , the rate of accumulation of amino acid replacements will be temporarily elevated as the reassorted sets of genes coadapt to each other . To address this , we inferred , for the genes coded by each of the 8 segments , the phylogenetic positions of all amino acid replacements , and studied those replacements that were descendant to at least one RCB . When a reassortment and a replacement occurred on the same phylogenetic branch , it was impossible to deduce which came first . To avoid this ambiguity , we considered a replacement to be descendant to a particular reassortment if it occurred on a phylogenetic branch descendant to an RCB , but not on the RCB itself . ( Including the substitutions on the RCBs as descendants to reassortments gave similar results; see below . ) Between 33% and 50% ( depending on the segment ) of the RCBs were terminal branches , and thus their effect on subsequent replacements could not be studied . 14% to 33% more of the RCBs were pre-terminal branches . Amino acid replacements on the terminal branches , especially in RNA viruses , tend to be more deleterious , and may follow evolutionary patterns distinct from the replacements elsewhere on the phylogeny [25]–[30]; thus , we chose to exclude such replacements from our analyses . Therefore , the RCBs on pre-terminal branches could also have no effect on replacements in our dataset . The remaining 33% to 42% of RCBs could be followed by amino acid replacements , and in fact , most of the replacements occurred in reassortant clades . In NA , for example , 320 out of 411 ( 78% ) observed non-terminal amino acid replacements were descendant to at least one RCB ( Table 1 ) . To assess the effect of reassortments on subsequent accumulation of amino acid replacements , we measured the phylogenetic distance between each replacement and its most recent ancestral RCB . We then compared these distances to those expected if the phylogenetic positions of post-reassortment replacements were random in respect to reassortments . This approach is conservative , in that it ignores any possible long-term effect of reassortments spanning phylogenetic distances comparable with the height of the phylogenetic tree . In two of the genes , NA and PB1 , the mean distance was significantly lower than expected ( Table 1 ) , indicating that RCBs were followed by a transient increase in the rate of amino acid replacements . Since the inferred number of RCBs for each gene is moderate ( e . g . , 20 for NA , only 14 of which could have descendant replacements ) , individual RCBs could have a disproportionate effect on the distribution of distances . For the NA segment , we asked whether the observed accelerated evolution after reassortments is due to some single reassortment event . To this end , we repeated the analysis 14 times , each time excluding one of the RCBs , and comparing the observed and expected distances between the amino acid replacements and the remaining RCBs . In all 14 comparisons , the results remained significant , indicating that the acceleration of replacements is a phenomenon to which multiple reassortments contribute ( Table 2 ) . To further test its robustness , we repeated the analyses using alternative methods of mapping reassortments onto the tree ( Tables S2 , S3 , S4 , S5 , S6 , S7 ) . These methods differ in the strength of the required statistical evidence for inference of reassortments , and in the ways non-monophyletic sets of reassortant taxa are treated ( see below ) . For NA , acceleration of evolution after reassortments was significant in 5 out of 6 analyses; in the sixth analysis , it was marginally significant ( p = 0 . 09 ) . For PB1 , acceleration was significant in 3 out of 6 analyses . Moreover , in a number of analyses , acceleration was also observed in several other genes for which no significant result is observed in Table 1: M1 in 4 tests , and HA and NS1 in 3 tests each . Thus , 5 out of 8 genes show evidence for acceleration at least in half of the tests . Still , the results for NA are the most robust , and this is the gene for which the evidence for reassortments-caused-acceleration of evolution is the strongest . When the substitutions on the RCB itself were also counted as descendant to the reassortment , the results were similar ( Tables S8 , S9 , S10 , S11 , S12 , S13 ) . In PB1 , the increase in the rate of amino acid replacements after the RCB is rather long-lived: it spans a phylogenetic distance of ∼0 . 04 ds units , although the excess is not significant for most individual distance bins ( Figure 2 ) . In contrast , in NA , this increase is very brief , with most of the excess replacements observed on the very short phylogenetic branches that immediately follow the RCB ( Figure 3 ) . In particular , we observe 30 such replacements at phylogenetic distances up to ∼0 . 003 ds units ( which is just above the time it takes the NA gene to obtain a single synonymous replacement , and is therefore the highest phylogenetic resolution we can achieve; leftmost bin in Figure 3 ) . Because virtually no such replacements would be expected to occur in such a short period of time if they had been independent of reassortments , all these excess 30 replacements are reassortment-provoked . Therefore , at least ∼9% ( 30/320 ) of all amino acid replacements in NA were caused by reassortments; since some of the later replacements descendant to the RCBs could also be reassortment-provoked , the actual number is probably higher . Moreover , the fact that these replacements occur so fast implies that most of them are facilitated by positive selection , and thus comprise a post-reassortment adaptive walk [31] , [32] . What is the length of such adaptive walks , i . e . , the characteristic number of amino acid replacements provoked by an individual reassortment ? The 30 “fast” replacements were descendant to 3 individual RCBs ( Table 2 ) ; 11 more RCBs were not followed by replacements so soon , although they could be followed by reassortment-associated replacements later . Therefore , an average reassortment provoked ∼2 . 1 ( 30/14 ) amino acid replacements in its descendant clade . However , these replacements could occur in multiple independent lineages , so that the number of post-reassortment replacements per lineage was lower . Indeed , by the time the phylogenetic distance of 0 . 003 ds units after an RCB was reached , the descendant subtree had often multifurcated , so that these 3 RCBs gave rise to a total of 33 individual descendant lineages; 36 more lineages originated from the remaining 11 RCBs . In 2 out of the 3 RCBs that provoked replacements , different post-RCB lineages accumulated replacements independently ( Table 2 ) . As a result , over the evolutionary time of 0 . 003 ds units , an average post-RCB lineage attained 0 . 43 ( 30/69 ) reassortment-provoked replacements ( Table 2 ) . This is the lower boundary for the length of the “adaptive walk” per lineage associated with a reassortment event , as it excludes any effect of reassortments over longer timescales . We asked whether the post-RCB amino acid replacements are enriched in particular classes of mutations , compared to the rest of the replacements . In this analysis , we considered the NA gene , because in it , the effect of post-reassortment adaptive walk is the most robust; and also the HA gene , because it is the other primary determinant of antigenic properties , is known to evolve under continuous positive selection , and is highly epidemiologically relevant . Several categories of mutations had biased phylogenetic distances from the RCBs , compared with the complementary sets ( Table 3 ) . Firstly , in NA , the replacements at amino acid sites experiencing positive selection tended to be farther from RCBs , while no such difference was observed for HA . Secondly , the replacements at sites that distinguish the antigenic clusters [33] of HA tended to occur farther from RCBs . Thirdly , parallel replacements had a strong tendency to occur soon after the RCBs both in NA and HA . Fourthly , reversions in NA , but not in HA , occurred soon after the RCBs . Fifthly , the sites previously shown to be involved in intragenic epistatic interactions [34] as “leading” occurred father from RCBs both in NA and in HA , while the “trailing” sites occurred soon after the RCBs . Neither the NA nor the HA genes showed any bias for replacements at the epitopic sites . As was the case with the previous analysis , most of these results were also supported by alternative methods of inferring RCBs ( Tables S14 , S15 , S16 , S17 , S18 , S19 ) ; the exception were reversions in NA , which gave discordant results in different tests .
Our study provides the first systematic analysis of association between reassortments and amino acid-level changes in influenza A . We show that a reassortment involving a particular segment provokes a transient increase in the rate of amino acid replacements at the gene encoded on this segment; these replacements tend to occur at sites that do not normally experience positive selection , and often involve parallel replacements . One way to estimate the effect of RCBs on subsequent accumulation of amino acid replacements would be to directly compare the replacement rates on branches that had descended from reassortments and those that had not . However , this is not feasible in Influenza A , because in most segments , the majority of branches ( 50–85% ) descend from at least one RCB ( Table 1 ) ; the remaining branches tended to be deep-lying , and thus likely have biased patterns of replacements [29] , [30] . Instead , we searched for a transient increase in the rate of amino acid changes on the branches descendant to the internal RCBs . This analysis could miss some of the very rapid reassortment-provoked changes that had happened on the same branches as the reassortments themselves . We compared the phylogenetic distances between the amino acid replacements and the preceding RCBs to the null distribution expected under the assumption that the amino acid replacements were distributed over post-RCB , non-terminal branches randomly , with probability of a replacement to fall onto a particular branch proportional to the branch length . In NA and PB1 genes , the replacements occur sooner after the RCBs than in the null model . When alternative strategies of mapping reassortments were used , three other genes – M1 , HA and NS1 – also show the same pattern in at least half of the tests . Since it is hard to validate the reassortment-mapping algorithms , the results of each individual analysis for each individual gene should be taken with some caution . Together , however , they provide strong support for the reassortment-caused accumulation of changes , especially in the NA gene . The post-RCB excess of amino acid replacements was not exclusively dependent on a single particular RCB ( Table 2 ) ; thus , it seems to be a universal phenomenon . Non-uniformity of the substitution rate has been described previously for Influenza , and has been attributed to episodic action of positive selection [35] , to simultaneous fixation of multiple interacting advantageous mutations [36] or to frequent selective sweeps under clonal interference . There is no obvious mechanistic association between reassortments and either of these factors . If reassortments themselves are adaptive , they can spread through population rapidly by means of positive selection [22] . Strong positive selection favoring an allelic variant may also drive to fixation neutral and even mildly deleterious point mutations linked with it [37]; this phenomenon is prevalent in Influenza A which evolves , on the within-segment level , under nearly complete linkage [38] , and where clonal dynamics is largely determined by linkage with beneficial alleles [38] , [39] . This phenomenon could cause an excess of replacements on the same branches as the reassortment events , which is , however , not observed in our data ( data not shown ) . There is no way hitchhiking can cause accumulation of replacements on the branches descendant to RCBs . Therefore , the only feasible explanation for the post-reassortment increase of the rate of amino acid replacements seems to be that they constitute an adaptive walk [31] , [32] , i . e . , a burst of positively selected adaptive changes provoked by a shift in the fitness landscape . In NA , we observe a radical increase in the rate of amino acid replacements immediately after reassortments . This excess spans only a short period of time: it is mostly over by 0 . 003 ds units , i . e . , by the time a single synonymous replacement occurs somewhere in the gene ( which takes less than a year; [40] ) . Thus , at least NA , and very likely other genes , experience transient positive selection after reassortment events . The excess of replacements at phylogenetic distances up to 0 . 003 ds ( i . e . , up to the time a single neutral replacement is expected to be accumulated at the gene ) in NA suggest that a total of ∼30 amino acid replacements in the entire phylogeny , or ∼0 . 43 replacements per lineage , were facilitated by preceding RCBs . Therefore , reassortments are responsible for at least ∼9% of all amino acid replacements in this gene . In fact , this may be an underestimate for at least two reasons . First , some of the reassortments were likely to have been undetected . Second , a fraction of the adaptive walks could have spanned longer phylogenetic distances than this threshold ( Figures 2 , 3 ) . What is the cause of the post-reassortment accumulation of positively selected replacements ? Influenza A is a model system for studying positive selection , with most of the selective pressure exerted by the host immune system . Positive selection is most pronounced in the genes coding for the surface glycoproteins ( NA and HA ) , and within these genes , at the epitopic sites which are most involved in the immune response . Conceivably , the immunity-driven positive selection could increase after a reassortment . We observe , however , that the post-RCB adaptive walk is mostly manifested at sites other than those under constant positive selection , or responsible for antigenic properties , and is not affected by the epitopic vs . non-epitopic location of the site ( Table 3 ) . This suggests that the post-reassortment adaptive walks are not driven by the pressure to evade the host immune system . Rather , these replacements are probably associated with epistatic interactions between genes [41] . In general , reassortments and host shifts lead to changes in the patterns of both synonymous and nonsynonymous substitutions , probably due to joint effects of changes in the mutation and selection patterns [29] , [42]–[44] . After a reassortment event , a gene finds itself in a novel genetic environment which may , through epistatic interactions , exert novel selective pressures on its amino acid sites , facilitating further amino acid changes . An adaptive walk could compensate for the loss of fitness associated with the preceding reassortment [2]; however , as the reassortments themselves tend to be adaptive [22] , it seems more likely that these replacements could exploit novel fitness peaks that have become newly accessible after the reassortments . The evidence for the post-reassortment adaptive walk is the most robust for the NA gene . Sequence evolution of influenza NA does not always lead to changes in antigenic properties [45] , and may be caused by other forces instead . Indeed , antibody-driven affinity-changing mutations in HA can be compensated by substitutions changing the activity of NA [46] , [47]; this indicates that the choice of the optimal NA genotype is dependent at least on HA , and possibly on other genes as well . Furthermore , reassortments often involve a currently circulating strain and an older strain [48] , and a reassortment between HA and NA frequently involves an up-to-date variant of HA and an older variant of NA , as suggested by less discordance between sampling time and phylogenetic position of HA sequences than of NA sequences ( [7] and our data ) . Therefore , while the immune escape is the primary factor of evolution of HA , much of the NA evolution may be epistatic and , in particular , compensatory . Parallel replacements are overrepresented after reassortments ( Table 3 ) . Overall , the rate of parallelism in Influenza A evolution is high [49] , [50] , probably due to similar selective pressures exerted on different strains . The high parallelism observed in this study suggests that the replacements involved in a post-reassortment adaptive walk may also be adaptive in other contexts . Epistatic interactions both within [51]–[54] and between segments [55] are wide-spread in Influenza A . One evolutionary manifestation of this phenomenon is positive epistasis between replacements: a replacement can facilitate subsequent replacements at different sites of the same protein [34] . The sites involved in such epistatic interactions can be classified as “leading” or “trailing” , depending on whether replacements in them tend to come as first or second in epistatic pairs; for example , replacements at leading sites can introduce radical changes to protein structure , while replacements at trailing sites may compensate those changes [34] . We find that , while the replacements at leading sites are remote from reassortments , the replacements at trailing sites occur sooner after reassortments than expected . Therefore , the sites experiencing post-reassortment replacements are the same sites that also react to the change of the protein structure due to replacements elsewhere in the protein . This suggests that the class of sites denoted as “trailing” in [34] , and involved in post-adaptive walk in this study , may be responsible for adaptation to novel genetic environment that stems from changes in the same gene as well as in other genes . Association between reassortments and the rate of subsequent accumulation of amino acid mutations may be important for predicting future pandemic strains . For example , the avian H5N1 influenza is among the most likely candidates for the agent of a future pandemia [56]–[58] . Naturally occurring strains of A/H5N1 are not transmittable between mammals; however , to become transmittable , they require just five additional mutations [59] or a reassortment with just four additional mutations [60] . Two of these mutations are already frequent among the A/H5N1 viruses [61] . If a reassortment commonly leads to accelerated accumulation of amino acid replacements , gaining the remaining mutations and evolving a natural mammalian-transmittable H5N1 strain may take less time than predicted by simple models [61] .
We downloaded all complete human H3N2 influenza A genotype sequences ( N = 2205 ) available on 27 . 10 . 2011 from the flu database [62] . Nucleotide sequences for each segment were aligned using muscle [63] , [64] . Genotypes containing truncated sequences , multiple unidentified nucleotides , or indels were discarded . We used CD-HIT [65] to cluster genotypes that had identical sequences of NA segments , and retained one random sequence from each cluster , thus retaining 1379 genotypes for further analysis . For segments encoding PB1 , M1 and NS1 that contain overlapping ORFs , we excluded the overlapping regions , and analyzed the longest remaining ORF . All alignments are available at http://makarich . fbb . msu . ru/flu_walks/ . For each segment , we Bayes-sampled the 1 , 000 phylogenetic trees using MrBayes MPI version [66]–[68] with the following settings: GTR+I+G model , 22 million iterations , sampling each 22 , 000th iteration . Three isolates: A/Ontario/RV123/2005 , A/Ontario/1252/2007 and A/Indiana/08/2011 were excluded from analyses , because we found the branch leading to the clade formed by them to be , for several segments of non-human origin ( NP , M , NS , PB2 and PA ) , too long for a meaningful estimation of evolutionary parameters; these isolates are SOIV triple reassortants ( see also [69] ) . Each phylogenetic tree was rooted by the isolate A/Albany/18/1968 . These 1 , 000 trees were used to infer the reassortment events ( see below ) . For each segment , the consensus tree of the 1 , 000 MrBayes-sampled trees was used as input for HyPhy [70] to estimate the evolutionary parameters and to restore the ancestral sequences . The branch-specific dS values were estimated using the local MG94xHKY85 [71] model . The ancestral sequences were reconstructed using the GTR+I+G global nucleotide model . As an alternative approach , we also repeated our analyses using maximum likelihood trees constructed with PhyML [72] instead on consensus Bayesian trees , and obtained similar results . For subsequent analyses , we rescaled the lengths of all branches in the consensus trees in the units of dS; this allowed us to study the distribution of nonsynonymous replacements independently of branch lengths . To obtain the gene-specific dN/dS values , we used global MG94xHKY85 model in HyPhy . For phylogenetic mapping of reassortments , we used a two-step procedure . First , we computationally predicted the subsets of taxa that occupied incompatible positions in phylogenies of different segments using GIRAF software [12] running on a cluster node with 512 Gb of RAM . GiRaF automatically predicted the subsets of taxa originating from each ancestral reassortment event on the basis of the MrBayes sampled trees for all eight segments . Importantly , not all segments were necessarily involved in each inferred reassortment; although in reality each reassortment splits all segments into two subsets ( the retained and the acquired segments ) , for some of the segments , the phylogenetic signal was often too weak to allow GiRaF to ascribe them to one of the two mutually reassorting sets of segments [12] . For such segments , no reassortment event could then be inferred; as a result , although the same sets of taxa were considered for each segment , the number of reassortment events per segment varied ( Table 1 ) , and the number of segments involved in each reassortment ( on either side ) was usually under 8 , with some of the segments “abstaining” . We used two approaches to quality filtering of the predicted reassortments . In the first approach ( “joint reassortments” , recommended in [12] and used in the main text ) , we acquired the subsets of taxa involved in reassortments from the “catalog file” produced by GiRaF . This file includes only those reassortments that involved inconsistencies between at least 3 pairs of segments; therefore , each reassortment involved between 4 and 8 segments . In the second approach ( “high-confidence reassortments” ) , we used the reassortment subsets involving any number of pairs of segments ( i . e . , between 2 and 8 segments ) , but required the GiRaF-predicted confidence level for the reassortments of 1 . 0 . Second , we inferred , on the basis of these lists of reassortant taxa , the phylogenetic positions of the RCBs . In theory , each reassortment event should give rise to a monophyletic set of taxa; the last common ancestral branch to this clade is then the RCB . In reality , however , many of the predicted subsets of taxa were not monophyletic . This occurs because , under complex histories of sequential reassortments , GIRAF can either split the taxa descendant to a common reassortment event into multiple sets , or join the taxa descendent from multiple reassortment events into a single set [12] . In such cases , the inference of RCBs is ambiguous . We used three different approaches to infer the phylogenetic position of the RCBs . While these approaches produced identical results for monophyletic sets of reassortants , they differed in the way they treated non-monophyletic sets . For each set of reassortants , we inferred as RCB ( s ) ( i ) the ( single ) branch leading to the most recent common ancestor of all reassortants ( “one-point inference” , used in the main text ) ; ( ii ) the set of branches leading to the most recent common ancestors of each clade involving only reassortants ( “two-point inference” ) ; or ( iii ) the union of ( i ) and ( ii ) ( “three-point inference” ) . Arguably , each approach has its merits . Under ( i ) , the number of inferred RCBs is minimal ( and equal to the number of sets of reassortant taxa ) , and so this approach is most parsimonious; conversely , under ( ii ) and ( iii ) , a single subset of reassortant taxa could give rise to multiple RCBs on the same phylogenetic tree . Under ( ii ) , only reassortant taxa are descendants to RCBs; finally , ( iii ) may be best for inference of sequences of nested reassortments such that a later reassortment affects a subset of lineages that were also involved in an earlier reassortment , and GIRAF underpredicts the set of reassortant taxa for the earlier reassortment . All GiRaF output files and the phylogenetic trees in Nexus FigTree ( http://tree . bio . ed . ac . uk/software/figtree/ ) format with reassortments mapped onto them are available at http://makarich . fbb . msu . ru/flu_walks/ . For NA , we used two approaches for validation of the observed reassortments . First , we compared our RCBs with the reassortments inferred previously on the basis of manual analysis of a much smaller dataset [7] . Out of the 6 reassortments reported in [7] , 5 mapped precisely to 4 of our RCBs ( accounting for the strains missing in [7] , and including one pre-terminal RCB; two of the reassortments mapped to the same RCB ) . The remaining reassortment mapped to a branch adjacent to an RCB ( Table S20 ) . The reassortments from [7] are included for comparison with our RCBs in the phylogenetic trees available at http://makarich . fbb . msu . ru/flu_walks/ . Second , we analyzed the inconsistencies in the dates of sampling of strains . In general , the sampling dates of Influenza strains are highly correlated with their distance from the root on the “cactus-shaped” Influenza phylogeny [33] , consistently with the major role of continuous positive selection on immunity avoidance shaping it [73] . However , recombination may lead to disruptions of this order , because a strain nested deep in the phylogeny may have a recent sampling date if it has reassorted with another recent strain . We inferred the inconsistencies in the sampling dates as follows ( the approach is similar to that used in [48] ) . We split all our strains ( including those identical by NA sequence ) into subsets depending on their most recent RCB , and further into smaller subsets based on the year of sampling . We then built a consensus sequence for each of these subsets , excluding those that carried fewer than 10 sequences , and used these consensuses to construct a new ML phylogenetic tree with branches scaled in ds units as described above . We then rotated the branches of the tree to order them by the sampling year . Some of the branches could not be thus ordered; i . e . , had sampling dates inconsistent with their phylogenetic position . Specifically , seven of the branches , descendant to four inferred reassortments , had sampling dates later than some of the branches found to the right of them on the tree ( Figure S1 ) , supporting their origin from reassortment events . The reassortments that corresponded to these branches were among the top-ranking in our analysis ( Table 2 ) . The method of inference of reassortments based on inconsistencies in sampling dates of a single segment is orthogonal to our main approach based on inconsistencies between phylogenies of different segments; therefore , it provides an independent validation for the reassortments that we detect . Using the reconstructed ancestral sequences , we inferred , for each segment , the phylogenetic positions of all nonsynonymous replacements . We then measured the distance between each nonsynonymous replacement and its ancestral reassortment . If multiple RCBs were ancestral to a given replacement , we considered the most recent one . In measurements of phylogenetic distances , we assumed that reassortments occurred at the middles of the RCBs , and that each nonsynonymous replacement occurred at the middle of the corresponding phylogenetic branch . We used two approaches for dealing with replacements on the RCBs themselves: they were considered to be either ancestral to this reassortment , and thus excluded from the list of its descendants ( Tables 1–3 , S2 , 3 , S4 , S5 , S6 , S7 , S14 , S15 , S16 , S17 , S18 , S19 ) ; or descendant to this reassortment , with the distance between the reassortment and the replacement equal to zero ( Tables S8 , S9 , S10 , S11 , S12 , S13 ) . The distances were measured in dS units . To obtain the expected phylogenetic distribution of the replacements , we , in 10 , 000 Monte Carlo trials , redistributed the replacements among the tree branches . The probability of a replacement to fall onto a given branch was taken to be proportional to its dS value . We excluded from reshuffling the branches that were not descendant to at least one RCB , the terminal branches , and ( for Tables 1–3 , S2 , S3 , S4 , S5 , S6 , S7 , S14 , S15 , S16 , S17 , S18 , S19 ) the RCBs themselves; since replacements on those branches are not included in corresponding analyses , the number of analyzed replacements was thus conserved . For the analysis of the length of adaptive walk in NA , we considered a replacement reassortment-provoked if it occurred within 0 . 003 ds units after reassortment . We traced the number of phylogenetic lineages descendant to each RCB at this time point , and calculated the per-lineage length of the adaptive walk by dividing the number of replacements by the number of lineages . Post-RCB replacements were considered phylogenetically independent if none of them were descendant to any of the remaining ones; the number of such replacements equaled the number of lineages carrying replacements at distance of 0 . 003 ds . Positively selected sites were inferred by IFEL [70] and MEME [74] methods from the HyPhy package; a site was considered positively selected if it was predicted by either of these methods . Epitopic sites were taken from [34] , and sites distinguishing the HA antigenic clusters , from [33] . Replacements from a particular ancestral amino acid to a particular descendant one that occurred at more than one lineage on the phylogeny were categorized as parallel , and replacements that reverted to a once-ancestral state , reversing . All manipulations with phylogenetic trees were done using the Perl Bio::Phylo package [75] . The statistical analyses were performed with R [76] . | Influenza A is a rapidly evolving virus with genome composed of eight distinct RNA molecules called segments . This genetic structure allows formation of new combinations of segments when a cell is coinfected by multiple viral strains , in a process called reassortment . While “antigenic drift” – the process of continuous accumulation of point mutations that change the antigenic properties of the viral proteins – is mainly responsible for the seasonal flu , the heaviest pandemics were caused by spread of novel reassortant strains and the associated radical “antigenic shift” . However , the association between these two types of processes has not been studied systematically . Here , we use the extensive available complete-genome sequencing data for Influenza A H3N2 subtype to infer the evolutionary timings of within-subtype reassortment events , and study the patterns of point amino acid-changing replacements that followed reassortments . We find that reassortments were often rapidly followed by replacements , which possibly compensated for the loss of fitness associated with the reassortment or explored newly accessible fitness peaks . These findings may be relevant for prediction of future pandemic strains of Influenza A . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"organismal",
"evolution",
"genome",
"evolution",
"population",
"genetics",
"microbiology",
"parallel",
"evolution",
"mutation",
"microbial",
"evolution",
"forms",
"of",
"evolution",
"comparative",
"genomics",
"biology",
"evolutionary",
"theory",
"evolutionary",
"genetics",
"viral",
"evolution",
"natural",
"selection",
"virology",
"co-infections",
"genomics",
"evolutionary",
"biology",
"genomic",
"evolution",
"evolutionary",
"processes"
]
| 2014 | Intrasubtype Reassortments Cause Adaptive Amino Acid Replacements in H3N2 Influenza Genes |
The Ufm1 conjugation system is an ubiquitin-like modification system that consists of Ufm1 , Uba5 ( E1 ) , Ufc1 ( E2 ) , and less defined E3 ligase ( s ) and targets . The biological importance of this system is highlighted by its essential role in embryogenesis and erythroid development , but the underlying mechanism is poorly understood . UFBP1 ( Ufm1 binding protein 1 , also known as DDRGK1 , Dashurin and C20orf116 ) is a putative Ufm1 target , yet its exact physiological function and impact of its ufmylation remain largely undefined . In this study , we report that UFBP1 is indispensable for embryonic development and hematopoiesis . While germ-line deletion of UFBP1 caused defective erythroid development and embryonic lethality , somatic ablation of UFBP1 impaired adult hematopoiesis , resulting in pancytopenia and animal death . At the cellular level , UFBP1 deficiency led to elevated ER ( endoplasmic reticulum ) stress and activation of unfolded protein response ( UPR ) , and consequently cell death of hematopoietic stem/progenitor cells . In addition , loss of UFBP1 suppressed expression of erythroid transcription factors GATA-1 and KLF1 and blocked erythroid differentiation from CFU-Es ( colony forming unit-erythroid ) to proerythroblasts . Interestingly , depletion of Uba5 , a Ufm1 E1 enzyme , also caused elevation of ER stress and under-expression of erythroid transcription factors in erythroleukemia K562 cells . By contrast , knockdown of ASC1 , a newly identified Ufm1 target that functions as a transcriptional co-activator of hormone receptors , led to down-regulation of erythroid transcription factors , but did not elevate basal ER stress . Furthermore , we found that ASC1 was associated with the promoters of GATA-1 and Klf1 in a UFBP1-dependent manner . Taken together , our findings suggest that UFBP1 , along with ASC1 and other ufmylation components , play pleiotropic roles in regulation of hematopoietic cell survival and differentiation via modulating ER homeostasis and erythroid lineage-specific gene expression . Modulating the activity of this novel ubiquitin-like system may represent a novel approach to treat blood-related diseases such as anemia .
The Ufm1 ( Ubiquitin-fold modifier 1 ) conjugation system is a novel ubiquitin-like ( Ubl ) modification system that mediates protein modification by a small protein modifier Ufm1 [1] . Ufmylation is catalyzed by a series of enzymes , consisting of Ufm1-activating E1 enzyme ( Uba5 ) , Ufm1-conjugating E2 enzyme ( Ufc1 ) , and a newly identified E3 ligase Ufl1 [1–3] . The genetic study using Uba5 knockout mice demonstrates that Uba5 is essential for embryogenesis and erythroid development , highlighting the importance of this system in animal development [4] . Yet , its downstream effector ( s ) and molecular mechanism remain poorly understood . UFBP1 ( Ufm1 binding protein 1 with a PCI domain , also known as C20orf116 , Dashurin , and DDRGK1 ) is a highly conserved protein whose orthologues are found in metazoan and plants but not in yeast , indicating its important role in multicellular organisms [2 , 5–7] . It is ubiquitously expressed in multiple tissues and organs with high- level expression in secretory cells [7] . Human UFBP1 contains 314 amino acid residues with a putative signal peptide at its N-terminus that anchors it to the endoplasmic reticulum ( ER ) [2 , 6] . It also contains a partial PCI domain that is frequently found in the subunits of the proteasome , COP9 signalosome and translation initiation factors . UFBP1 is found to associate with two other proteins namely C53 ( also designated as LZAP , Cdk5rap3 ) and Ufl1 ( also known as KIAA0776 , NLBP , RCAD and Maxer ) [2 , 6] . Interestingly , Tatsumi et al found that UFBP1 was a Ufm1 target and its ufmylation at the highly conserved residue Lys267 was promoted by the Ufm1 E3 ligase Ufl1 [2] . Moreover , Yoo et al have recently demonstrated that the ufmylated form of UFBP1 is a co-component of Ufl1 E3 ligase that promotes ufmylation of nuclear receptor co-activator ASC1 , an event that is crucial for estrogen receptor signaling and breast cancer development [8] . Nonetheless , little is known concerning the in vivo function of UFBP1 and the functional impact of its ufmylation . In this study , we report the essential role of UFBP1 in murine development and hematopoiesis . Ablation of UFBP1 leads to impaired embryogenesis and defective hematopoiesis . Our results strongly suggest that UFBP1 functions as a key player in the ufmylation pathway that is crucial for cellular homeostasis and erythroid development .
In an attempt to understand UFBP1’s in vivo function , we generated UFBP1 knockout mice . The murine UFBP1 gene is located in chromosome 2 and consists of 9 exons ( Fig 1A ) . A gene trap cassette flanked by two FRT sites was inserted into the intron between exon 2 and 3 and followed by floxed exons 3 and 4 , thereby creating a UFBP1Trap-F allele ( Fig 1A ) . Insertion of the gene trap was confirmed by genomic PCR ( Fig 1B ) . Disruption of UFBP1 expression was confirmed by the complete absence of UFBP1 protein in the embryos with homozygous trapped alleles ( Fig 1C ) . Therefore , the mice with homozygous trapped alleles ( UFBP1Trap-F/Trap-F ) were designated as UFBP1-/- or UFBP1 KO mice . Among 158 adult mice , we failed to obtain UFBP1-/- animals while 51 UFBP1+/+ and 107 UFBP1+/- mice were born healthy and appeared normal , indicating a possible embryonic lethality caused by loss of UFBP1 . By analyzing the embryos of timed-pregnant mice , we found that most UFBP1-/- embryos died at E12 . 5 ( Fig 1D ) , a phenotype which is similar to the ones of Uba5 and Ufl1 KO mice but slightly delayed compared to Ufl1 KO mice [9] . E11 . 5 UFBP1-/- embryos appeared smaller and paler than their wild-type littermates ( Fig 1E and S1 Fig ) . Extensive cell death ( TUNEL staining ) was observed in brain , liver and other tissues ( Figs 1F , 1G and 2B ) . Taken together , our results suggest that UFBP1 is essential for proper embryogenesis and its deficiency leads to elevated cell death . Ablation of either Uba5 or Ufl1 leads to defective embryonic erythropoiesis [4 , 9] . Therefore , it is of great interest to determine if UFBP1 deficiency also impacts on embryonic erythroid development . At E11 . 5 , UFBP1 KO embryos had fewer peripheral circulating erythrocytes compared to wild-type embryos ( Fig 2A ) , while UFBP1 deficient yolk sac ( E8 . 5 ) contained a reduced number of primitive erythroid colony-forming cells ( EryP-CFC ) ( Fig 2B ) . Additionally , there were a large number of abnormal multinucleated erythroid cells in UFBP1-/- embryos , a phenotype reminiscent of Uba5 and Ufl1 null embryos ( Fig 2C ) [4 , 9] . Some multi-nucleated erythroid cells were TUNEL-positive , indicating occurrence of cell death of these cells ( Fig 2D ) . These results indicate an impairment of primitive erythropoiesis in UFBP1 deficient embryos . We also examined definitive erythropoiesis in fetal livers . UFBP1 KO embryos ( E11 . 5 ) had small and dis-organized fetal livers with less cellularity ( Fig 2E ) and increased cell death ( Fig 2F ) . Flow cytometry analysis of E11 . 5 fetal liver cells showed that UFBP1 deficiency significantly blocked differentiation of TER119low cells into TER119med+high cells ( proerythroblasts and basophilic erythroblasts ) cells ( Fig 2G ) . Furthermore , the numbers of erythroid and myeloid progenitor cells , including CFU-Es ( colony-forming units-erythrocytes ) , BFU-Es ( erythroid burst-forming units ) , CFU-GM ( colony-forming units- granulocyte/macrophage ) and CFU-GEMM ( colony-forming units- granulocyte/erythrocyte/macrophage/megakaryocyte ) , were significantly lower in UFBP1-/- fetal livers ( E11 . 5 ) ( Fig 2H ) . Taken together , our results suggest that UFBP1 plays an indispensable role in both primitive and definitive erythropoiesis during embryonic development . To investigate the in vivo role of UFBP1 in adult erythropoiesis , we generated inducible conditional KO ( CKO ) mice of UFBP1 via a two-step procedure: 1 ) removing the gene trap cassette by crossing UFBP1Trap-F/+ mice with FLPo deleter mice; and 2 ) crossing UFBP1 floxed mice with ROSA26-CreERT2 mice . As shown in Fig 3A , endogenous UFBP1 in bone marrow ( BM ) cells was depleted in tamoxifen ( TAM ) -treated UFBP1F/F:ROSA26-CreERT2 mice ( designated as UFBP1F/F:CreERT2 hereafter ) but not in UFBP1F/F mice ( Fig 3A ) . Interestingly , UFBP1-depleted mice suffered substantial loss of body weight and died around 3 weeks after initial TAM treatment , while RCADF/F mice were fairly normal ( Fig 3B ) . In the hematopoietic system , UFBP1 deficient mice exhibited severe pancytopenia ( Fig 3C ) . The major indices for erythrocytes , including RBC ( red blood cell ) count and Hgb ( hemoglobin ) content , were significantly lower in UFBP1 deficient mice . Moreover , the counts for lymphocytes , granulocytes , monocytes , as well as platelet counts were also dramatically decreased in UFBP1 deficient mice ( Fig 3C ) . This result strongly indicates an essential role of UFBP1 in adult hematopoiesis . To further define the role of UFBP1 in adult erythropoiesis , we analyzed erythroid development in TAM-treated UFBP1 CKO mice according to Pronk et al [10] . The myeloerythroid ( Lin-Sca-1-c-kit+IL-7R- ) fraction of BM cells was subject to flow cytometry analysis using indicated markers ( Fig 4A ) . As shown in Fig 4A , significant changes in myeloid and erythroid lineages were observed in UFBP1 deficient BM cells . The percentages of all types of erythroid progenitors , including Pre MegE ( 4 . 40 ± 0 . 81% vs 1 . 42 ± 0 . 16% ) , Pre CFU-E ( 0 . 89 ± 0 . 22% vs 0 . 25 ± 0 . 07% ) , and CFU-E plus proerythroblasts ( 41 . 23 ± 0 . 78% vs 10 . 43 ± 0 . 32% ) , were significantly decreased in TAM-treated UFBP1F/F:CreERT2 mice compared to UFBP1F/F mice ( Fig 4A and 4B ) . Moreover , differentiation from CFU-Es ( TER119low ) to proerythroblasts ( TER119high ) was almost completely blocked by loss of UFBP1 ( Fig 4A , the last panel ) . By contrast , the percentage of GMPs ( 9 . 24 ± 0 . 51% vs 57 . 40 ± 0 . 95% ) was substantially increased in UFBP1 deficient BM , while the percentage of Pre GMs ( 31 . 60 ± 0 . 17% vs 19 . 9 ± 0 . 64% ) was modestly decreased ( Fig 4A and 4B ) . The total cell number of erythroid progenitors ( CFU-Es + proerythroblasts ) was also reduced in UFBP1 deficient BM , suggesting that UFBP1 plays a crucial role in development of erythroid lineage . Interestingly , although the number of GMPs in the UFBP1-deleted BM was dramatically expanded ( Fig 4C ) , the number of mature monocytes/granulocytes in peripheral blood was decreased ( Fig 3C ) , indicating that UFBP1 may also have a role in development of myeloid lineages . In addition to impaired erythropoiesis , UFBP1 deficient mice also exhibited cytopenia of other types of hematopoietic cells ( Fig 3C ) , indicating a possible loss of HSC function . In line with this hypothesis , UFBP1 deficient BMs failed to rescue lethally irradiated recipient mice ( S2 Fig ) . To further assess the cell intrinsic role of UFBP1 in HSC function , we performed competitive repopulation assays . Unfractionated BM cells from either UFBP1F/F or UFBP1F/F:CreERT2 ( CD45 . 2 ) were mixed with wild-type ( CD45 . 1 ) BM cells at a 1:1 ratio , and co-transplanted into lethally irradiated recipient CD45 . 1 mice ( Fig 5A ) . Four weeks after transplantation , mice were treated with either oil or tamoxifen . Three weeks after initiation of the treatment , we examined the contribution of donor cells ( CD45 . 2 ) to HSCs and other progenitor cells in BM . As shown in S3 Fig , the contribution of UFBP1 deficient ( UFBP1F/F:CreERT2 ) cells to Lin-Sca-1+c-Kit+ ( LSK ) population was substantially diminished after TAM treatment comparing to oil treatment ( 32 . 78 ± 1 . 79% vs 3 . 35 ± 0 . 28% ) . By contrast , TAM treatment yielded no effect on the contribution of UFBP1F/F cells ( S3 Fig ) . Further analysis using CD150 marker showed that the contribution of UFBP1 deficient cells to both long-term HSCs ( L-S+K+CD150+ ) ( 5 . 95 ± 1 . 02% vs 0 . 73 + 0 . 30% ) and multipotent progenitors ( L-S+K+CD150- ) ( 27 . 58 ± 1 . 24% vs 3 . 13 ± 0 . 31% ) was significantly reduced ( Fig 5B ) . Similarly , oligopotent progenitors ( L-S-K+ ) ( 38 . 18 ± 1 . 08% vs 2 . 08 ± 0 . 21% ) contained much lower level of UFBP1 deficient cells ( Fig 5B ) . The result of competitive repopulation assays strongly suggests that UFBP1 is essential for HSC function in a cell-autonomous manner . UFBP1 and other components of the Ufm1 system play a cytoprotective role in ER stress-induced apoptosis [7 , 11] . Therefore , we examined the effect of UFBP1 deficiency on ER homeostasis . Expression of ER chaperones Grp78 and ERdj4 was up-regulated in UFBP1 deficient BM cells ( Fig 6A ) , and UFBP1 depletion induced Xbp-1 mRNA splicing ( Fig 6B ) and elevation of phosphorylation of eIF2α and JNK ( Fig 6C ) . These results indicate that UFBP1 deficiency indeed led to elevated ER stress and activation of the UPR . Furthermore , a subset of apoptotic genes associated with ER stress , including Bax , Bak , Noxa , Puma and DR5 , were up-regulated in UFBP1 deficient cells ( Fig 6D ) . In contrast to UFL1 deficient cells , however , we did not observe apparent inhibition of autophagic flux in UFBP1 deficient BM cells ( S4 Fig ) . We also examined the effect of UFBP1 depletion on LSK cells in vitro . Sorted UFBP1F/F:CreERT2 LSK cells were treated with 4-hydroxytamoxifen ( 4-OHT ) and cultured for indicated time . 4-OHT-induced depletion of UFBP1 led to a significant decrease of proliferation and increase of cell death of LSK cells ( Fig 6E and 6F ) . Consistent with the in vivo result , the UPR and cell death genes were up-regulated in UFBP1-depleted LSK cells ( Fig 6G and 6H ) . Collectively , our results suggest that UFBP1 plays an essential role in survival of LSK progenitor cells by maintaining ER homeostasis . It has been shown that Uba5 is essential for embryonic erythropoiesis , but the underlying mechanism remains poorly understood [4] . We examined whether Uba5 exerts its function in similar cellular pathways in which UFBP1 is involved . As shown in Fig 7A and 7B , knockdown of either Uba5 or UFBP1 in erythroleukemia K562 cells caused the increase of eIF2α phosphorylation and Xbp-1 mRNA processing , indicating activation of the UPR in these knockdown cells . This result suggests that Uba5 is also involved in regulation of ER homeostasis . Furthermore , depletion of either Uba5 or UFBP1 down-regulated key erythroid transcription factors GATA-1 and KLF1 in K562 cells ( Fig 7C ) , suggesting a lineage-specific role of ufmylation in erythropoiesis . Together , our results provide the first evidence suggesting that the Ufm1 system as a whole have pleiotropic roles in multiple cellular processes including maintenance of ER homeostasis and transcriptional regulation of erythroid development . ASC1 , a co-activator of hormone receptors , is a newly identified Ufm1 target [8] . Its ufmylation promotes recruitment of other transcription factors and thereby enhances estrogen receptor-mediated transcription [8] . Interestingly , in contrast to Uba5 and UFBP1 , depletion of ASC1 in K562 cells did not promote Xbp-1 mRNA splicing and Grp78 up-regulation , indicating that ASC1 knockdown does not cause elevation of basal ER stress ( Fig 8A and 8B ) . However , ASC1 knockdown led to significant under-expression of GATA-1 and KLF1 , suggesting a crucial role of ASC1 in regulation of erythroid transcription program ( Fig 8C and 8D ) . We further examined the effect of UFBP1 and Uba5 knockdown on ASC1 . Knockdown of either UFBP1 or Uba5 in K562 cells did not change ASC1 protein level ( Fig 8E ) , but caused the reduction of its nuclear presence ( Fig 8E ) . Given that ASC1 functions as a transcription co-activator , we speculated a direct role of ASC1 in transcriptional regulation of erythroid genes such as Gata-1 and Klf1 . To test this possibility , we performed chromatin-immunoprecipitation ( ChIP ) assay using UFBP1-depleted BM and spleen cells to examine whether ASC1 is associated with Gata-1 and Klf1 promoters . Depletion of UFBP1 did not alter ASC1 expression in UFBP1 deficient bone marrow cells ( S5A Fig ) . Both c-Myc and CCND1 ( cyclin D1 ) were reported to be ASC1 targets in breast cancer cells [8] . In agreement with the previous report , ASC1 was indeed associated with the promoters of c-Myc and CCND1 , and more importantly , the associations were significantly decreased in UFBP1 deficient BM cells ( Fig 8F ) . As a negative control , the beta-actin promoter sequence was slightly enriched in ASC1-DNA complex but the enrichment was not affected by UFBP1 depletion ( Fig 8F ) . Intriguingly , both Gata-1 and Klf1 promoter sequences were substantially enriched in ASC1-DNA complex , and the enrichment was significantly diminished by UFBP1 deficiency ( Fig 8F ) . Similar result was obtained using UFBP1 deficient spleen cells ( S5B Fig ) . Together , our results strongly suggest that ASC1 is one of the key downstream effectors of the Ufm1 system to transcriptionally regulate erythroid lineage development .
In this study , we report that UFBP1 , a key component of the Ufm1 conjugation system , is essential for murine erythropoiesis in both embryonic and adult stages . Germ-line inactivation of UFBP1 caused anemia and embryonic lethality around E12 . 5 ( Figs 1 and 2 ) , whereas induced ablation of UFBP1 in adult mice impaired hematopoiesis , resulting in severe pancytopenia and animal death ( Figs 3 and 4 ) . We found that defective hematopoiesis in UFBP1 deficient mice was in part caused by sustained ER stress-induced cell death of HSCs ( Figs 5 and 6 ) . Furthermore , ablation of ufmylation capacity , including depletion of Uba5 ( E1 ) , UFBP1 ( E3 co-activator ) or ASC1 ( Ufm1 target ) , significantly down-regulated expression of erythroid transcription factors ( Figs 7 and 8 ) . Interestingly , depletion of Ufm1 target ASC1 caused under-expression of GATA-1 and Klf1 , but not elevation of ER stress ( Fig 7 ) . Taken together , our findings strongly suggest that the Ufm1 system plays crucial pleiotropic roles in regulating both cell survival and differentiation during hematopoiesis . Accordingly , we propose a working model for ufmylation-mediated regulation of hematopoiesis: its activity in maintaining ER homeostasis is essential for survival of hematopoietic stem and progenitor cells , whereas ufmylation of ASC1 is crucial for high-level expression of erythroid transcription factors and erythroid lineage development ( Fig 8G ) Previous studies have shown that UFBP1 is a putative Ufm1 target , and its ufmylation is promoted by Ufm1 E3 ligase Ufl1 [2] . Furthermore , Ufm1-modified UFBP1 is required for ufmylation of another Ufm1 target ASC1 [8] , indicating that UFBP1 may function as a co-factor of Ufm1 E3 ligase . Interestingly , we found that UFBP1 KO mice shared extensive phenotypic similarities with Ufl1 and Uba5 KO mice in hematopoietic system . First , ablation of either gene impaired embryonic erythropoiesis and caused embryonic lethality around E10 . 5-E12 . 5 , even though death of UFBP1 null embryos appeared to occur slightly delayed compared to Ufl1 null embryos [4 , 9] . This delayed death of UFBP1 null embryos may be attributed to the fact that Ufl1 also regulates C53 protein that is essential for embryogenesis ( Fig 1 and [6 , 12 , 13] ) . Second , a high number of multi-nucleated erythrocytes were found is all KO embryos ( Fig 2 and [4 , 9] ) . Third , UFBP1 and Ufl1 KO mice exhibited nearly identical defects in erythroid development ( Fig 4 and [9] ) . Together , these phenotypic similarities strongly suggest that UFBP1 is indeed a key downstream effector of the Ufm1 system in ufmylation-mediated regulation of hematopoietic development . Interestingly , a genome-wide association study ( GWAS ) study has identified one SNP ( rs11697186 ) in DDRGK1 ( same as UFBP1 ) gene that is strongly associated with anemia and thrombocytopenia in peggylated interferon and ribavirin therapy for chronic hepatitis C patients , which provides clinical evidence for possible involvement of UFBP1 in regulation of erythropoiesis [14] . Nevertheless , it is also worth noting that UFBP1 deficient mice do have a subtle phenotypic difference compared to Ufl1 KO mice . Autophagic flux appeared to be normal in UFBP1 deficient cells ( S4 Fig ) , whereas Ufl1 deficiency caused inhibition of autophagic degradation [9] . It has been shown that Ufl1 knockdown results in significant depletion of both C53 and UFBP1 proteins [6 , 12] . Our result indicates that depletion of UFBP1 protein may not be the causative factor for blockage of autophagic flux in Ufl1 deficient cells . Similar to Ufl1 deficient cells , UFBP1-depeted LSK cells displayed elevated ER stress that in turn caused activation of the UPR and cell death program ( Fig 6 ) . ER homeostasis is crucial for survival and function of HSC and other progenitor cells [15] . The UPR is a highly coordinated program that facilitates protein folding , processing , export and degradation of proteins during cellular response to ER stress . It generally includes three signaling branches: IRE1 ( Inositol requiring enzyme 1 ) , PERK ( PKR-like ER kinase ) and ATF6 [16] . Although the UPR serves as a cytoprotective role to restore ER homeostasis , persistent activation of IRE1 and PERK signaling caused by unresolved ER stress also leads to cell death and tissue damage [16] . It has been reported that HSCs and other progenitor cells exhibit distinct cellular response to extracellular ER stressors , and the PERK-eIF2α-CHOP pathway predisposes human HSCs to ER stress-induced apoptosis [15] . Furthermore , overexpression of ER chaperone ERdj4 and RNA-binding protein Dppa5/Esg1 in HSCs reduced ER stress and thereby improved HSC survival and stemness function [15 , 17] . Interestingly , we found that depletion of either UFBP1 or Uba5 led to elevated ER stress and activation of the UPR , and sustained ER stress may contribute to cell death of HSCs ( Figs 6 and 7 ) . This result is line with the previous studies showing that the Ufm1 system protects pancreatic beta cells and macrophages from ER stress-induced apoptosis [7 , 11] . Our finding suggests that the Ufm1 system plays a cytoprotective role by maintaining ER homeostasis of HSCs . Therefore , manipulation of ufmylation activity may represent a novel approach to improve survival and function of HSCs . In addition to their role in maintaining ER homeostasis that appears to be a general function in most types of cells we examined , UFBP1 and Ufl1 also possess a lineage-specific activity , which was evidenced by the complete blockage of differentiation of CFU-Es to proerythroblasts in both UFBP1 and Ufl1 deficient bone morrows ( Fig 4A and [9] ) . This lineage-specific effect was also observed in Uba5 deficient embryos , in which development of MEP ( megakaryocyte erythroid progenitors ) , but not GMPs ( granulocyte macrophage progenitors ) , was impaired in fetal livers [4] . Yet , the molecular basis of this erythroid-specific activity remains largely unknown . In this study , we found that ASC1 , a transcription co-activator and Ufm1 target , was the key downstream effector to mediate this lineage-specific activity . ASC1 was originally identified as thyroid hormone receptor interactor 4 ( TRIP4 ) [18 , 19] . As a transcriptional co-activator , ASC1 enhances transcription by recruiting nuclear receptors ( such as ERα ) , transcriptional cofactors ( p300 and SRC1 ) , and basic transcriptional machinery ( TBP and TFIIA ) to specific promoters , and its ufmylation provides a platform for the recruitment of these factors [8] . Interestingly , it has been known that nuclear receptor signaling pathways play a crucial role in both normal and stress erythropoiesis . Altering the content of thyroid hormone receptor α ( TRα ) in chicken erythroblasts affects the balance between proliferation and differentiation [20] . TRα-/- mice display defective fetal and adult erythropoiesis and an altered stress erythropoiesis response to hemolytic anemia [21 , 22] . Moreover , glucocorticoid receptor ( GR ) signaling is important for stress erythropoiesis [23] , while retinoic acid ( RA ) has a broad impact on hematopoiesis [24] . Estrogen receptor was also reported to modulate HSC survival and erythropoiesis [25] . Therefore , it is quite plausible that ASC1 and its ufmylation may be directly involved in regulation of erythroid development . To support our hypothesis , we found that knockdown of ASC1 in K562 down-regulated key erythroid transcription factors GATA-1 and KLF1 ( Fig 8C ) . More importantly , ChIP analysis showed that ASC1 was associated with the promoters of GATA-1 and Klf1 , and this association was significantly attenuated by depletion of UFBP1 ( Fig 8E ) , suggesting that ufmylation of ASC1 is important for its association with the promoters of GATA-1 and Klf1 . Further studies are needed to identify and characterize both extracellular and intracellular signals leading to ASC1 ufmylation and its binding partners , and to determine the molecular mechanism of ASC1-mediated regulation of erythroid development . In addition , our results also indicate an important role of UFBP1 in development of other lineages such as myeloid cells . Tissue- and lineage-specific genetic models will be generated for thorough elucidation of its function and working mechanism in the hematopoietic system .
Our animal study was conducted according to the guidelines of Animal Research: Reporting In Vivo Experiments ( ARRIVE ) . IACUC of Georgia Regents University approved this study ( protocol #2011–0314 ) K562 cells were cultured in RPMI 1640 medium supplemented with 10% FBS and antibiotics . Tamoxifen and 4-hydroxytamoxifen and other chemical reagents were purchased from Sigma ( St . Louis , MO , USA ) . ES cell clone HEPD0618_2_D02 ( JM8A3 . N1 with C57BL/6N background ) containing trapped UFBP1 allele was purchased from the EUCOMM ( European Conditional Mouse Mutagenesis ) team . The ES cells were injected into the blastocysts of C57BL/6 mice ( Northwestern University Transgenic and Targeted Mutagenesis Laboratory ) . Chimeric mice were crossed with B6 ( Cg ) -TyrC-2J/J albino mice , and heterozygous offspring with germ-line transmission were confirmed by genotyping . To generate CKO mice , we crossed UFBP1Trap-F/+ mice with FLPo deleter mice ( B6 ( C3 ) -Tg ( Pgk1-FLPo ) 10Sykr/J , The Jackson Laboratory , Bar harbor , ME ) to remove the gene trap cassette . The floxed UFBP1 mice were crossed with ROSA26-CreERT2 mice ( B6 . 129-Gt ( ROSA ) 26Sor<tm1 ( cre/ERT2 ) Tyj>/J , The Jackson Laboratory ) in which CreERT2 was inserted into ROSA26 locus . Cre-mediated deletion of UFBP1 was induced by tamoxifen administration . Tamoxifen ( 20 mg/ml in corn oil , Sigma , St . Louis , MO ) was administrated by 5-day IP injection with an approximate dose of 75 mg tamoxifen/kg body weight . All animal procedures were approved by IACUC of Georgia Regents University . The following primers were used for PCR genotyping of UFBP1 KO embryos and mice: P1 ( TAGTACTTGAAGTCTGGCTTGGTA ) , P2 ( CACAACGGGTTCTTCTGTTAGT CC ) and P3: ( TAGTCAGGAACTGATGA GTGTCTC ) . A 35-cycle ( 92°C , 45 sec , 62°C , 45 sec , 72°C 45 sec ) PCR was performed in the present of 1% DMSO , and the determination of genotypes was described in Fig 1 . For the floxed UFBP1 allele , P1 and P3 primers were used in PCR genotyping . The floxed allele generates a 637 bp PCR product comparing to 440 bp WT product . Genotyping of Cre-ERT2 mice was performed according to the standard protocol of the Jackson Laboratory . Fetal liver cells ( 5 , 000 cells ) were plated in one ml of serum-free methylcellulose and IMDM ( M3234 , Stem Cell Technology , Vancouver , BC , Canada ) supplemented with 10% fetal bovine serum ( FBS ) . For CFU-Es , cells were cultured for 3 days in the presence of hEpo ( 2 units/ml , Epogen , Amgen , Thousand Oaks , CA ) and stained with benzidine dihydrochloride , and the number of colonies is scored . For BFU-Es , cells were cultured in the presence of hEpo ( 2 units/ml ) and mSCF ( 100 ng/ml , Peprotech , Rocky Hill , NJ ) for 7 days , stained and scored . For CFU-GMs , CFU-GEMMs , BM or fetal liver cells ( 20 , 000 ) were cultured in M3434 ( Stem Cell Technology ) for 5 days , and the numbers of the colonies were manually counted . For EryP-CFCs , yolk sac cells from E8 . 5 embryos was cultured in IMDM medium containing 10% plasma-derived serum , 5% PFHM-II , 0 . 025 mg/ml ascorbic acid , 4 . 5 x 10−4 M MTG , Epo ( 2 U/ml ) , and colonies were manually counted after 96 hours . Complete blood counts ( CBC ) of blood samples from adult mice were analyzed by in-house Horiba ABX Micro S60 analyzer ( HORIBA Medical , Irvine , CA , USA ) . Recipient mice ( CD45 . 1 ) were first irradiated with dose of 9 Gy . BM cells from UFBP1F/F and UFBP1F/F;ROSA-CreERT2 mice ( CD45 . 2 ) were isolated , mixed with BM cells from CD45 . 1 mice in a 1:1 ratio ( 1x106 cells total ) , and subsequently injected into the retro-orbital venous sinus of anesthetized recipient mice ( 5 mice for each group ) . After a 4-week recovery period , tamoxifen was administered in 5 consecutive days . Donor cell engraftment was monitored by flow cytometry of the blood samples from tail veins in a 2-week interval using lineage markers . After 3 weeks of tamoxifen injection , BM cells were subjected to flow analysis . Sorted BM HSC and myeloerythroid progenitor cells were cultured in IMDM supplemented with 10% fetal bovine serum , hEpo ( 2 units/ml , Amgen , Thousand Oaks , CA ) , mSCF ( 100 ng/ml , Peprotech , Rocky Hill , NJ ) , mIL-3 ( 10 ng/ml , Peprotech ) , Dexamethasone ( 1 μM , Sigma , St . Louis , MO , USA ) and IGF-1 ( 100 ng/ml , Invitrogen , Grand Island , NY ) . Deletion of UFBP1 was induced by addition of 4-hydroxytamoxifen ( 1 μM , Sigma ) . Freshly isolated BM cells were suspended in PBS with 1% fetal bovine serum and stained with BV510-Sca-1 ( Biolegend , San Diego , CA ) , APC780-c-Kit ( eBioscience , San Diego , CA ) , PE-Cy7-CD150 ( Biolegend ) , Alexa700-CD16/32 ( eBioscience ) , PE-IL-7R ( BD Biosciences ) , APC-CD41 ( BD Biosciences ) , BV650-CD105 ( Biolegend ) , FITC-CD71 ( BD Biosciences ) , BV421-TER119 ( Biolegend ) , and PerCP-Cy5 . 5 conjugated lineage markers including CD4 , CD8 , CD3 , CD5 , Gr-1 , CD11b , CD19 and B220 ( BD Biosciences ) . The samples were analyzed using BD LSR II SORP and Diva 7 . 0 software ( GRU Cancer Center Flow Cytometry Facility ) . For competitive repopulation assays , BV421-CD45 . 2 ( Biolegend ) was used to substitute BV421-TER119 . Cell sorting of HSCs and progenitor cells was performed using BD FACSAria II SORP . Total RNA from each sample was isolated with the GeneJET RNA Purification Kit ( Thermo Scientific ) , and then reversely transcribed using the SuperScript First-Strand Synthesis System ( Invitrogen ) according to the manufacturer’s instruction . RT-PCR was performed using the iTaq Universal SYBR Green Supermix kit ( BIO-RAD ) with 40 cycles of 95°C for 15 seconds and 60°C for 1 minute on StepOnePlus Real-Time PCR System ( Life Technologies ) . The results were analyzed by StepOne Software ( Version 2 . 1 , Life Technologies ) . Relative expression level of each transcript was normalized to murine beta-actin and GAPDH by using the 2^ ( -delta delta Ct ) method . The following is the list of primers used in this study: Genes Primer sequence Mouse Actin GACCTCTATGCCAACACAGT AGTACTTGCGCTCAGGAGGA Mouse UFBP-1 GAAGCCAGCAGAAGTTCACC GAAGCCGTTCCTCTTCCTTC Mouse Grp78 ACTTGGGGACCACCTATTCCT ATCGCCAATCAGACGCTCC Mouse ERdj4 TAAAAGCCCTGATGCTGAAGC TCCGACTATTGGCATCCGA Mouse CHOP GCATGAAGGAGAAGGAGCAG ATGGTGCTGGGTACACTTCC Mouse GADD34 AGAACATCAAGCCACGGAAG TCTCAGGTCCTCCTTCCTCA Mouse Bax TGCAGAGGATGATTGCTGAC AAGATGCTGTTGGGTTCCAG Mouse Bak AAAATGGCATCTGGACAAGG AAGATGCTGTTGGGTTCCAG Mouse Bim TGCAGAGGATGATTGCTGAC GATCAGCTCGGGCACTTTAG Mouse Noxa GGCAGAGCTACCACCTGAGT TTGAGCACACTCGTCCTTCA Mouse Puma GCCCAGCAGCACTTAGAGTC TGTCGATGCTGCTCTTCTTG Mouse DR5 TGACTACACCAGCCATTCCA AGTTCCTCTTCCCCGTCAGT Human Actin CCTGTACGCCAACACAGTGC ATACTCCTGCTTGCTGATCC Human Gata1 GATGGAATCCAGACGAGGAA GCCCTGACAGTACCACAGGT Human Klf1 CACGCACACGGGAGAGAAG CGTCAGTTCGTCTGAGCGAG Human UFBP-1 GAAGCCAGCAGAAGTTCACC GAAGCCGTTCCTCTTCCTTC Human Grp78 ACTTGGGGACCACCTATTCCT ATCGCCAATCAGACGCTCC Human ASC1 AGTGGGTTGACCACACAGGT CTGCCCTCCACCCTTTTAAT For ASC1 ChIP assay , UFBP1F/F ( control ) and UFBP1F/F;CreERT2 mice were injected with tamoxifen for 5 consecutive days . At day 4 , mice were injected with a single dose of phenylhydrazine ( 40 mg/kg ) . Bone marrow and spleen cells were harvested at day 8 , and subjected to ChIP and quantitative PCR analysis as previously described [26 , 27] . The following primers were used for detection of specific promoters: mKLF1-chip , AGCGACCAGGACTTTTCTTT ( F ) and CGATATGTGTGT GGTGTGCT ( R ) ; mGATA1-chip , CCAACTCCTGCAAGAACA GT ( F ) and GTTGACACTA GCCCCACAGT ( R ) . mActin-chip , TCTCCAAAAGTGCCTGACTC ( F ) and CCCTTCTGCTGTGGTTCTAA ( R ) ; mMyc-chip , CGACCAAAGGCAAAATACAC ( F ) and CCCTGCGTATATCAGTCACC ( R ) ; mCCND1-chip , CTCCCCCAACTCAAGACTG ( F ) and AATTCCAGCAACAGCTCAAG ( R ) . Xbp-1 mRNA splicing assay was conducted as described previously [26] , using the following primers: for human Xbp-1: 5’ GGAGTTAAGACAGCGC TTGG 3’ ( F ) and 5’ ACTGGGTCCAAGTTGTCCAG 3’ ( R ) ; and for murine Xbp-1 , 5’ ACACGCTTG GGAAT GACAC 3’ ( F ) and 5’ CCATGGGAAGATGTTCTGGG 3’ ( R ) . Lentiviral vectors expressing specific shRNAs were constructed using pLKO . 1 vector . Lentiviruses were prepared using 293FT packaging cell line according to the manufacturer's instruction ( Invitrogen ) . The following shRNAs were used: hUba5 shRNA: CCTCAGTGTGATGACAGAAAT; hUFBP1 shRNA#1: GTGTCGAGAAGCCAGCGGAAA; hUFBP1 shRNA#2: GGAGCTAAGAAACTGCGGAAG; hASC1 shRNA#1: GCAGAGTATCATAGCAGACTA; hASC1 shRNA#2: GGACTAGAGTTCAACTCATTT . For knockdown assays , K562 cells were infected with lentiviruses expressing either scrambled or gene-specific shRNAs for 24 hours , and then selected with puromycin ( 1 . 5 μg/ml ) and culture for 3 to 4 days . Knockdown efficiency was evaluated by either immunoblotting or quantitative RT-PCR . Immunofluorescent staining and Immunoblottings were performed as described previously6 . Confocal images were acquired using Zeiss 510 META confocal microscope with Zen software ( Carl Zeiss Microscopy GmbH , Jena , Germany ) , while epifluorescence images were obtained using Zeiss Observer D1 with AxioVision 4 . 8 software ( Carl Zeiss Microscopy GmbH , Jena , Germany ) . The antibodies used in this study include: UFBP1 and Uba5 rat antibody ( Li lab and [6 , 26] ) , eIF2α , phospho-S51-eIF2α ( Cell Signaling , Danvers , MA , USA ) , ASC1 ( A300-843A , Bethyl Laboratories , Montgomery , TX , USA ) , JNK and phospho-JNK ( Santa Cruz Biotechnology Inc . , Dallas , TX , USA ) , GATA-1 and KLF1 ( Abcam , Cambridge , MA , USA ) , α-tubulin and β-actin ( Sigma ) . All affinity-purified and species-specific HRP- and fluorophore-conjugated secondary antibodies were obtained from Jackson ImmunoResearch ( West Grove , PA , USA ) . TUNEL staining was performed using In situ cell death detection kit ( Fluorescein , Roche , Indianapolis , IN , USA ) according to the manufacturer’s instruction . For hematoxylin and eosin ( H & E ) staining of embryo sections , the embryos were fixed in 10% formalin overnight , rinsed with PBS and embedded in paraffin . Embryo sections were deparaffinized and hydrated , and subsequently stained with hematoxylin solution for 6 minutes . After rinse in water for 15 minutes , the sections were stained with eosin-Y solution for 3 minutes . The sections were then rinsed in water , dehydrated and mounted with Permount . | Protein modification by Ubiquitin ( Ub ) and Ubiquitin-like proteins ( Ubl ) plays pivotal roles in a wide range of cellular functions and signaling pathways . The Ufm1 conjugation system is a novel ubiquitin-like system , yet its biological functions and working mechanism remains poorly understood . UFBP1 is a putative Ufm1 target that has been implicated in several signaling pathways but little is known regarding its in vivo function . In this report , by using multiple knockout mouse models , we demonstrate that UFBP1 is essential for murine development and blood cell development . While germ-line deletion of UFBP1 caused defective red blood cell development and embryonic lethality , somatic ablation of UFBP1 impaired production of mature red blood cells and other types of hematopoietic cells . We found that depletion of UFBP1 led to elevated stress in the endoplasmic reticulum that in turn caused cell death of hematopoietic stem cells . Furthermore , UFBP1 deficiency diminished expression of key transcription factors essential for red blood cell development . Taken together , our study provides strong genetic evidence for the essential role of UFBP1 as well as other components of the Ufm1 system in hematopoietic development . Therefore , the ufmylation pathway may represent a novel therapeutic target in treatment of blood diseases . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| []
| 2015 | UFBP1, a Key Component of the Ufm1 Conjugation System, Is Essential for Ufmylation-Mediated Regulation of Erythroid Development |
Previous studies have shown that EBLV-1 strains exclusively hosted by Eptesicus isabellinus bats in the Iberian Peninsula cluster in a specific monophyletic group that is related to the EBLV-1b lineage found in the rest of Europe . More recently , enhanced passive surveillance has allowed the detection of the first EBLV-1 strains associated to Eptesicus serotinus south of the Pyrenees . The aim of this study is the reconstruction of the EBLV-1 phylogeny and phylodynamics in the Iberian Peninsula in the context of the European continent . We have sequenced 23 EBLV-1 strains detected on nine E . serotinus and 14 E . isabellinus . Phylogenetic analyses were performed on the first 400-bp-5’ fragment of the Nucleoprotein ( N ) gene together with other 162 sequences from Europe . Besides , fragments of the variable region of the phosphoprotein ( P ) gene and the glycoprotein-polymerase ( G-L ) intergenic region were studied on Spanish samples . Phylogenies show that two of the new EBLV-1a strains from Iberian E . serotinus clustered together with French strains from the North of the Pyrenees , suggesting a recent expansion southwards of this subtype . The remaining seven Iberian strains from E . serotinus grouped , instead , within the cluster linked , so far , to E . isabellinus , indicating that spatial distribution prevails over species specificity in explaining rabies distribution and supporting interspecific transmission . The structure found within the Iberian Peninsula for EBLV-1b is in concordance with that described previously for E . isabellinus . Finally , we have found that the current EBLV-1 European strains could have emerged only 175 years ago according to our evolutionary dynamics analyses .
Rabies is caused by viruses of the genus Lyssavirus , which includes so far fourteen species recognized by the International Committee on Taxonomy of Viruses ( ICTV ) , plus the recently proposed Lleida bat Lyssavirus ( LLEBV ) [1] and Gannoruwa bat lyssavirus ( GBLV ) [2] . Three of the viruses of the phylogroup 1 are associated with European bats of the family Vespertilionidae: the European bat Lyssaviruses 1 and 2 ( EBLV-1 and EBLV-2 ) and the Bokeloh bat Lyssavirus ( BBLV ) . Besides , the other two lyssaviruses described so far in Europe ( West Caucasian bat virus , WCBV; and LLEBV ) are hosted by the cave bat Miniopterus schreibersii ( Family Miniopteridae ) [1 , 3] . Out of all of them , only EBLV-1 and EBLV-2 have caused rabies in humans , while more than 90% of the bat rabies cases have been reported from serotine bats ( E . serotinus ) infected by EBLV-1 [4] . EBLV-1 is divided in two main subtypes: EBLV-1a and EBLV-1b . The first one is found in an East-west axis from Ukraine to the north of France , while EBLV-1b is reported from France , southern Germany and the Netherlands [5] . EBLV-1a has been recently described also from Southern France [6] . Nevertheless , the evolutionary relationships among these viral subtypes in Western Europe are only partially known and previous attempts of evolutionary analyses [7] were hampered either by uneven available sampling or the dearth of molecular markers . Two cryptic species of serotine bats ( E . serotinus and E . isabellinus ) have been lately found within the genus Eptesicus in the Iberian Peninsula [8] . While E . serotinus is distributed across Northern Iberia as well as the rest of Western Europe , the sibling species . E . isabellinus is restricted to the Southern half of the Iberian Peninsula and Northern Africa [9] . So far , EBLV-1 infection has been declared only from E . isabellinus in Spain [7] , forming an exclusive monophyletic clade , possibly related to the EBLV-1b subtype . In this study , we described the first EBLV-1 associated to E . serotinus South of the Pyrenees , and show how the inclusion of these new strains in phylogenetic and phylodynamic analyses together with the use of newly developed molecular markers have substantially improved our present knowledge of EBLV-1 molecular epidemiology in Europe and the Iberian Peninsula .
No live animals were used for this study . All the work has been done with bat carcasses from , either animals directly submitted to the Laboratory by public health services , or admitted in wildlife care Centers , which submitted to the laboratory those bats which could not be recovered after treatment . Consequently , the need for approval by an IACUC/ethics committee does not apply . We have sequenced three fragments of the viral genome of 23 EBLV-1 strains detected in brains of nine E . serotinus and 14 E . isabellinus sent to the National Center for Microbiology ( Majadahonda , Spain ) for rabies diagnosis and that were found positive for Lyssavirus antigen by the fluorescence antibody test and real time polymerase chain reaction ( RT-PCR ) [10] . Morphological bat identification was confirmed in most cases by genomic sequencing of several diagnostic mtDNA fragments [9]: 69R99 ( MG211681 ) , 80R99 ( MG211682 ) , 56R00 ( MG211683 ) , 69R00Eis ( MG211684 ) , 211R07 ( MG211685 ) , 86R08 ( MG211686 ) , 2011Riglos ( MG211687 ) , 201127004 ( MG211688 ) , 201149008 ( MG211689 ) , 201238163 ( MG211690 ) , 201325895 ( MG211691 ) , 200928458 ( MG211692 ) , 201544034 ( MG211693 ) , 201548093 ( MG211694 ) , 201539228 ( MG211695 ) , 201427094 ( MG211696 ) , 201539226 ( MG211697 ) , 44R02 ( MG211698 ) , 292R07 ( MG211699 ) . The 400-bp-5' terminal sequence of the nucleoprotein ( N-400 ) gene was amplified and sequenced as described previously [7] . Besides , a 686-bp variable region of the phosphoprotein ( P ) and a 763-bp fragment including the glycoprotein-polymerase ( G-L ) intergenic region were amplified by EBLV-1 specific primers in nested RT-PCR reactions specific for each region ( Table 1 ) . All 23 N-400 sequences from Spain were aligned using the MAFFT software ( http://mafft . cbrc . jp/alignment/software/ ) together with 162 sequences from other European countries available on GenBank ( S1 Table ) . The fittest nucleotide substitution model was selected according to the Bayesian information criterion ( BIC ) using jModelTest v . 2 . 1 . 10 ( http://darwin . uvigo . es/our-software/ ) . For the Iberian samples we selected the substitution models HKY , K80 + I , K80 for the alignments of the P , G-L and N400 markers respectively . Besides , a GTR model was selected for the alignment of all Iberian and European N400 fragments . Only for the Iberian samples , the N , P and G-L fragments were concatenated into a 1809 base pair ( bp ) alignment using Sea View v . 4 . 6 . ( Table 1 ) and a HKY nucleotide substitution model was selected from this concatenated alignment using Partition Finder v . 1 . 1 . 1 . ( http://www . robertlanfear . com/partitionfinder/ ) . Phylogenetic relationships were reconstructed for each alignment using a Bayesian approach with the program MrBayes ( version 3 . 1 . 2 ) with two runs of 1 x 106 generations . Additional phylogenies were reconstructed using other three optimality criteria: Minimum Evolution ( ME ) , Maximum Likelihood ( ML ) and Maximum Parsimony ( MP ) . ME was obtained through a Neighbor-Joining algorithm with MEGA7 ( http://www . megasoftware . net/ ) , ML was constructed with PHYML software ( http://www . atgc-montpellier . fr/phyml/ ) and finally , MP trees were inferred with PAUP* 4 . 0b10 ( http://paup . csit . fsu . edu/ ) weighting transversions differentially according to the transitions/transversion ratio of the selected substitution model . Confidence in the topologies was inferred after 1 , 000 bootstrap replicates . Trees were edited using the FigTree v1 . 4 . 1 program ( http://tree . bio . ed . ac . uk/software/figtree/ ) . Viral genetic geographic structure across Europe was examined for each lineage data set , EBLV-1a and EBLV-1b , through parsimony-based haplotype networks employing the Median Joining ( MJ ) algorithm implemented in Network ( v5 . 0 , Fluxus Technology ( http://www . fluxus-engineering . com/sharenet . htm ) , without star-contraction pre-processing or post-processing clean-up options . To reconstruct phylodynamic patterns of EBLV-1 we analysed the N-400 sequence dataset in a Bayesian framework implemented in the BEAST software ( Bayesian Evolutionary Analysis Sampling Trees ( http://beast . bio . ed . ac . uk/ ) [11] . For model parameters we follow Hughes 2008’s analyses that use similar sequences of the same marker and virus . Accordingly , evolution rates were assumed to fit an uncorrelated lognormal molecular clock [12] and two MCMC chains were run for 5 x 107 generations each with trees and parameters sampled every 1000 steps . We used the SRD06 option for codon partition and substitution model according to the recommendation of Hughes 2008 . This option includes an HKY substitution model and four gamma rates with two codon partitions for the alignment: first and second codons were analysed together and separately from the third codon . Analyses started with a random starting tree with a constant population size prior , according to Hughes 2008 . Then , chains were run to an effective sample size ( ESS ) of parameters higher than 200 ( following BEAST’s recommendations ) and chains convergence was assessed from standard deviation and likelihood values using Tracer ( http://tree . bio . ed . ac . uk/software/tracer/ ) with the first 10% of the trees discarded as burning . Bayesian credible intervals or 95% Highest Posterior Density ( 95% HDP ) were inferred as uncertainty evaluation . Finally , the consensus phylogenetic tree was built with TreeAnnotator ( BEAST software package ) and edited in FigTree ( http://tree . bio . ed . ac . uk/software/figtree/ ) .
EBLV-1 is mostly found associated with the bat E . serotinus north of the high mountain ranges of Southern Europe , such as the Alps or the Pyrenees , which seem to hinder virus expansion to the south . However , a group of sequences from Southern Spain were found to cluster in a monophyletic group associated with a different bat , the sibling species E . isabellinus [7] . In this study , we report for the first time nine EBLV-1 strains found in E . serotinus south of the Pyrenees and describe their evolutionary relationships ( Fig 1A and 1B ) . Two of them grouped within the recently described cluster of EBLV-1a sequences from Southern France [6] , extending this group now to Northern Spain . The EBLV-1a subtype is otherwise typically distributed through Northern Europe , along an east-west axis from The Netherlands to Ukraine [5] . The shallow genetic differentiation ( Figs 1B and 2A ) found between the strains on both slopes of the Pyrenees of this Southern group of EBLV-1a suggests a recent geographical expansion of this subtype across southern France , with very recent arrival to the Iberian Peninsula . On the other hand , one of the new EBLV-1a strains was found only 24 km away from the nearest Iberian EBLV-1b strain ( Fig 1A ) , with no significant geographical barriers between them . This supports a hypothesis of current southwards expansion of EBLV-1a , as was proposed for Western Europe by Vázquez-Morón et al . [7] . The clear star-like structure observed for EBLV-1a in the haplotype network provides additional support for this hypothesis ( Fig 2A ) . The remaining seven new EBLV-1 sequences detected in E . serotinus cluster together with the Iberian clade that had been exclusively found in E . isabellinus in the Southern half of the Iberian Peninsula [7] ( Fig 1B ) . This Iberian clade was considered to be differentiated from both EBLV-1a and EBLV-1b [7] . However , our new phylogenetic analyses show that this Iberian lineage shared by the two species of Eptesicus is actually part of the EBLV-1b subtype . This fact points again to a more predominant role of geographical factors than the host species in determining EBLV-1’s phylogenetic structure in Iberia . Both Eptesicus species are phylogenetically close and show some overlap in their geographical distribution in Iberia [13] . Interestingly , phylogenetic relatedness and geographic contact , were considered the two main conditions for interspecific transmission of rabies among bats in America [14] . The haplotype network ( Fig 2B ) shows a much more complex structure and higher haplotype diversity for EBLV-1b than for EBLV-1a , suggesting a much longer evolutionary history . The analyses also suggest that Iberian EBLV-1b has evolved isolated from the rest of European EBLV-1b , from a still unknown common ancestor , although more geographic coverage is still necessary to fully reconstruct this event . Our coalescence based analysis ( S3 Fig ) suggests that the Most Recent Common Ancestor ( tMRCA ) for the entire EBLV-1 lineage currently circulating in Eptesicus bats dates to 175 years ago ( [47; 260] HDP ) . This value is remarkably similar ( just five years older ) to the date previously proposed [12] , but with narrower 95% HDP intervals . The tMRCA estimated for the whole EBLV-1a lineage was approximately 70 years ( [59; 158] HDP ) , while the Iberian EBLV-1a strains could have emerged less than 15 years ago , ( [4; 13] HDP ) giving additional support to the hypothesis of a recent southward expansion from France across the Pyrenees . On the other hand , the tMRCA for the subtype EBLV-1b is estimated be around 110 years ago , ( [53; 187] HDP ) . This result suggests a longer evolutionary history than EBLV-1a in agreement with the Network analysis . The Iberian EBLV-1b strains would have spawned approximately 55 years ago , ( [31; 82] HDP ) . Nevertheless , we suggest caution on this dating since some uncertainty for the marginal posterior distribution of all the tMRCA was noticed , which will , most probably , be solved when additional cases of Iberian EBLV-1 infection are available . In addition to the nucleoprotein gene , the two hypervariable regions studied enlighten the internal structure within the Iberian EBLV-1b cluster . The strains associated with E . serotinus seem particularly related to those associated with E . isabellinus from the South West ( Fig 3 ) . In fact , the presence of a strain ( 69R00_SE ) from Seville ( South West ) in the new group of viruses associated with E . serotinus ( North East ) points to an inter-specific transmission of EBLV-1 ( Fig 1A and Fig 3 ) and suggests that this particular Iberian lineage has extended northwards from E . isabellinus to E . serotinus . In fact , the Network shows the Northern Iberian EBLV-1b strains associated with E . serotinus branching off from a common Southern haplotype ( in an expansion star-like shape ) , providing additional support for this hypothesis ( Fig 2B ) . On the other hand , the Iberian South West seems to concentrate the highest variability within the Iberian EBLV-1b lineage ( Fig 1A ) , including a sub-lineage represented by two sequences ( 155R99 , R75 ) which occupied the most basal position and have never been detected after 1999 ( Fig 3 ) . An east-west discrimination of two groups of viruses in the South is in agreement with the east-west population subdivision described for the host bat E . isabellinus [9] . Interestingly , this study suggests a close genetic relatedness between South Western Iberian and North-African populations for E . isabellinus [9] . This genetic pattern strongly supports the possibility of the presence of EBLV-1 in North Africa from where it has not been reported yet . In conclusion , the increasing involvement of wild life rescue centers in bat rabies surveillance has allowed us to detect the first cases in Iberian E . serotinus and we hope it will allow us to increase the overall number of Iberian strains in the upcoming years . Unfortunately , active screening of healthy bats captured in colonies by testing oral fluids and blood does not usually provide enough genomic material to perform this kind of study [15] . The use of highly variable genomic regions ( Fig 3 , S3A and S3B Fig ) has clearly improved the resolution of the virus evolutionary phylogenetic reconstructions [16] , and has provided a better understanding of the phylogenetic structure of the Iberian EBLV-1 populations , pointing to evolutionary patterns at a continental level that deserve further research . | Rabies is caused by at least fourteen different viruses of the genus Lyssavirus . Although the classical rabies virus transmitted by the dog accounts for most human cases , most lyssaviruses are hosted by bats , which are able to transmit the disease to humans . The European bat lyssaviruses 1 ( EBLV-1 ) and 2 ( EBLV-2 ) have caused human rabies in Europe , while more than 90% of the bat rabies cases have been reported from serotine bats ( Eptesicus serotinus ) infected by EBLV-1 . These cases clearly accumulate in certain areas of Europe , but the factors driving this distribution are unknown . The evolutionary relationships among these viral strains provide valuable information , however , they are only partially known . The south of the Iberian Peninsula is the only location within Europe with an additional reservoir for EBLV-1 , the isabelline serotine bat ( Eptesicus isabellinus ) which is present also in North Africa . In this study we have reconstructed the EBLV-1 phylogeny and phylodynamics in the Iberian Peninsula in the context of the European continent . Our results suggest that some lineages have longer evolutionary history in their distribution areas , than others which seem to be in the process of a geographical expansion . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"and",
"discussion"
]
| [
"taxonomy",
"organismal",
"evolution",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"geographical",
"locations",
"microbiology",
"vertebrates",
"animals",
"mammals",
"genetic",
"mapping",
"viruses",
"phylogenetics",
"data",
"management",
"rna",
"viruses",
"phylogenetic",
"analysis",
"microbial",
"evolution",
"europe",
"rabies",
"virus",
"computer",
"and",
"information",
"sciences",
"medical",
"microbiology",
"microbial",
"pathogens",
"evolutionary",
"systematics",
"viral",
"evolution",
"people",
"and",
"places",
"lyssavirus",
"haplotypes",
"eukaryota",
"virology",
"viral",
"pathogens",
"bats",
"genetics",
"heredity",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"amniotes",
"organisms"
]
| 2018 | First cases of European bat lyssavirus type 1 in Iberian serotine bats: Implications for the molecular epidemiology of bat rabies in Europe |
Electrophysiological evidence suggested primarily the involvement of the middle temporal ( MT ) area in depth cue integration in macaques , as opposed to human imaging data pinpointing area V3B/kinetic occipital area ( V3B/KO ) . To clarify this conundrum , we decoded monkey functional MRI ( fMRI ) responses evoked by stimuli signaling near or far depths defined by binocular disparity , relative motion , and their combination , and we compared results with those from an identical experiment previously performed in humans . Responses in macaque area MT are more discriminable when two cues concurrently signal depth , and information provided by one cue is diagnostic of depth indicated by the other . This suggests that monkey area MT computes fusion of disparity and motion depth signals , exactly as shown for human area V3B/KO . Hence , these data reconcile previously reported discrepancies between depth processing in human and monkey by showing the involvement of the dorsal stream in depth cue integration using the same technique , despite the engagement of different regions .
Visual environments provide a range of cues that allow the brain to extract depth structure from the ambiguous images projected onto the two-dimensional ( 2D ) retinas . A fundamental challenge in visual neuroscience is to understand how this 2D information is processed and integrated , to allow the viewer to perceive and act in a three-dimensional ( 3D ) world . While multiple regions of the macaque [1] and human [2] brain have been found to respond to images that depict depth , it is only recently that we have begun to understand how information from different signals is fused together . While many studies demonstrated that regions of cortex could respond to information conveyed by two different cues ( such as depth from binocular disparity , and depth from motion ) , this alone does not imply that information is fused into a common representation . For instance , information from the two cues might be locally segregated within the cortex . Differentiating responses to fused versus independent signals requires careful assessment of neural responses to presentations of stimuli in which information from the two different cues is manipulated independently . Electrophysiological recordings from macaque middle temporal ( MT ) area have suggested representation of surface structures defined by combinations of binocular disparity and relative motion cues [3] . This is consistent with a large number of studies that have highlighted the importance of area MT in signalling structured motion information [4–6] and disparity signals [7–9] . Surprisingly , however , human imaging identified a different neural locus for the fusion of depth signals from disparity and motion [10 , 11] . Specifically , dorsal visual area V3B/kinetic occipital area ( V3B/KO ) [12 , 13] which is located more caudally relative to human MT , showed neuronal responses that match the predictions of cue fusion using multiple tests . What lies behind the apparent discrepancy between the neural locus of fusion in humans and macaques ? Is it a difference between species or is it merely technique related: functional MRI ( fMRI ) versus single unit electrophysiology ? It is known that putatively homologous areas in human and macaque ( i . e . , macaque MT and human MT+ ) may carry partially different functions [14–16] . It is an open question , however , whether this also holds for cue fusion mechanisms . Here , we aim to rule out differences induced by technique by using comparative brain imaging [5 , 17 , 18] , thereby exploiting stimuli , experimental designs , and analysis tools previously used in human studies [10] to test for responses in the macaque brain . Identifying neural responses for integrated depth signals requires an experimental design that allows us to differentiate collocated , but independent , responses to two different depth cues from an integrated representation that fuses the signals together . To this end , we used random dot displays depicting near or far depth positions of a planar target square relative to its surround . We represent these stimuli as bivariate probability density functions in the space of disparity-motion stimuli ( i . e . , green and magenta blobs in Fig 1A ) . By manipulating dot positions in the two eyes ( binocular disparity ) and differences in the target’s speed relative to its surround ( relative motion ) we could produce different impressions of depth . Using this stimulus space , we created four conditions in which the target’s near versus far depth was defined by ( 1 ) Disparity ( where the motion cue indicated zero depth ) ; ( 2 ) Motion ( where the disparity cue indicated a flat surface in the fronto-parallel plane ) ; ( 3 ) Congruent cues ( where disparity and motion both indicated the same depth ) ; or ( 4 ) Incongruent cues ( where disparity indicated one depth position [e . g . , near] and motion indicated the other [e . g . , far] ) . To understand the experimental logic , we outline the way in which fusion versus independence representational schemes should be engaged by these stimuli . First , it is important to understand that when two cues specify the same depth arrangement ( Congruent condition ) , performance is expected to be best under both scenarios , but for different reasons . For an optimal fusion mechanism , information from disparity and motion is averaged together to produce a depth estimate with lower variance ( Fig 1A , left ) . By contrast , an optimal independence mechanism uses the outputs of separate detectors for the two cues . This corresponds to finding the maximal separation between the two stimuli ( Fig 1A , right ) , which can be intuitively computed as the Pythagorean quadratic sum of the separations along the disparity and motion dimensions . Thus , stimuli are more discriminable because their effective separation is increased ( Fig 1A , right ) . To distinguish fusion from independence , we can measure performance when we manipulate the conflict between the depth information specified by the two cues . First , “single” cue performance ( i . e . , conditions [1] Disparity or [2] Motion ) are useful because they involve one cue indicating no difference in depth between pairs of stimuli ( Fig 1B , left ) . In this case , the independence mechanism effectively ignores the cue signalling zero depth ( e . g . , geometrically , the hypotenuse can never be shorter than one of the catheti ) , while the fusion mechanism is compromised because it averages together one signal that specifies depth ( e . g . , near ) with another that indicates a flat surface . Second , incongruent stimuli can cause radically different responses from the two mechanisms . For an independence mechanism , sensitivity should be comparable to that of congruent stimuli: the separation between the two can become greater for incongruent stimuli , for which the cues specify an opposite depth sign ( Fig 1B , right ) . The independence mechanism is not affected by incongruence , as the depth sign is effectively ignored ( i . e . , the Pythagorean separation still increases whether the cues agree or disagree ) . However , the fusion mechanism is affected: a strict fusion mechanism could be completely insensitive , although ( more realistically ) a robust fusion mechanism would revert to the sensitivity of a single cue component . Using the responses to the different stimuli , we can generate a set of predictions for performance under the fusion and independence scenarios ( Fig 1C ) . For the fusion mechanism , we would expect sensitivity to depth differences in the congruent case to exceed the quadratic summation of performance for the single cue cases . This is because the fusion mechanism is compromised by the conflicts in the single cue stimuli . In addition , performance for incongruent stimuli will be similar to that for single cues ( assuming a robust fusion mechanism ) . By contrast , for the independence mechanism , we would expect congruent cue performance to match the quadratic summation prediction established from the single cue conditions , and no difference in performance for incongruent cues ( Fig 1C , right ) . While a neuronal population might be weighted towards a depth cue integration mechanism , it would be unrealistic to expect a pure fusion-tuned region . In Fig 1C ( middle ) we show an example of the response of a hypothetical hybrid population response , in which fusion and independent units are collocated . Here , we test for cortical regions that respond on the basis of cue fusion using functional brain imaging in monkeys and compare that with previous results obtained in humans [10] . We use multivoxel pattern analysis to quantify sensitivity to differences in brain activity evoked by stimuli depicting different depth configurations and contrast empirical performance with the predictions for independence versus fusion .
To better assess depth processing in the monkey brain , we adapted methods from a previous human study [10] and confined the fMRI responses to independently defined regions of interest ( ROIs; see Materials and methods for definitions ) . We then trained a machine learning classifier ( SVM ) to distinguish between near and far voxel activation patterns for each depth cue and across ROIs . For each area , we calculated the performance of the classifier in decoding depth from an independent data set using a leave-one-out cross-validation approach . Consistent with the searchlight analyses described above , depth defined by congruent stimuli was reliably decoded in most of the ROIs ( Fig 4A ) , but performance varied across areas . This widespread sensitivity to different cues throughout visual cortex does not indicate explicit encoding of depth . The SVM may decode low-level image features , rather than depth per se . Discrimination performance was higher across the three conditions in early visual areas ( V1 , V2 ) , ventral area V4A , dorsal V3 , and MT ( and its satellites ) . Although most of the areas showed lower sensitivity for both single cues compared with the concurrent stimulus , motion classification was particularly poor ( <56% ) in the more dorsal regions ( V3A , DP , PIP , and CIP ) . Our main interest was not to compare classification accuracies of single and combined cue conditions across regions ( because these are influenced by a range of factors besides neural activity , such as our ability to measure fMRI activity with the same sensitivity in different anatomical locations—which basically holds for all fMRI studies comparing signals across areas ) but to evaluate the relative performance within each area of the condition in which both disparity and motion concurrently signalled depth . We compared prediction accuracies for the congruent stimulus relative to a minimum bound for fusion prediction based on the quadratic summation of decoding accuracies for the single cues ( see Figs 1 and 4A ) . We found that fMRI responses were higher than quadratic summation ( P < 0 . 01 ) in areas of the MT cluster ( fundus superior temporal [FST] , medial superior temporal [MST] , and MT ) and ventral V3 . We quantified the extent of integration across areas using a bootstrapped index ( Materials and methods , Eq 1 ) . Values close to zero correspond to the performance expected if information from disparity and motion are collocated but processed independently . A higher value would indicate that a fusion mechanism may be present ( see Fig 1C ) . We found that the integration index in mid-level areas MST and MT significantly exceeded zero ( Bonferroni correction for multiple comparisons , P < 0 . 01 ) ( Fig 4B ) . In addition , we found integration indices above zero ( uncorrected threshold ) in areas FST and V3v ( Table 1 ) ( for human results , see S5 Fig and Ban and colleagues , 2012 ) . Our findings in the MT complex are consistent with decoding improvements for the congruent condition as a result of the fusion of disparity and motion cues ( see Fig 1 ) —although this test alone cannot rule out independences . We added the incongruent condition to perform a second test to assess cue integration mechanisms . Both cues , motion and disparity , were presented simultaneously , but , in this case , depicting opposite depths ( one cue signalled “near” and the other “far” ) . If the representation of depth defined by the two stimulus dimensions ( disparity and motion ) is independent , this conflicting situation should have no effect on classification performance . Thus , the performance of a machine learning classifier in distinguishing between voxel patterns should be similar for both congruent and incongruent conditions and comparable to the quadratic sum of the component cues . In contrast to this expectation , if fusion is present , the discrimination performance should be lower when motion and disparity conflict . Under strict fusion , performance would diminish below that of either component cue , whereas in robust fusion , performance would revert to the level of one of the two components . Area MT exhibited a significant drop in performance for the incongruent condition in comparison to the congruent stimulus ( Fig 5A ) . However , classification accuracy was still slightly higher than for single cue conditions . The visual posterior inferotemporal area ( PIT ) also showed lower prediction accuracy for the incongruent stimulus , although this was not significant after Bonferroni correction . These findings suggest that MT may house a mixed population that contains both units tuned to independent and to fused cues . This may help support a robust fusion process that prevents viewers from becoming completely insensitive to conflicting cues . To assess for brain areas involved in a generalized depth representation , we performed a cross-cue transfer test . In particular , we evaluated whether activity patterns evoked by one depth cue provide information about the other . We trained a machine learning classifier to discriminate depth configurations using one cue ( for example , disparity ) , and tested classifier’s prediction using depth responses elicited by the other cue ( for example , motion ) . For each ROI , we compared the average performance of the classifier in predicting the near versus far stimulus configuration within cues ( trained and tested with the same cue ) and across cues ( trained with one cue and tested with the other ) ( Fig 5B ) . To quantify the transfer of information between cues we calculated a bootstrapped index ( Materials and methods , Eq 2 ) . A value of one would indicate that prediction accuracy is equal for both within- and between-cue classification , meaning that there is 100% transfer of information . To provide a baseline for the transfer that arises by chance , we conducted the transfer test on randomly permuted data . We calculated random transfer performance 1 , 000 times for each ROI , and we chose the 95th percentile of the resulting distribution of transfer indices as the significance threshold . We found transfer indices above the baseline in areas V1 , V2 , V3d , and MT , although only V3d and MT exceeded the threshold significantly ( Bonferroni corrected , P < 0 . 01 ) ( Table 1 ) . Transfer performance of MT was around 55% of that obtained when trained and tested on the same cue ( Fig 5C ) . These results suggest that area MT may play a key role in a more generic representation of depth in the monkey . The same test in humans highlighted dorsal areas V3d and V3B/KO , but did not yield significant results in hMT+ ( S6 Fig ) . So far , we considered scenarios in which neuronal population responses relate either to fusion or independence mechanisms . However , even in areas where we found evidence for depth cue integration ( monkey MT and human V3B/KO ) , it is unlikely that we sampled voxels that respond exclusively to fused signals . To evaluate how different population mixtures might affect decoding results , we used simulations in which we systematically varied the composition of the neuronal population and compared the simulation results with the empirical data ( S7 Fig ) , exactly as in Ban and colleagues , 2012 ( see their Fig 6 and Methods section ) . Whereas in humans the estimated number of fusion units ranges between 50% and 70% in V3B/KO , our simulations suggested that approximately 35% of the neural population in monkey MT might be tuned to fusion . This difference in neuronal composition across species might explain the relative differences in classification performance between human V3B/KO ( Fig 6A of Ban and colleagues , 2012 ) and monkey MT ( S7B Fig ) . In particular , whereas classification performance for the incongruent condition is comparable to that of the single cues in human V3B/KO , the lower percentage of units that contribute to fusion in monkey MT might have caused higher performance for the incongruent condition compared with single cues . Despite these differences , the crucial point is whether the sensitivity for the incongruent condition exceeds the quadratic summation of the single cues . Our statistical tests showed that sensitivity for the incongruent condition was not significantly greater than the quadratic summation in monkey MT nor in human V3B/KO ( S10 Fig ) . In sum , our analyses showed that the ratio of fusion-tuned units in monkey MT is sufficiently high to observe significant differences between the classification performances for congruent and incongruent conditions . In addition to performing our ROI analysis , we used the searchlight multivoxel pattern analysis ( MVPA ) approach to confirm that we did not miss important areas involved in cue integration in monkeys ( Fig 6A ) . We first used the integration index and confirmed our ROI-based analysis of area MT as well as revealing responses in CIP and a confined region in the dorsal part of V2 and V3 . Searchlight analysis of the human brain ( Fig 6C ) pointed to area V3B/KO as the only consistent locus for cue integration . Second , we used a searchlight analysis of the cross-cue transfer test . We found that areas V3d and MT supported significant classification of near-far patterns across cues ( Fig 6B ) . In addition to these two areas , V3A also exhibited high transfer indices . For humans , the highest transfer indices were found around the dorsal areas V3d , V3A , and V3B ( Fig 6D ) . To highlight the implication of areas across the five tests performed ( selectivity for disparity , motion and congruent stimuli , and integration and transfer indices ) , we computed a probabilistic summary map . This color coded map summarizes each voxel that reached significance in each of the five tests ( Fig 7 ) . Taken together , these results suggest that monkey area MT and human V3B/KO are the prime candidate regions for fusing motion and disparity cues in the monkey and human brains , respectively . To assess whether monkeys perceived the depth stimuli in a similar manner as humans ( Ban and colleagues , 2012 ) , we performed a depth discrimination task for each of the four conditions in two rhesus macaques . Two planes with different depths were sequentially presented , and monkeys had to indicate whether the second stimulus ( target depth ) was nearer or farther compared with the first stimulus ( reference depth ) by making saccadic movements to one of two dots on the left and right sides of the screen . Differences between the reference and target depth stimuli ranged between 0 and 6 . 3 arcmin . Both monkeys were able to discriminate between the two depth planes ( P < 0 . 01 ) for all conditions at depth differences higher than 1 . 8 arcmin ( Fig 8A ) . Remarkably , one of the monkeys was able to classify between stimuli even for the smallest tested depth difference ( 0 . 3 arcmin ) when the congruent condition was presented . In general , when depths were discriminable , monkeys showed higher sensitivity ( d’ ) to the congruent stimulus compared with the other conditions . Sensitivity to the relative motion condition was slightly lower than for disparity in both monkeys , while discrimination for the incongruent condition was comparable to that of the single cues . Overall , when pooling the performances across all depth levels and subjects , classification accuracies were above chance for all stimuli and significantly higher for the congruent condition than for the three other conditions ( Fig 8B ) . Moreover , we calculated sensitivity based on just noticeable difference ( j . n . d . ) thresholds [10] . As in humans , we found that monkeys were most sensitive when disparity and motion concurrently signalled depth differences , and they were least sensitive for relative motion–related differences ( Fig 8C ) . We used sensitivities to single cue conditions to calculate their quadratic summation . In line with fusion and the previous human results [10 , 11] , performance for congruent cues exceeded both the quadratic summation and that of the incongruent cues . In summary , these results suggest similar depth perception across conditions between monkeys and humans , with increased performance for the congruent condition and significantly lower discrimination for relative motion–defined depths compared with depths defined by disparity . We analyzed horizontal eye movements from our subjects to assess possible differences between the stimulus conditions . First , we observed that the distribution of eye positions during each condition ( averaged across blocks ) was centered within a 0 . 25° window surrounding the fixation point ( 0° ) for all eight conditions ( S11A Fig ) . Second , for each condition , we computed the mean eye position within each presentation block . No significant differences were observed across conditions in eye positions ( Kruskal–Wallis test , P = 0 . 1315 ) ( S11B Fig ) . Hence , differences in fixation are an unlikely explanation of our findings . These results are in line with previous human data using identical stimuli [10] , also showing that eye positions were not different across conditions , and unlikely to explain the observed differences in fMRI signals .
Depth representation from individual cues ( binocular disparity , motion , texture , shape , shading , etc . ) has been widely studied; however , surprisingly little is known about how the primate brain integrates the information from different depth cues . Previous fMRI studies in humans reported sensitivity to binocular disparity at multiple levels of the visual hierarchy , from early visual cortex to parietal and temporal areas [22–29] . Likewise , large portions of human extrastriate visual cortex are activated by 3D structure from motion [30–33] . Nevertheless , this extensive sensitivity for both cues does not necessarily imply depth selectivity: for instance , it might represent low-level image features , local disparities , or speed of movement ( although this was controlled for in some experiments [5] ) . Depth processing has also been investigated in macaques by measuring fMRI responses to both binocular and monocular cues [5 , 34–39] . Similar to humans , these studies suggested a widely distributed network for 3D representations , implicating monkey occipital , temporal , and parietal cortices . However , the human intraparietal sulcus ( IPS ) has shown much stronger sensitivity to motion-defined 3D than the corresponding monkey region , suggesting that some visuospatial processing areas might not be present in the macaque cortex [5 , 36 , 40] or that similar functions in both species rely on different regions [15] . Behavioral experiments have been performed to compare cue integration in human and nonhuman primates ( macaques ) . Notably , Schiller and colleagues [21] investigated how disparity , motion parallax , and shading depth cues were perceived and integrated . They showed that both species effectively utilized these cues to perceive depth . Moreover , they found that performance was significantly increased when all three cues were presented together , indicating that these separate cues are integrated at yet unknown sites in the primate brain . In addition , consistent with our behavioral data , previous work reported less effective depth discrimination for motion than for disparity in primates [41] . In line with these findings , results from our depth discrimination task in monkeys using disparity and relative motion , together with those previously obtained in adult and juvenile humans [10 , 11] , confirmed the role of these cues in depth perception and that depth cue integration takes place in the primate brain . Overall , similar behavioral results in both species also indicate that monkeys are an excellent animal model for the study of the underlying neural mechanisms of depth perception . Neurophysiological evidence in nonhuman primates suggests sensitivity to binocular disparity throughout visual cortex [40 , 42 , 43] . Recordings in monkey V2 and MT have shown neurons selective to different disparities and clustered organization according to their disparity preference [9 , 44] . Furthermore , monkey MT has also been related to depth representation defined by motion [4 , 6] . A recent study [3] showed that single MT neurons have depth-sign–tuned responses for both binocular disparity and motion cues . Specifically , they observed that 68 cells ( approximately 51% ) preferred either disparity or motion alone , while the remaining cells showed significant selectivity for both conditions . Among these neurons selective for both cues , 37 cells ( approximately 28% of the population ) had the same depth-sign preference ( congruent cells ) , and 29 cells ( approximately 21% of the population ) had opposite depth-sign preference ( incongruent cells ) . Thus , disparity tuning at the single cell level did not always match across cues , supporting the existence of congruent and opposite depth-sign–tuned cells . Moreover , responses of congruent neurons to the congruent combination of binocular disparity and motion cues showed an increased selectivity compared with that from single cues , but no enhancement was found in the incongruent cells ( when disparity and motion were presented congruently ) . In Fig 1C we depicted the two ideal scenarios for fusion and independence; however , in light of previous findings , it is unrealistic to expect such a pure neuronal assembly , but a hybrid arrangement instead . On one hand , if a sizeable neuronal subpopulation contributes to fusion , we should be able to detect a sensitivity increase caused by the reduced variance , ( only ) when the congruent stimulus is presented . At the same time , in a mixed population , separate portions of neurons respond independently to each of the single cues , and their aggregate neuronal response patterns will also be more discriminable when cues are presented together . Moreover , the increase in discriminability by these independent assemblies should equally affect the incongruent condition , because the improvement corresponds to the quadratic sum of the discriminabilities of the marginal distributions . Critically , the contribution of the fusion mechanism will elicit a significant sensitivity improvement that will exceed the quadratic summation of the single cues for the congruent condition , but not for the incongruent stimulus . The combination of both effects can be seen in our results in monkey MT ( S10 Fig , left ) , where ( i ) a fusion-tuned subpopulation contributes with a substantial enhancement that exceeds the performance of the incongruent condition and the quadratic summation of the individual cues; and ( ii ) separate single cue–tuned neural subpopulations contribute with higher responses for the incongruent condition compared with single cues , although not higher than their quadratic summation . Altogether , the findings of Nadler and colleagues [3] are in agreement with our results , because we obtained higher classification performance when the two cues are presented congruently ( S10 Fig ) . To the best of our knowledge , responses of congruent cells to the incongruent condition and responses of incongruent cells to both congruent and incongruent conditions were not tested in Nadler and colleagues [3] , nor elsewhere [45] . These findings support MT as a potential candidate for integrating depth cues , although the role of opposite cells is currently not understood . Additional electrophysiological studies , in which both congruent and conflicting combinations of disparity and motion cues are tested , are required to further our understanding of cue integration mechanisms at the single cell level in MT . At a mechanistic level , it is important to note that contributions made by either “congruent” or “incongruent” neurons , when estimating environmental properties from the combination of signals , are unknown . Recently , however , we proposed a model whereby both congruent and incongruent responses contribute to estimating the most likely depth of the scene ( at a population level ) [46] . In particular , a population “best guess” estimate is computed by combing the outputs of congruent neurons that drive excitation and incongruent neurons that drive inhibition . This process allows robust perceptual estimates—accounting for improved performance when signals are consistent , and robust reversion to one of the signals under conflict . This model explains why many neurons are incongruent , and how the use of such neurons supports robust perceptual estimations by the brain . Our fMRI logic is based on separating “independence” versus “fusion . ” It is not specific to the way in which fusion is implemented , and , as corroborated by the new model , it is likely to involve the contributions of both “congruent” and “incongruent” neurons . A given stimulus will evoke activity within the population of fusion of neurons ( some will appear “congruent” and others “incongruent , ” while , in reality , all encode information about the likelihood of the viewed stimulus ) . Thus , there is nothing special about the response that will be evoked by a consistent or inconsistent stimulus in terms of the types of neurons that they will stimulate . However , there is a difference in the statistical evidence in favor of a particular depth interpretation: this will be higher when cues are consistent . The present results invite future studies on the neural implementation of the new model [46] at the single cell level in monkey area MT . Recent tests for cue fusion revealed V3B/KO as the main cortical locus in the human cortex for the integration of depth defined by binocular disparity and motion cues [10 , 11] . Complementary fMRI studies that used texture [47] , shading [48] , and gloss [49] supported the capability of human area V3B/KO to integrate qualitatively different depth cues . However , no evidence of integration was found in human MT+ . Considering that no conclusive evidence yet exists concerning the existence of a macaque homologue of human V3B/KO , one could argue that cue fusion for motion and disparity in monkey occurs in a region more caudally relative to MT , in a location that corresponds at least topographically with V3B/KO . In this respect , it is also worth mentioning that the parcellation of dorsal extrastriate visual cortex of the monkey is highly complex and that the exact areal definitions are highly disputed . Nevertheless , our results suggest an areal difference across primate species for depth cue integration and a potential homology between human V3B/KO and monkey MT for this specific functionality . Then , can we claim that human MT+ is not an exact homologue to monkey MT ? Of course , it is unreasonable to expect that 100% identical areas can be found in two different species , as ecological pressure must have triggered ( slightly ) different functional properties in individual species . There is evidence that human MT+ and monkey MT might share a large range of functions . For example , when we previously showed monkeys and humans identical videos and correlated the “free-viewing” fMRI signals from independently identified monkey MT with all signals from all voxels of the human cortex , we found significant correlations not only in human area MT+ but also in dorsal areas of the visual cortex . However , when seeding in human MT+ , the correlations we found were surprisingly well confined to area MT in the monkey [15] . These entirely data-driven results suggest that MT shares a number of functional properties across species , but not all . While human MT+ is functionally more closely related to MT than other areas in the monkey cortex , monkey MT might be carrying some functionalities that through evolutionary pressure are distributed across several regions of the human visual cortex ( S8 Fig ) . Thus , according to our results and those of Ban and colleagues [10] , human MT+ cannot be considered an exact homologue of monkey MT , at least with regard to depth cue integration . Response patterns across species differed substantially in MT ( see S9A Fig ) . Although responses for disparity , relative motion , and their combination are discriminable in both human MT+ and monkey MT ( indicating that there is not a lack of responsivity in MT of both species ) , human MT+ did not show significant increases for the congruent condition compared with the quadratic summation of the single cues , nor a significant transfer of depth information across cues—which are required to support fusion . Moreover , when comparing the relative performances of MT across species , we observed that monkey MT exceeded human MT+ significantly ( S9B Fig ) . Considering these and previous findings , monkey MT share depth cue fusion processes more with human V3B/KO , compared with human MT . However , this specific correspondence does not rule out other functional correspondences between human and monkey MT . CIP has also been implicated in processing depth from different cues ( disparity and texture ) and integration [50 , 51] . In our study , we observed higher classification accuracy for the congruent condition compared with the quadratic summation of the single cues and the incongruent condition in area CIP; however , results were not significant ( Fig 4 ) . In addition , near-far discriminability for the motion-defined depth condition dropped to chance level in area CIP , as well as in the rest of parietal regions except for LIP . Accordingly , human V7 , the putative corresponding area of monkey CIP [52] , did not show selectivity for depth cues integration [10 , 47] . To depict depth structure from motion , we used a display in which near and far conditions were defined by differences in speed between a target ( center ) and reference plane ( surround ) . On one hand , the reference plane always moved horizontally ( and sinusoidally , from left to right and right to left ) with a constant speed . Importantly , the direction of the reference plane movement was unambiguous due to a static reference background surrounding the reference plane ( Fig 1 ) . When the centrally presented target plane moved faster than the reference plane , it appeared to protrude towards the viewer ( near percept ) . The opposite effect ( far percept ) was established when the target plane moved slower compared with the reference . Because there is no ambiguity on the direction of the reference plane movement , the differential movement velocity disambiguated the sign of depth in our stimuli [53 , 54] . Moreover , the occlusion of dot patterns at the edge between target and reference planes may also contribute to solve depth-sign ambiguity [55 , 56] . It has been shown that accretion–deletion in the presence of relative motion provides sufficient information to disambiguate depth sign . However , accretion–deletion alone is unable to provide any depth perception [56] . Results from our depth discrimination task in monkeys ( Fig 8 ) , together with those previously reported in humans [10] , suggest that both human and nonhuman primates are capable to discriminate between near versus far depth positions on the basis of relative motion defined in this study . As in any fMRI experiment , responses from voxels reflect a complex mixture of neural signals ( originating from thousands of neurons ) , most of them even unknown . It is therefore natural to ask whether the fMRI decoding performance that we measure reflects differences in the perceived depth , as opposed to simpler differences in the speed of motion . In particular , it is well known that MT contains neurons that respond differentially to different speeds of motion [57] . Our data contain a number of lines of evidence suggesting that speed per se is unlikely to be responsible for the results that we report . First , the congruent and incongruent stimulus conditions both contained stimuli that differed in disparity-defined depth and the speed of motion . An explanation for our data premised on motion speed alone would not predict the differential decoding performance we find for congruent versus incongruent stimuli in monkey MT . Specifically , the speed of motion for the congruent condition , near disparity and near motion ( NN ) , is matched to that of the incongruent condition , far disparity and near motion ( FN ) , and the speed in the congruent condition , far disparity and far motion ( FF ) , is matched with that of the incongruent condition , near disparity and far motion ( NF ) . Thus , if discrimination relied on speed , the classification of congruent conditions ( NN versus FF ) and incongruent conditions ( NF versus FN ) should be similar . Our results in monkeys ( MT ) and humans ( V3B/KO ) , however , showed differences in performance between congruent and incongruent conditions . This suggests that there is something particular about the difference in speed , in that it produces an impression of depth compatible with the depth signal provided by the disparity cue . Second , if the classification were a mere consequence of speed differences , we would not have expected to see transfer of decoding performance between the different cue conditions ( Fig 5 ) . It might be possible to explain this finding based on an association between speed and disparity preference in the underlying selectivities of individual neurons . However , although MT neurons with strong speed tuning also tend to have strong disparity tuning , there is no evidence for a correlation between the preferred magnitude of disparity and velocity [58] . Thus , there is little evidence to suggest that retinal speed signals are confounded with the disparity sign of the stimuli . Although we are aware that fMRI responses contain a complex mixture of neural signals with different properties , such as speed preferences , we believe that the tests performed in our experiment are sufficiently robust to overcome such difficulty . Together , this suggests that the results we report in monkey MT and human V3B/KO are likely to be due to the impression of depth evoked by different speeds of motion , rather than speed . We found that fMRI responses are more discriminable when the two cues signal depth concurrently , and that depth information provided by one cue might be diagnostic of depth indicated by the other . We revealed that monkey area MT shows fMRI signals consistent with a fusion mechanism of independent depth cues . These results may reconcile the human imaging data with previous monkey electrophysiological studies implicating area MT in depth perception based on motion and binocular disparity signals . Our findings , together with those obtained in humans , provide evidence for a fusion mechanism for depth perception in the dorsal stream of primates . The fusion of depth cues , however , appears to be computed in different areas in humans ( V3B/KO ) and monkeys ( MT ) . Therefore , it is tempting to speculate that human V3B/KO may have been part of the MT cluster in an ancestor of monkeys and humans , which has drifted in a caudo-dorsal direction during human evolution .
Four rhesus monkeys ( Macaca mulatta , two female ) , 4–6 years old , weighing between 4 and 7 kg , participated in the experiments ( two underwent fMRI and the other two performed the behavioral task ) . Animal care and experimental procedures met the Belgian and European guidelines and were approved by the ethical committee of the KU Leuven Medical School ( Protocols P103/2008 and P022/2014 ) . Animals were born in captivity and were pair- or group-housed ( two to five animals per group; cage size at least 16–32 m3 ) with cage enrichment ( toys , foraging devices ) , outside views , and natural day-night cycles ( throughout the year , supplemented with an artificial 12/12-hour light/dark cycle ) at the primate facility of the KU Leuven Medical School . They were daily fed with standard primate chow supplemented with bread , nuts , raisins , prunes , and fruits . The animals received their water supply either during the experiments or in the cages before and after the experiments . Monkeys had previous experience performing behavioral tasks and were prepared for fMRI sessions . Prior to scanning , monkeys that underwent fMRI were trained daily ( 2–4 weeks ) to perform a passive fixation task while in a sphinx position with their head rigidly fixed in a plastic primate chair . For the discrimination task , subjects were trained to answer using saccadic eye movements to left or right targets . Details concerning head-post surgery and training procedures have been previously described [59] . During the experimental period , access to water was restricted , but animals were allowed to drink until fully satiated during the daily training and scanning sessions . The human data were the same as those presented in Ban and colleagues [10] , although additional data analyses were performed on this data set ( see below ) . Because the basic cue-fusion results observed in human V3B/KO described in Ban and colleagues [10] have already been replicated using exactly the same stimuli [11] and with entirely different stimuli [47 , 48] , we deemed it unnecessary to replicate them in the present study . Despite our efforts to perform an identical experiment , there are always unavoidable differences between monkey and human fMRI experiments , which are summarized in S1 Table . One difference is that humans performed a subjective assessment of eye vergence [10] , while monkeys received liquid reward to maintain fixation to a fixation point . However , it is rather unlikely that differences in the fixation task greatly impacted the sensory-driven activations . In particular , previous experiments showed that activity patterns can be similar during anesthesia versus awake fixation experiments [60] . Stereoscopic presentation and display parameters were matched as closely as possible to those used in the previous human study [10] and will be only summarized briefly . For a detailed table listing every difference between the monkey and the human study , see S1 Table . Stimuli consisted of two planes , the reference plane and the target plane , defined by random patterns of dots ( 0 . 15° size , 15 dots/deg2 density ) . The target plane was represented by a square ( 10 × 10° ) located in the center of the screen , superimposed on the rectangular reference plane ( 18 × 14° ) . In addition , this random dot region was surrounded by a static grid ( 43 × 32° ) of black and white squares that provided an unambiguous background ( permanently present on the screen on top of a gray flat surface ) . At the center of the screen , a fixation square ( 0 . 5° × 0 . 5° ) was presented within a circular mask ( 1° diameter ) . The reference plane remained the same across all conditions , with no stereoscopic structure and a constant motion profile . The stereoscopic structure of the dots and the motion of the target plane were altered to depict depth relative to the reference plane . We defined eight stimuli and four different conditions: depth defined ( i ) by disparity , ( ii ) by motion , ( iii ) by congruent disparity and motion , and ( iv ) by incongruent disparity and motion . Disparity-defined stimuli were rendered as red-cyan anaglyphs . To create the random dot stereograms ( RDSs ) , we first measured the distance between the centers of the two eyes of our subject to adjust the dot displacements between the left- and right-eye images to the interpupillary distance of each monkey . The luminance of the red dots through the red filter was 11 cd/m2 and 0 . 2 cd/m2 through the cyan filter , and the luminance of cyan dots was 10 . 1 cd/m2 through the cyan filter and 0 . 0 cd/m2 through the red filter . For the disparity-alone condition , the target plane was given a horizontal binocular disparity of ±9 arcmin to signal near or far depth relative to the reference plane that presented no binocular disparity structure . Both target and reference planes moved rigidly ( same speed and phase ) with a sinusoidal horizontal movement of 0 . 9° amplitude and 1-second period ( i . e . , in 1 second , the planes covered a displacement of 0 . 9° from left to right and from right to left ) . To depict depth from relative motion alone , the target plane moved horizontally with different amplitudes ( 1 . 32° for near and 0 . 29° for far ) relative to the reference plane that kept the same movement amplitude ( 0 . 9° ) . Both planes moved in phase and followed a sinusoidal velocity profile with a 1-second period ( Fig 1 ) . The direction of the movement was always obvious due to a static background surrounding the moving reference plane ( see above ) . The relative motion of the target plane with respect to the reference plane gave rise to a pattern of deletion and accretion of the reference ( near stimuli ) or target ( far stimuli ) dots , as the two planes translated back and forth across the screen . Because there is no ambiguity about the direction of the reference plane movement due to the static background , the speed difference between planes supports the disambiguation of depth in our stimuli ( i . e . , when the target moves faster than the reference , it always appears to be protruding toward the viewer ) [53] . The near/far speeds of the relative motion stimuli were defined based on the psychophysical experiments [10] , during which human observers were asked to judge the impression of depth perceived from relative motion differences of the target and reference planes ( the reference plane speed was fixed ) and depth perceived from the binocular disparities . We took the psychophysical matching point between the disparity and relative motion stimuli for performing the main fMRI experiments . Even when the eye and/or head movements are restricted , as in the present study and that of Ban and colleagues [10] , the depth sign from relative motion was not ambiguous for our stimuli . Our introspection of having viewed these stimuli many times , and the reports of the human observers , indicate that our stimuli were effective in depicting near versus far depth positions on the basis of relative motion . For stimuli concurrently depicting disparity and motion-defined depth , the target plane had a disparity of ±9 arcmin and a movement amplitude of either 1 . 32° or 0 . 29° . In the congruent condition both disparity and motion signalled the same depth ( near-near or far-far ) . For the incongruent case , disparity and motion simultaneously provided opposing information ( near-far or far-near ) . Monkeys were placed within the bore of the magnet in sphinx position inside a plastic primate chair using a physical head restraint . Images were projected ( Barco 6300 LCD projector ) on a translucent screen located at a distance of 57 cm from the monkey . Subjects had to fixate passively on a square presented in the center of the screen . Eye position was monitored at 120 Hz using a pupil-corneal reflection tracking system ( Iscan ) . To encourage monkeys to maintain fixation and remain quiet , liquid reward was delivered through a plastic tube located just inside their mouths . Stimuli were presented to the monkeys through colored filters ( red-cyan anaglyph goggles , as in Durand and colleagues [34] and Tsao and colleagues [38] ) placed in front of the eyes . Before each scanning session , a contrast agent , MION , was injected into the femoral/saphenous vein ( 6–11 mg/kg ) to improve the contrast-to-noise ratio [59 , 61] . We presented the stimuli in blocks of 16 seconds . In each block , stimuli were picked randomly from a set of 24 example stimuli ( per subject ) that differed in the random placement of dots making up the stereogram . Individual stimuli were presented for 1 second , followed by a 1-second fixation period . Three blocks of each stimulus type were randomly presented during an individual run ( 24 stimulus blocks ) , and the scan started and ended with a 16-second fixation interval that served as baseline . Each scan lasted 416 seconds , during which 208 volumes were acquired . Monkey data were acquired with a 3T MR Siemens Trio scanner with an AC88-insert head gradient . Functional images were collected using a gradient-echo T2*-weighted echo-planar imaging sequence ( repetition time [TR] = 2 , 000 ms , echo time [TE] = 17 ms , 52 slices , voxel size = 1 mm isotropic , flip angle = 75° ) . Monkeys were scanned with a custom-built , eight-channel , implanted phased-array receive coil [62] and a saddle-shaped , radial transmit-only surface coil . To provide an anatomical reference for the functional scans , high-resolution T1-weighted images were acquired for each monkey during a separate session under ketamine-xylazine anesthesia , using a single radial transmit-receive surface coil and a MP-RAGE sequence ( TR = 2 , 200 ms , TE = 4 . 05 ms , flip angle = 13° , 208 slices , voxel size = 0 . 4 mm isotropic ) . During the session , 12–15 whole-brain volumes were obtained and averaged to improve signal-to-noise ratio . Functional volumes were reconstructed online using Siemens GRAPPA image reconstruction . We preprocessed the fMRI data using custom Matlab ( MathWorks ) scripts and the SPM5 software package ( Wellcome Department of Cognitive Neurology , London , UK; http://www . fil . ion . ucl . ac . uk/spm/ ) . We only analyzed runs in which monkeys performed more than 97% of fixation within a 2 × 2 deg . fixation window . Temporal preprocessing was applied to correct for linear trends and spin-excitation history effects caused by head motion . Spatial preprocessing consisted of motion correction and rigid coregistration to the individual anatomical template . To compensate for echo-planar distortions and intersession variance , functional images were matched to the anatomy , applying nonlinear warping in JIP ( www . nitrc . org/projects/jip [63] ) . No spatial smoothing was performed on the data for the analysis . For each monkey , we used retinotopic mapping procedures to define the ROIs in the occipital cortex , IT , and posterior parietal cortex . Retinotopic organization in occipital visual areas and IT cortex ( V1 , V2 , V3v , V3d , V4 , V4A , V4t , OT , MT , MST , FST , and PIT ) of both subjects was found to be highly consistent with previous reports [64] . We used individual retinotopic maps along with previous parcellation schemes [65] to delineate V3A , DP area , and posterior parietal cortex areas ( PIP , CIP , and LIP ) . The anatomical labels were drawn on the inflated cortical surface of each monkey using FreeSurfer ( http://surfer . nmr . mgh . harvard . edu ) and then projected into the volume space . To ensure that distinct ROIs were not overlapping , we performed an automatic correction using customized Matlab code . Preprocessing of the fMRI responses and ROI definitions in human are described in Ban and colleagues [10] We first performed a t test to contrast the response to all stimulus conditions versus the fixation baseline across all the runs . Within each ROI , we selected gray matter voxels from both hemispheres and sorted them by their t statistic . We then selected the top 150 voxels from each ROI . For smaller cortical areas with less than 150 voxels , we selected all the voxels with a t value > 0 [25] . We normalized ( z-score ) each voxel’s time course separately for each experimental run to minimize baseline differences across runs . The data vectors for the multivariate analysis were generated by shifting the fMRI time series by two volumes , to account for the hemodynamic response delay . We then averaged all data points within each block to obtain the voxel patterns . To control for the possibility that classification accuracy was due to a univariate baseline difference , we normalized each pattern vector by subtracting the mean voxel amplitude . We used a linear SVM ( LibSVM ) to classify between fMRI response patterns evoked by near versus far stimulus presentations for each condition separately and performed an n-fold leave-one-run-out cross-validation , in which data from all runs but one were used as training patterns ( 24 patterns per run , 6 per condition ) and data from the remaining run were used as test patterns . Then , the mean accuracy across cross-validations was used . To quantify differences in SVM prediction accuracies between combined-cue conditions and the minimum bound prediction , we calculated an fMRI integration index ( ϕ ) : ϕ=d′D+Md′D2+d′M2−1 ( 1 ) where d′D+M is the classifier’s performance in the congruent condition , and d′D and d′M are performances for single cue conditions in d-prime units , calculated with the following formula: d′=2×erfinv ( 2p-1 ) ( 2 ) where erfinv is the inverse error function and p is the proportion of correct predictions . To conduct the transfer test analysis , we first used a recursive feature elimination ( RFE ) method [66] to select the voxels in each ROI with the highest discriminative power . Then , we used a standard SVM to compute within- and between-cue prediction accuracies . To assess the relationship between transfer classification performance and the mean performance for single cues , we calculated a transfer index ( T ) : T=2d′Td′D+d′M ( 3 ) where d′T is the transfer classification performance , and d′D and d′M are the performances for single cue conditions . Statistical significance of the results was evaluated using bootstrapped resampling with 10 , 000 samples . We performed a searchlight classification analysis [67] in volume space for each monkey by selecting spherical volumes of cortical voxels within a 3-mm radius , moving voxel-wise through the entire cortical volume . Prior to the analysis , a mask was applied to exclude voxels outside the cortex . We ran three SVM classification analyses ( disparity , motion , and congruent and incongruent conditions ) discriminating between near-far patterns . We then calculated the fMRI integration index at each voxel location ( Eq 1 ) . To assess transfer classification accuracies , we conducted between-cue classification analysis . Finally , we computed group maps for each condition and projected the result onto a representative cortical surface ( inherently causing some smoothing ) . This analysis confirmed that all the relevant voxels in the classification analyses were captured by the ROIs previously defined . In addition to the ROI-based analysis from Ban and colleagues , 2012 [10] , we ran the searchlight ( 8-mm radius ) analysis in humans following the same procedures as described for the monkey . We then computed groups maps and projected the results onto a flattened cortical surface of the human brain . As in the fMRI experiments , random dot patterns depicting depth from binocular disparity , relative motion , and the combination of both ( congruent and incongruent ) were presented to two monkeys . Because the monkeys of the fMRI experiment were not available anymore , two different animals ( Monkey B and Monkey T ) , were trained on this task . Monkeys were placed in sphinx position in front of an LCD screen ( 57-cm distance ) and viewed the stimuli through colored filters ( red-cyan ) . In each trial , two stimuli with a slight depth difference between them were sequentially shown , and subjects indicated which of the two planes was farther by making a saccade to one of two targets ( 1° dots at 10 deg eccentricity ) located on the left and right sides of the screen . The first stimulus ( 1-second duration , reference depth ) systematically showed the same depth as the far stimuli during the fMRI experiment ( 9 arcmin for disparity and/or 0 . 29° movement amplitude ) , whereas the second plane ( 1 second , target depth ) was presented at different levels of depth compared with the reference plane ( 9 ± 0–6 . 3 arcmin for disparity and 0 . 29 ± 0–0 . 18° for motion ) . For the case of the incongruent condition , disparity far and motion near were depicted , as in the previous human study [10 , 11] . Each condition was presented in alternate runs , which consisted of 360 trials ( 15 trials per 12 levels for near or far depths ) in which different depth levels of the same condition were randomly presented . A total of 140 runs were acquired . Eye positions of one eye were monitored in both monkeys ( Iscan , 120 Hz ) while they performed the passive fixation task during fMRI sessions . To assess possible differences in eye positions between conditions , we analyzed horizontal eye movements in detail—which are most relevant for depth-defining stimuli . First , for each block ( 16 seconds ) , we measured eye position and calculated the average eye position for each stimulus across trials . The mean eye traces were used to compute the distribution of the eye position for each condition . Second , the mean eye position within each presentation block ( across the 16 seconds of trial duration ) was computed . Then , fixation per condition and 95% confidence intervals were calculated . A Kruskal–Wallis test was used to calculate significant differences between conditions . | In everyday life , we interact with a three-dimensional world that we perceive via our two-dimensional retinas . Our brain can reconstruct the third dimension from these flat retinal images using multiple sources of visual information , or cues . The horizontal displacement of the two retinal images , known as binocular disparity , and the relative motion between different objects are two important depth cues . However , to make the most of the information provided by each cue , our brains must efficiently integrate across them . To examine this process , we used neuroimaging in monkeys to record brain responses evoked by stimuli signaling depths defined by either binocular disparity or relative motion in isolation , and also when the two cues are combined congruently or incongruently . We found that cortical area MT in monkeys is involved in the fusion of these two particular depth cues , in contrast to previous human imaging data that pinpoint a more posterior cortical area , V3B/KO . Our findings support the existence of depth cue integration mechanisms in primates; however , this fusion appears to be computed in slightly different areas in humans and monkeys . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"functional",
"magnetic",
"resonance",
"imaging",
"brain",
"vertebrates",
"social",
"sciences",
"neuroscience",
"animals",
"mammals",
"magnetic",
"resonance",
"imaging",
"primates",
"brain",
"mapping",
"animal",
"management",
"animal",
"performance",
"neuroimaging",
"old",
"world",
"monkeys",
"research",
"and",
"analysis",
"methods",
"imaging",
"techniques",
"monkeys",
"animal",
"cells",
"visual",
"cortex",
"agriculture",
"macaque",
"psychology",
"radiology",
"and",
"imaging",
"eukaryota",
"diagnostic",
"medicine",
"sensory",
"cues",
"cell",
"biology",
"cellular",
"neuroscience",
"anatomy",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"sensory",
"perception",
"amniotes",
"organisms"
]
| 2019 | Areal differences in depth cue integration between monkey and human |
Cytomegaloviruses ( CMVs ) encode cellular homologs to evade host immune functions . In this study , we analyzed the roles of GP33 , a guinea pig CMV ( GPCMV ) -encoded G protein-coupled receptor ( GPCR ) homolog , in cellular signaling , viral growth and pathogenesis . The cDNA structure of GP33 was determined by RACE . The effects of GP33 on some signaling pathways were analyzed in transient transfection assays . The redET two-step recombination system for a BAC containing the GPCMV genome was used to construct a mutant GPCMV containing an early stop codon in the GP33 gene ( Δ33 ) and a rescued GPCMV ( r33 ) . We found the following: 1 ) GP33 activated the CRE- and NFAT- , but not the NFκB-mediated signaling pathway . 2 ) GP33 was dispensable for infection in tissue cultures and in normal animals . 3 ) In pregnant animals , viral loads of r33 in the livers , lungs , spleens , and placentas at 6 days post-infection were higher than those of Δ33 , although the viruses were cleared by 3 weeks post-infection . 4 ) The presence of GP33 was associated with frequent lesions , including alveolar hemorrhage in the lungs , and inflammation in the lungs , livers , and spleens of the dams . Our findings suggest that GP33 has critical roles in the pathogenesis of GPCMV during pregnancy . We hypothesize that GP33-mediated signaling activates cytokine secretion from the infected cells , which results in inflammation in some of the maternal organs and the placentas . Alternatively , GP33 may facilitate transient inflammation that is induced by the chemokine network specific to the pregnancy .
Congenital infection of human cytomegalovirus ( HCMV ) occurs in 0 . 2–1% of all births and causes birth defects and developmental abnormalities , including neurological sequela [1–4] . In contrast to murine CMV ( MCMV ) and rat CMV ( RCMV ) , guinea pig CMV ( GPCMV ) causes infection in utero and pathogenic outcomes similar to those caused by HCMV , which makes GPCMV animal models useful for studies on congenital CMV diseases [5–8] . Virus-encoded G protein-coupled receptor homologs ( vGPCRs ) are characteristic of beta- and gamma-herpesviruses and are considered potential therapeutic targets . HCMV encodes four vGPCRs: UL33 and UL78 are conserved in all beta-herpesviruses , including MCMV , RCMV and GPCMV , whereas US28/US27 homologs are restricted to primate beta-herpesviruses . Although US28 and UL78 are expressed early after infection , US27 and UL33 are expressed during the late phase of infection [9] . It is well known that stimulation of GPCR activates heterotrimeric G proteins , which are composed of α , β , and γ subunits , and dissociates them into α and βγ subunits , resulting in production of second messengers for transcriptional gene regulation . Among the Gα family , Gαs stimulates cyclic AMP ( cAMP ) production , resulting in cAMP response element ( CRE ) activation , while Gαi inhibits cAMP production . Gαq stimulates phospholipase C ( PLC ) , resulting in activation of protein kinase C ( PKC ) and nuclear factor of activated T-cells ( NFAT ) . GPCRs associate with G proteins through a DRY ( Asp-Arg-Tyr ) motif located at the cytosolic end of the third membrane-spanning segment , and the motif plays an important role in the conformational transition between active and inactive states of GPCRs [10] . Although UL33 contains the conserved DRY motif , it activates PLC , p38 , and CRE binding protein ( CREB ) constitutively [11] . A structural study on US28 suggests that the DRY motif and its immediate environment affect constitutive activity of viral GPCRs [12] . M33 , an MCMV homolog of UL33 , displays ligand-independent , constitutive signaling through the Gq/PLC pathway [13] . M33 also activates CREB in a ligand-independent , constitutive manner . M33 contains a NRY motif at the amino acid ( aa ) position 130–132 in place of the DRY motif , and the alteration R131A but not N130D in the motif abolished the constitutive signaling [14] . M33 and R33 , but not UL33 , activate the NFκB pathway . This activation by M33 is G protein-independent , but that by R33 is G protein-dependent [15 , 16] . Functional similarities in signaling activities between M33 and US28 , including activation of the NFκB pathway and vascular smooth muscle cell migration , have been reported [13 , 17] . UL33 and UL78 form heterodimers with human CC chemokine receptor 5 ( CCR5 ) . The heterodimerization of UL33 impairs the CC chemokine ligand 5 ( CCL5 ) -induced internalization of CCR5 and the CCR5-mediated cell migration and HIV entry [18] . The roles of MCMV and RCMV vGPCRs in vivo have been studied intensively . After CMV infection , acute-phase replication in the primary organs is followed by dissemination to the secondary sites , such as the salivary glands , where the virus may replicate for several days , affording an opportunity for horizontal viral spread . The lack of M33-dependent signaling decreased viral loads in the salivary glands [14 , 19] . A recombinant MCMV expressing US28 partially complemented the defect due to the lack of M33 in BALB/c mice infected intraperitoneally ( i . p . ) [20] . A recent study demonstrated that M33 is required for viral replication in the salivary glands in NOD/scid-/infg- ( NSG ) mice infected i . p . [21] . Intranasal ( i . n . ) administration of MCMV lacking M33 also demonstrated that transmission from the lymph nodes ( LNs ) to the lungs requires M33 functions [22] . As there are few studies on GPCMV GP33 and as the GPCMV animal model can address the question of whether GP33 plays any role in congenital infection , this study sought to characterize the roles of GP33 in cellular signaling and viral pathogenesis , particularly under pregnant conditions .
RNA samples were prepared from guinea pig lung fibroblasts ( GPL ) at 3 days after infection with GPCMV . The cDNA structure of the GP33 gene was determined by RACE analyses using the primers listed in S1 Table . The GP33 gene consists of three exons that encode a 424 amino acid ( aa ) product ( Fig 1A ) and contain the conserved DRY motif at the aa positions 130–132 . Homologies of the GP33 aa sequence with GP33 homologs of other herpesviruses are more than 85% with similar distances from GP33 to UL33 and to M33 ( Fig 1B ) . A plasmid encoding GP33 fused with EGFP at the carboxyl side was constructed . Immunofluorescence assay ( IFA ) of the cells that were transfected with the plasmid and fixed with paraformaldehyde demonstrated that GP33 was localized mainly in some vesicles in the cytoplasm and slightly in the plasma membrane ( Fig 1C ) , as shown in cells expressing UL33 [11 , 23] . Plasmids encoding the GP33 ORF without and with a FLAG-tag at the carboxyl end were also constructed . Immunoblotting of the transfected cell lysate with anti-FLAG antibodies demonstrated that GP33 formed a ~50 kDa band but was also heavily aggregated ( Fig 1D ) . Next , transient transfection of a CRE reporter plasmid along with GP33s lacking exons 1 and 2 of GP33 , GP33 cDNA , and HCMV UL33 , respectively , demonstrated that GP33 weakly activated CRE and that an 11-aa sequence from the amino-terminal end of GP33 , which is absent in GP33s , was important for this CRE activation ( Fig 2A ) , probably due to a difference in the efficiency of correct subcellular trafficking [11 , 14] or of glycosylation at an NXT motif just before the cysteine residue of GP33 , which is conserved among the UL33 homologs , near the amino-terminal region [24] . In addition , the NFAT- but not the NFκB-signal pathway was activated by GP33 ( Fig 2B and 2C ) . As M33 but not UL33 activates the NFκB pathway [13] and as the homology distances of GP33 to UL33 and M33 are similar , GP33 may function in manner more closely resembling that of UL33 . To elucidate the roles of GP33 in the viral dissemination and pathogenesis , GFP-expressing GPCMV with a stop codon mutation in the GP33 ORF ( Δ33 ) and that with the rescued wild-type GP33 ORF ( r33 ) were generated using the two-step redET recombination method . Rawlinson and his colleagues demonstrated upregulation of some pro-inflammatory cytokines in the placenta and amniotic fluid of congenitally infected fetuses in human as well as increase of MCP-1 and TNF-α expression in HCMV-infected ex vivo placental explant [25 , 26] . In addition to those findings , since it is well known that NFAT signaling activates TNF-α expression in some cell types and that NFAT can interact with several transcription factors , including those downstream of GPCR signaling , for synergistic activation ( reviewed in [27–29] ) , we analyzed the effect of GP33 on TNF-α expression . As shown in Fig 2D , real-time reverse transcription ( RT ) PCR assays demonstrated that TNF-α mRNA level in r33-infected GPL cells was higher than that in Δ33-infected GPL cells ( p<0 . 05 ) . In GPL cells , GPCMV r33 and Δ33 were found to grow to a similar extent ( Fig 3A ) . In addition , both viruses infected guinea pig epithelial cell clone GPE-7 as well as macrophages prepared by TPA-treatment of splenic monocytes at a similar level ( means and standard errors of means ( SEMs ) of GFP+ cells/well were r33 = 1 , 010 ± 135 vs . Δ33 = 1 , 206 ± 107 and r33 = 2 , 406 ± 388 vs . Δ33 = 1 , 741 ± 114 , respectively ) ( Fig 3B ) . These results indicate that GP33 is dispensable for infection and growth in vitro . Female 5-week old normal guinea pigs ( Hartley ) were infected i . p . with 1x106 or 2x107 infectious units ( IUs ) /animal of GPCMV ( r33 or Δ33 ) and sacrificed at 1 week post-infection ( p . i . ) . There were no apparent abnormalities in the normal animals infected with r33 or with Δ33 . No significant differences in the appearance of any organs were observed irrespective of the dose , low ( 1x106 IUs/animal ) or high ( 2x107 IUs/animal ) , and there were no differences in viral loads in their livers , spleens , lungs , kidneys , or plasma specimens ( Fig 4 ) . As it appears that GP33 is not involved in disease pathogenesis in normal animals , dams ( Hartley ) at 4 weeks’ gestational age ( GA ) were infected subcutaneously ( s . c . ) with GPCMV r33 or with Δ33 ( n = 4 each; 1 . 2 x 107 IUs/animal ) and sacrificed at 6 days p . i . to see whether GP33 has any roles in the pathogenesis of congenital CMV diseases . There were no statistical differences in the body weight of r33- and Δ33-infected dams ( Fig 5A ) or in the weight of the placentas and fetuses from r33- and Δ33-infected dams ( Fig 5B and 5C ) . However , viral loads in the spleens and livers of the r33-infected dams were significantly higher ( p<0 . 05 ) than those of the Δ33-infected dams , although those in lungs , pancreas , and kidneys did not reach statistical significance ( Fig 5D–5H ) . In addition , viral loads in the placentas of the r33-infected dams were significantly higher ( p<0 . 01 ) than those of the Δ33-infected dams ( Fig 5I ) . Due to the small size of the infected fetuses at 5 weeks’ GA , collection of the fetal organs from most fetuses was impossible . In addition to the higher viral load described above , severe abnormalities were apparent in the lungs , including inflammation and alveolar hemorrhage , of the r33-infected dams ( Fig 6A ) . In addition to the lungs , the livers of the dams infected with r33 showed changes in surface color from red to ochre and in texture from smooth to rough , and one of the livers had a yellowish necrotic lesion ( Fig 6B ) . HE staining analyses demonstrated more frequent infiltration of inflammatory cells into the lungs and livers , especially around the portal regions , of the r33-infected dams ( Fig 7A–7D ) . In addition , immunohistochemical analysis with a monoclonal antibody against a GPCMV early antigen ( g-1 ) detected viral antigens in the liver and lymph nodes of one of the r33-infected dams , although not the dam with the necrotic lesion in the liver ( Fig 7C inset and Fig 7H ) . Importantly , GPCMV antigen-positive cells were detected frequently , especially around white pulps , in the spleens of all 4 dams infected with r33 but in none of the dams infected with Δ33 ( Fig 7F , 7G , 7I and 7J ) . However , the TNF-α mRNA levels in the livers and lungs were similar in r33- and Δ33-infected dams ( means ± SEM of threshold cycle numbers after normalization to β-actin for the livers and lungs were r33 = 30 . 8 ± 0 . 45 vs . Δ33 = 30 . 1 ± 0 . 43 and r33 = 32 . 9 ± 1 . 88 vs . Δ33 = 30 . 6 ± 0 . 44 , respectively ) , probably because the infected cells represented only a fraction of the total cells in those organs . Next , dams ( Hartley ) at 4 weeks’ GA were infected s . c . with GPCMV r33 or with Δ33 ( n = 4 each; 2 x 107 IUs/animal ) and sacrificed at 3 weeks p . i . There were no significant differences in the weights of the dams or of their placentas or fetuses ( Fig 8A and 8B ) . In addition , there were no significant differences in viral loads in any of the examined maternal organs ( except for the livers , p<0 . 05 ) , the placentas , or any of the examined fetal organs , between GPCMV r33- and Δ33-infected animals ( Fig 8C–8I ) . As described below , the viral loads in the salivary glands were too small to evaluate the effect of GP33 on viral replication in the salivary glands . In spite of the negligible differences in viral loads as shown in Fig 8 , more severe abnormalities were apparent in the lungs of the dams infected with r33 than in those infected with Δ33 ( Fig 9A ) . The abnormalities were histologically analyzed , and irrespective of the GP33 status , the lungs of the dams showed infiltration of inflammatory cells and alveolar hemorrhage . However , the number of lesions in the lungs of the dams infected with r33 and with Δ33 were 10 . 75 ± 1 . 53 and 4 . 00 ± 0 . 91 per section , respectively , and this 2 . 7-fold difference was statistically significant ( p<0 . 01 ) ( Fig 9B and 9C ) . In addition , the pathologically obtained number of lesions and the visible lesions on the surface of the lungs scored as described in the Method section showed a good correlation ( correlation coefficient r = 0 . 91 , p<0 . 005 ) ( Fig 9D ) . Virus antigens were below the detection level by immunohistochemistry with monoclonal anti-GPCMV early antigen ( g-1 ) , as viruses are usually cleared from most organs by 3 weeks p . i . Schleiss and his colleagues reported that the lack of the GP1 gene encoding a macrophage inflammatory protein 1 ( MIP-1 ) homolog attenuated in vivo phenotypes of GPCMV and that the MIP-1 homolog signals via CCR1 [30 , 31] . GPCMVΔ9K WT and its derivatives r33 and Δ33 were prepared from pBAC-GPCMVΔ9K that lacks a 9 kb region of the GPCMV genome [32] . At least MIP-1 homolog is encoded in the 9 kb region , although any other ORFs are not identified . To confirm that the lack of the 9 kb region results in the reported attenuated phenotypes and to see whether the lack of the region increases lung and liver pathology of pregnant guinea pigs , dams ( Hartley ) at 4 weeks’ GA were infected s . c . with GPCMVΔ9K WT lacking the 9 kb region or with GPCMV SG , a strain that contains the 9 kb region and has the full capacity to grow in animals [33] ( n = 4 each; 3x106 IUs ) and sacrificed at 6 days p . i . As GPCMV SG was more virulent than GPCMVΔ9K WT in preliminary experiments , animals were infected with about one fourth of the dose used in Fig 5 . We found that GPCMV SG showed more virulent viral growth phenotype , including more severe inflammatory lesions in their lungs ( Fig 10A ) , a severe large necrotic lesion in one of the livers ( Fig 10B ) , and higher viral loads in the livers , lungs , and salivary glands as compared to GPCMVΔ9K WT ( Fig 10D–10F ) . In addition , in a separate experiment , viral genome copies per 106 cells ( Log10 ) in the salivary glands of dams at 4 weeks’ GA ( n = 4 ) that were infected s . c . with GPCMV SG ( 3 x 106 IUs/animal ) and sacrificed at 3 weeks p . i . were 6 . 5 ± 0 . 9 , whereas those of dams infected with the BAC-derived strains were negligible as shown in Fig 8 . Thus , the lack of the 9 kb region attenuated GPCMV infection significantly but it did not increase lung and liver pathology .
In this study , we determined the cDNA structure of GP33 and demonstrated that GP33 activated the CRE- and NFAT- , but not the NFκB-signal pathway . Phylogenetically , GPCMV lies between MCMV and HCMV [33 , 34] , but in the case of GP33 homologs , it is likely that GP33 is functionally closer to HCMV UL33 , as i ) GP33 and UL33 contain the DRY motif , whereas M33 and R33 contain the NRY motif , and ii ) GP33 and UL33 activate the CRE- and NFAT- , but not the NFκB-signal pathway , whereas M33 and R33 activate all three pathways . Although M33 and R33 , but not UL33 , induce migration of smooth muscle cells [17 , 35] , it remains unclear whether GP33 behaves in a similar manner to rodent GP33 homologs or to UL33 . GP33 was dispensable for infection both in tissue cultures and in normal animals . In pregnant animals , however , r33 encoding GP33 induced significantly more frequent lesions and higher viral loads in the maternal organs at 6 days p . i . Potential factors affecting the difference in disease pathogenesis between normal and pregnant animals could include i ) age of animals , ii ) immune responses , iii ) chemokine network , and iv ) routes of viral inoculation . First , it is unlikely that older animals , in this case , the dams , are more susceptible to GPCMV , as younger animals were more prone to body weight loss after virus challenge [8] . While there is a recent study demonstrating functional impairment of CMV-reactive cellular immunity in pregnant women based on an IFN-γ ELISPOT assay [36] , issues surrounding immune responses and susceptibility to infectious diseases during the pregnancy have long been controversial [37] . As the chemokine network in pregnancy significantly contributes to the maintenance of pregnancy , especially to maternal-fetal tolerance and to placentation [38 , 39] , the chemokine profiles specific to the pregnancy may have impact on GP33-mediated pathogenesis . M33 plays an essential role in smooth muscle cell migration in a CCL5-dependent manner , despite its ligand-independent activation of signal pathways in transiently transfection experiments [17] . As an analogy , GP33 could respond to a certain ligand in vivo , such as in the context of pregnancy , in spite of ligand-independent activation shown in this work . Finally , infection via different inoculation routes yields different disease outcomes ( reviewed in [40] ) . For example , CX3CR1 deficiency of mice compromised dissemination of MCMV to the salivary glands from a local footpad inoculation but not from systemic intraperitoneal inoculation [41] . A recent study demonstrated that the lack of M33 hampered viral spread from lymph nodes to the lungs after the i . n . administration of MCMV [22] . One of the limitations of our study is the artificial routes of infection , i . p . for non-pregnant normal animals and s . c . for pregnant animals , while i . n . infection apparently represents the natural infection route . The most striking phenotype for M33 or R33 mutant viruses is the inability to replicate in the salivary glands [14 , 19 , 42 , 43] . However , we could not evaluate whether GP33 is essential for replication in the salivary glands , as the GPCMV strains used in this study showed significantly reduced viral loads in the salivary glands , probably due to the lack of the MIP-1 homolog . Although the lack of the 9 kb region encoding the MIP-1 homolog resulted attenuated phenotypes , it did not increase lung and liver pathology in pregnant animals . Therefore , it is clear that the deletion of GP33 confers additional attenuation that could be observed in pregnant animals . The maternal organs of the dams infected with GPCMV r33 showed higher viral loads and active viral replication at 6 days p . i . and more severe inflammation in the lung at 3 weeks p . i . , although viral loads in the maternal organs were almost undetectable at 3 weeks p . i . Our hypothetical view would be that GP33 transiently increases viral activities , which triggers inflammation in the maternal visceral organs , but the balance between viral activities and viral clearance by immune responses determines pathological outcomes . Further studies on immune correlates will be required . In addition to the maternal organs , higher viral loads in the placentas of the dams at 6 days p . i . suggest that GP33 has some pathological roles in congenital diseases . However , due to the small size of the infected fetuses , we could not obtain organs from a sufficient number of fetuses to analyze fetal transmission at 6 days p . i . In addition , the lack of the difference in the weights of the placenta and fetus as well as in viral loads of any of the examined fetal organs , between r33- and Δ33-infected dams at 3 weeks p . i . , makes it hard to conclude that GP33 is involved in fetal transmission . Not only direct infection of fetuses causes congenital diseases but also infection of the placenta may impair the normal growth of fetuses indirectly , for example , by CMV-induced cytokine modulation [25 , 44] . Also vascular insufficiency due to CMV-mediated injury of endothelial cells could compromise perfusion to the developing fetus or to the placenta , resulting in hypoxia and growth restriction [45 , 46] . Therefore , it would be essential to analyze pup survival and pup viral loads to see whether GP33 contributes to any of the pathogenic aspects of congenital infection . The importance of our study lies in the demonstration of GP33 involvement in the pathological effects in some organs . It is intriguing that GP33 appears to enhance the accumulation of GPCMV-infected cells in the spleen . We hypothesize that an increase of inflammatory cytokines by GP33 allows the infection and migration of monocytes/macrophages in the spleen to the organs for viral dissemination as well as for the induction of pathological lesions . Alternatively , GP33 may facilitate transient inflammation that is induced by the chemokine network specific to the pregnancy . Expression of GP33 , such as by transduction with a recombinant adenovirus vector , in the absence of GPCMV infection or in the presence of Δ33 infection might help the understanding of the observed phenomena . In addition to GP33 , GPCMV encodes GP78 , the UL78 homolog . It would be interesting to see whether the presence of GP78 compensates for the GP33 defect and whether GP33 forms a heterodimer with GP78 or other GPCR homologs for unexpected functions . Although our study had some limitations , our observations regarding the pathogenic role of GP33 during pregnancy warrants further studies on the functions of GPCMV vGPCRs . In conclusion , GP33 , a GPCR encoded by GPCMV , contributes to increasing viral activities transiently in the maternal spleens and livers and in the placentas as well as to inducing inflammation in some maternal organs during pregnancy .
GPL ( CCL-158 , ATCC USA ) were cultured in F-12 medium supplemented with 10% fetal bovine serum ( FBS , HyClone , USA ) and infected with GPCMV . Two days later , RNA samples were extracted from the infected cells with RNeasy Mini Kit ( Qiagen ) . The cDNA was synthesized from the extracted RNA with SMARTer RACE 5´/3´Kit ( Clontech ) . The obtained PCR products were purified using a DNA fragment purification kit ( MagExtractor , TOYOBO , Japan ) , and their nucleotide sequences were determined . The UL33 ORF prepared from HCMV-infected fibroblasts , the ORF predicted initially as GP33 ( GP33s ) , the GP33 ORF in the cDNA obtained by the RACE analyses , and the GP33 ORF with FLAG-tag at the carboxyl end were cloned into pcDNA3 . 0 ( Invitrogen ) , resulting in pcDNA-UL33 , -GP33s , -GP33 and -GP33F , respectively . The GP33 ORF was also cloned between EcoRI and BamHI sites of pEGFP-N1 ( Clontech ) , resulting in pEGFP-GP33 to express GP33-EGFP fusion protein . Cells transfected with pEGFP-GP33 were fixed with paraformaldehyde and observed under a confocal microscope . COS-7 cells ( product no . CRL1651 , ATCC ) in 96-well plates were cultured in Dulbecco’s modified medium supplemented with 10% FBS and penicillin/streptomycin , and transiently transfected with a reporter plasmid [pCRE-luc , pNFAT-luc , or pNFκB-luc ( Clontech ) ] , an internal control plasmid ( pRL-EF1α ) [47] , and an effector plasmid ( pcDNA-GP33s and -GP33 ) or positive control plasmids , by adding 200-fold diluted lipofectamine 2000 ( Invitrogen ) and cultured for 6 h . Total amount of DNA for each condition was adjusted by adding the empty vector . The cells were then cultured in a serum-free medium for 24 h and rinsed with PBS . The cells were treated with a Dual-luciferase reporter assay kit ( Promega ) , and the luciferase activities of the cells were measured by GloMax ( Promega ) . pF4A-CMV-GαqQ209L and a plasmid expressing FLAG-MyD88 , kindly provided by H . Ueda ( Gifu Univ . ) and A . Takaoka ( Hokkaido Univ . ) , respectively , were used as positive control plasmids for the NFAT- and NFκB-pathways , respectively . pBAC-GPCMVΔ9K , a BAC that contains a GFP expression cassette and the GPCMV genome lacking an 8 . 9-kb sequence [32] , was used for the generation of GP33 mutants as follows . GPCMV with a knock-out mutation in the GP33 gene and that with a wild-type sequence were prepared from the same backbone as follows . First , the rpsL-neo cassette was amplified by PCR using primers GP33rpsL-F and GP33rpsL-R ( S1 Table ) that add a 50-bp GPCMV sequence flanking the GP33 ORF and used for recombination with the GPCMV region 54 , 807–56 , 090 ( the numbers are based on Genbank acc . no . AB592928;[33] ) of pBAC-GPCMVΔ9K using the RedET recombination system ( GeneBridges ) , resulting in pBAC-GPCMVΔ9K::rpsL-neo/GP33 . A fragment containing the GPCMV region 54 , 698–56181 was amplified by PCR with primers GP33BS-F and -R , and cloned into the BamHI and EcoRI sites of pBluescriptII KS ( + ) . Next , the resultant plasmid was mutagenized using a commercial kit ( QuickChange Lighting , Agilent ) with the primers GP33KO-F and -R ( S1 Table ) to introduce a “TGAT” sequence , which inserts a stop-codon mutation at amino acid position 30 , accompanied by a frame-shift mutation , into the GP33 ORF ( Fig 1A ) . The rpsL-neo cassette in pBAC-GPCMVΔ9K::rpsL-neo/GP33 was replaced with the GPCMV sequences containing the mutated “TGAT” or wild-type sequence by electroporation of the PCR fragments ( 54 , 698–56 , 179 ) amplified from the plasmids using primers GP33-F and -R ( S1 Table ) , resulting in pBAC-GPCMVΔ9K-r33 and -Δ33 . GPCMV WT , r33 and Δ33 , were recovered by transfection of BAC DNAs , the parental pBAC-GPCMVΔ9K and the manipulated pBAC-GPCMVΔ9K-r33 and -Δ33 , respectively , into GPL cells . After observation of plaque formation , the recovered viruses were amplified a couple of times by the co-culturing of infected cells with uninfected cells . Finally , cell-free virus stocks were prepared and concentrated by ultracentrifugation ( 20 , 000 g for 2 h ) in a 20% sucrose step gradient as described previously [32] . RFLP analyses of the GPCMV-BACs with EcoRI and HindIII demonstrated no apparent differences in the genome structures among GPCMV WT , r33 and Δ33 . In addition , the integrity of the locus encoding GP129 , GP131 , and GP133 was confirmed . As the viral stocks prepared from the two BAC clones for each strain behaved similarly in the following experiments , the results obtained from only one of them are shown . Titers of the GPCMV-BAC stocks were determined by counting GFP-positive foci after infection of GPL cells . Although the infectivity depended on cell type , we used those titers to calculate multiplicities of infection ( MOIs ) . Four- or 5-week-old female normal guinea pigs ( strain Hartley ) and pregnant guinea pigs at 4 weeks’ GA were all purchased from a commercial breeder ( SLC Inc . , Hamamatsu , Japan ) . Since the company allows mating for two weeks , 4 weeks’ GA means 4 ± 1 weeks’ GA . Animals were randomly grouped based on the body weights . Blood specimens were drawn from ear veins using heparinized micro-hematocrit capillaries ( Harschmann , Germany ) for serological tests . All tested animals were seronegative for GPCMV . Normal and pregnant animals were infected with GPCMV i . p . and s . c . , respectively . Animals were euthanized at the indicated days p . i . The visible lesions on the surface of the lungs were scored blindly from “1” to “10” based on numbers and sizes of lesions , by 5 lab members who worked in the bacteriology field and were not involved in this study . Guinea pig epithelial cell clone GPE-7 was established by immortalization of primary renal cells obtained from guinea pigs with SV40 T antigen as described previously [48] . Splenic monocytes were prepared and treated with 100 nM phorbol 12-myristate 13-acetate ( PMA ) for 2 days to obtain macrophages as described previously [32] . Total RNA samples were prepared from infected GPL cells at 24 hrs p . i . or from guinea pig tissues using a commercial kit ( RNeasy Plus Mini kit , Qiagen ) , and reverse-transcribed using PrimeScript RT master mix ( Takara Bio ) . Real-time PCR of the reverse transcribed products was performed using a commercial kit ( GoTaq qPCR Master Mix , Promega ) with primers amplifying a region containing an exon junction of the guinea pig TNF-α or β-actin gene ( S1 Table ) . Total DNA samples were prepared using Maxwell 16 Tissue DNA purification kit ( Promega ) , and the quantification of cellular β-actin and GPCMV GP83 genes was performed as described previously [8] . All organs obtained from sacrificed animals were fixed in 10% buffered formalin . Formalin-fixed specimens were embedded in paraffin , sectioned , and stained with hematoxylin and eosin ( HE ) , as described previously [8] . Serial sections were used for HE staining and immunohistochemistry . Immunohistochemical analysis was performed using the monoclonal antibody g-1 , which detects a GPCMV early antigen [5] , as a primary antibody . For the second- and third-phase immunostaining reagents , a biotinylated F ( ab ) 2 fragment of anti-mouse immunoglobulin ( DAKO ) and peroxidase-conjugated streptavidin ( DAKO ) were used . DAB was used as a chromogen , and the slides were counterstained with hematoxylin . Statistical analyses were performed using GraphPad Prism ver . 7 ( GraphPad Software , Inc . , CA , US ) . The statistical significances of differences in the weights and viral loads between the two groups were calculated using the Mann-Whitney U test , with values of p<0 . 05 considered significant . The correlation between the two types of lesion scores was evaluated by Pearson correlation coefficient ( r ) . All animal procedures , including euthanasia using CO2 , described in this study were approved by the Animal Care and Use Committee of Gifu Pharmaceutical University ( approval no . 2016–132 , -220 , -320 , 2017–021 , -079 , -214 , 2018–117 ) . The same procedures were also approved by the Animal Care and Use Committee of Gifu University ( approval no . 28–57 , -75 , -96 , 29–37 , -49 , -91 , 30–94 ) , as we used a P2 animal facility in the Life Science Research Center of Gifu University for infection of guinea pigs with GPCMV . All of the above mentioned animal procedures are in accordance with the following Japanese laws and guidelines: “Action on Welfare and Management of Animals” ( law ) and “Fundamental Guidelines for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions under the jurisdiction of the Ministry of Education , Culture , Sports , Science and Technology” ( guidelines ) , “Act on the Conservation and Sustainable Use of Biological Diversity through Regulations on the Use of Living Modified Organisms” ( law ) , and “Act on the Prevention of Infectious Diseases and Medical Care for Patients with Infectious Diseases” ( law ) . | Cytomegalovirus ( CMV ) is a major pathogen that causes congenital diseases , including birth defects and developmental abnormalities in newborns . Better understanding of the immune evasion mechanisms may open the way to the development of new types of live attenuated vaccines for congenital CMV infection . In contrast to murine and rat CMVs , guinea pig CMV ( GPCMV ) causes infection in utero , which makes GPCMV animal models a useful tool for understanding the pathogenesis of congenital infection and evaluation of vaccine strategies . By constructing a GPCMV mutant lacking GP33 , a viral G protein-coupled receptor homolog , this study found that GP33 was involved in the induction of significant inflammatory responses in pregnant but not in normal animals . As GP33 activated the NFAT- and CRE- , but not the NFκB-signal pathway , it is plausible that GP33 enhanced cytokine expression , which results in pathogenic outcomes in the maternal organs and placentas . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
]
| [
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"immune",
"physiology",
"cytokines",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"spleen",
"immunology",
"microbiology",
"vertebrates",
"cytomegalovirus",
"infection",
"animals",
"mammals",
"animal",
"models",
"developmental",
"biology",
"signs",
"and",
"symptoms",
"experimental",
"organism",
"systems",
"molecular",
"development",
"viral",
"load",
"digestive",
"system",
"research",
"and",
"analysis",
"methods",
"embryology",
"infectious",
"diseases",
"inflammation",
"placenta",
"animal",
"studies",
"exocrine",
"glands",
"immune",
"response",
"guinea",
"pigs",
"immune",
"system",
"rodents",
"eukaryota",
"diagnostic",
"medicine",
"anatomy",
"virology",
"physiology",
"salivary",
"glands",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"amniotes",
"organisms"
]
| 2018 | Roles of GP33, a guinea pig cytomegalovirus-encoded G protein-coupled receptor homolog, in cellular signaling, viral growth and inflammation in vitro and in vivo |
With the rising development of bacterial resistance the search for new medical treatments beyond conventional antimicrobials has become a key aim of public health research . Possible innovative strategies include the inhibition of bacterial virulence . However , consideration must be given to the evolutionary and environmental consequences of such new interventions . Virulence and cooperative social behaviour of the bacterium Pseudomonas aeruginosa rely on the quorum-sensing ( QS ) controlled production of extracellular products ( public goods ) . Hence QS is an attractive target for anti-virulence interventions . During colonization , non-cooperating ( and hence less virulent ) P . aeruginosa QS-mutants , benefiting from public goods provided by wild type isolates , naturally increase in frequency providing a relative protection from invasive infection . We hypothesized that inhibition of QS-mediated gene expression removes this growth advantage and selection of less virulent QS-mutants , and maintains the predominance of more virulent QS-wild type bacteria . We addressed this possibility in a placebo-controlled trial investigating the anti-QS properties of azithromycin , a macrolide antibiotic devoid of bactericidal activity on P . aeruginosa , but interfering with QS , in intubated patients colonized by P . aeruginosa . In the absence of azithromycin , non-cooperating ( and hence less virulent ) lasR ( QS ) -mutants increased in frequency over time . Azithromycin significantly reduced QS-gene expression measured directly in tracheal aspirates . Concomitantly the advantage of lasR-mutants was lost and virulent wild-type isolates predominated during azithromycin treatment . We confirmed these results in vitro with fitness and invasion experiments . Azithromycin reduced growth rate of the wild-type , but not of the lasR-mutant . Furthermore , the lasR-mutant efficiently invaded wild-type populations in the absence , but not in the presence of azithromycin . These in vivo and in vitro results demonstrate that anti-virulence interventions based on QS-blockade diminish natural selection towards reduced virulence and therefore may increase the prevalence of more virulent genotypes in the Hospital environment . More generally , the impact of intervention on the evolution of virulence of pathogenic bacteria should be assessed . Trial Registration: ClinicalTrials . gov NCT00610623
Anti-virulence therapies have been recently suggested as alternative strategies to circumvent the growing problem of antibiotic resistance [1] , [2] . In P . aeruginosa inhibition of Quorum-Sensing ( QS ) seems particularly attractive as QS regulates many virulence determinants of this pathogen [3] . Azithromycin is a widely used macrolide antibiotic without significant bactericidal activity on P . aeruginosa [4] . Recent studies suggest azithromycin might be of benefit against this bacterium because it interferes with the QS-circuit and thereby inhibits the expression of a wide range of extracellular virulence factors [5] . Inhibition of QS is likely to have important evolutionary consequences for P . aeruginosa . Both in vitro and in vivo studies suggest that mutants ( QS-cheats ) that don't respond to QS ( specifically , mutants that are defective in one of the QS-receptors , LasR ) can have a selective advantage in the presence of QS-wildtypes [6] , [7] . This has been recently demonstrated during colonization of untreated colonized patients in whom QS-cheats accumulated over time [8] . The most likely explanation for this advantage is that the mutants exploit the wild type public goods , without paying the metabolic cost of their production [9]–[11]; although other direct costs of QS in clinical contexts can't be ruled out [12]–[14] . Regardless of the reasons why QS-mutants have a fitness advantage , this advantage is unlikely to be realised if QS is blocked in wild type bacteria . Azithromycin ( or any QS-blocker ) will therefore reduce , or remove , selection for less virulent QS-cheats and maintain the predominance of more virulent QS-wild type bacteria . We tested this hypothesis by following the evolutionary dynamics of QS ( lasR ) mutants in intubated patients colonised by P . aeruginosa , during a placebo controlled clinical trial evaluating the prevention of pneumonia by azithromycin . Whereas the proportion of lasR mutants rapidly increased in the untreated control patients , the proportion did not change in the azithromycin-treated patients . This fitness advantage in the absence , but not the presence , of azithromycin was similarly observed in vitro . More generally , the impact of intervention on the evolution of virulence of pathogenic bacteria should be assessed [15] .
We monitored QS-gene expression directly in tracheal aspirates to document the “in patient” QS-inhibition by azithromycin . Azithromycin significantly reduced the expression of both QS-circuit ( lasI; Mann-Whitney test , P = 0 . 006 ) as well as QS-target ( rhlA; P = 0 . 005 ) genes , whereas it did not affect expression of the QS-independent gene trpD ( P>0 . 2 ) ( Figure 2 ) . It is of course possible that azithromycin inhibited the expression of some other genes unrelated to QS . However microarray data have shown that QS-regulated genes were among those whose expression was most severely affected by azithromycin [16] . We determined the evolution of P . aeruginosa QS in patients by first measuring the production of elastase , which is under the control of the lasR QS-system [17] from the 650 collected isolates ( data not shown ) . Variations in elastase activity correlated with mutations in lasR between independent wild type and mutant lasR alleles ( Mann-Whitney: P<0 . 001 ) , as determined by sequencing this gene in the first isolate obtained from each patient , and then in subsequent isolates presenting a different QS-phenotype ( Figure 3 ) . Mutations in lasR were therefore reducing the expression of elastase , and by inference , other lasR-regulated genes . Whereas the proportion of lasR mutants significantly increased through time in the 31 control group patients ( Figure 3a; logistic regression: F1 , 10 = 65 . 36 , P<0 . 001 ) , there was a small , decline in the proportion of lasR mutants in the 30 azithromycin treated patients ( Figure 3a; F1 , 10 = 32 . 58 , P<0 . 001; test of whether slopes differ ( treatment by time interaction ) in full model: F1 , 20 = 77 . 6; P<0 . 001 ) . In agreement with this observation , isolates from placebo patients showed decreasing mean elastase levels in vitro ( Figure 3b; F1 , 10 = 41 . 12 , P<0 . 001 ) , while isolates from the azithromycin treated group showed a concomitant increase through time ( Figure 3b; F1 , 10 = 41 . 12 , P<0 . 001; treatment by time interaction in full model: F1 , 20 = 26 . 46 , P<0 . 001 ) . These data are consistent with the hypothesis that azithromycin treatment removes any advantage of QS-mutants , because QS is blocked in the wildtype population . There are , however , a number of alternative explanations , particularly as bacterial densities , based on mean P . aeruginosa genomic copy numbers , were twice as large in the placebo compared to the azithromycin group ( 7×106 and 1 . 2×107; t = 1 . 96 , P = 0 . 06 ) . First , azithromycin-imposed reduction in densities could reflect reductions in growth rate , and this could simply have slowed down the rate at which lasR mutants change in frequency . We can however rule this out as a primary explanation for our data , because azithromycin did not only cause a quantitative change in the frequency of lasR mutants , but also a qualitative change: lasR mutants decreased in frequency during azithromycin treatment , whereas they increased in the placebo group ( Figure 3 ) . Second , it is possible that azithromycin may reduce selection for lasR mutants if reductions in QS-mediated public goods production results from reductions in bacterial density caused by azithromycin . Third , azithromycin may directly inhibit the growth of lasR mutants more than wildtype bacteria , explaining why there was a small reduction in the frequency of lasR mutants following azithromycin treatment . We address these possibilities below . Furthermore we cannot exclude that azithromycin influenced the structure of the resident bacterial flora of the patients which could in turn influence the P . aeruginosa population and its virulence properties [18] . To aid the interpretation of the clinical data , we carried out in vitro experiments with wild type P . aeruginosa ( PAO1 ) and an isogenic lasR mutant in the presence and absence of azithromycin . Danesi et al . [19] measured azithromycin concentrations of 9 mg/kg in lung tissue of patients receiving a comparable dosing regimen as those in our study , hence we used similar concentrations ( 5 and 10 mg/l ) for our in vitro experiments . We first measured growth rates in media where the primary nutrient source is protein ( BSA ) , making lasR-controlled expression of proteases necessary for the production of useable amino acids [6] . Consistent with previous studies [6] , growth rate of the lasR mutant in monoculture was reduced relative to the wildtype ( by approximately 50% ) in the absence of azithromycin , demonstrating an advantage of QS in this environment ( Figure 4a; 2-sample t-test: P<0 . 05 ) . The addition of azithromycin reduced densities of both genotypes ( linear effect of azithromycin: F1 , 29 = 71 . 2 , P<0 . 001 ) , but this reduction was much greater for the wildtype than the lasR mutant ( interaction between concentration and genotype: F1 , 29 = 6 . 92 , P = 0 . 013 ) , confirming a role of azithromycin in suppressing the production of QS-controlled exoproducts . Given that azithromycin inhibits wildtype growth more than that of the lasR mutant , we are unable to explain the slight drop in the frequency of lasR mutants in the patients . We next measured the fitness of the lasR mutant invading wildtype populations ( 1∶100 ratio ) . Consistent with the in vivo results , we found that the fitness advantage of lasR mutants was decreased with increasing azithromycin concentration ( Figure 4b; Linear effect of azithromycin concentration of fitness of lasR mutant: F1 , 16 = 48 . 41 , P<0 . 001 ) . Unlike in the clinical context , the lasR mutant still had a slight fitness advantage over the wildtype at the highest concentration of azithromycin used ( 10 mg/l ) , suggesting , unsurprisingly , that additional variables influence the fitness of lasR mutants in vivo . These results strongly suggest that the major advantage of lasR mutants in vitro ( and presumably in vivo ) is their ability to exploit wildtype public goods [8] , and that azithromycin removes this advantage because less public good ( elastase ) is produced . However , the data do not distinguish between azithromycin directly inhibiting elastase production , or indirectly through density reductions , or both . We have shown that azithromycin treatment can prevent selection for lasR mutants , and consequently increase the proportion of wild type P . aeruginosa in colonized patients . A number of not mutually exclusive factors may help to explain this pattern , but the data suggests that the primary reason is because azithromycin blocks QS . Blocking QS prevents lasR mutants from exploiting the public goods provided by the wildtype , and reduces any direct costs associated with QS , such as production of extracellular products that are not of benefit in this particular environment [8] . Both in vitro and in patient data obtained from the clinical trial suggest a key role of QS-dependent virulence for the development of infection ( Köhler et al . , submitted ) , and support the general consensus from studies using animal models showing that QS-expression ( and public goods production in general ) is associated with increased virulence [2] , [6] , [20]–[23] . Azithromycin is therefore likely to be of clinical benefit to a treated patient by inhibiting the QS-dependent virulence during the course of the treatment . However , when treatment is discontinued the patient is at risk of being colonized by highly virulent bacteria , with the potential of late onset infections . Moreover a wider use of such anti-virulence interventions may also increase the prevalence of highly virulent QS-wild type isolates within the hospital . More generally , any intervention that reduces bacterial densities is also likely to result in a reduced selective advantage of less virulent mutants that do not make public goods [24] . The study highlights the need to carefully consider both the short and longer term implications of anti-virulence therapy and other interventions ( such as vaccines [15] ) on pathogen virulence .
We obtained approval for this study by the “Commission Centrale d'Ethique de la Recherche sur l'Etre Humain des Hôpitaux Universitaire de Genève” . Written informed consent from all patients or their legal representatives was obtained according to legal and ethical considerations . This randomized , placebo-controlled , double blind study ( ANB 006#2001 , ClinicalTrials . gov ID#NCT00610623 ) was designed to assess the efficacy of azithromycin as a quorum-sensing inhibitor in preventing the occurrence of P . aeruginosa pneumonia in ventilated patients with documented colonization . Twenty-one European centers participated in this trial; eight in France , four in Spain , three in Belgium , three in Poland , two in Serbia and one in Switzerland . We screened mechanically-ventilated patients for respiratory tract colonization by P . aeruginosa every 48 hours . Neutropenic patients and patients treated with immunosuppressive drugs were not eligible . Patients with ongoing P . aeruginosa infection , having received macrolides or antibiotics active against the colonizing P . aeruginosa isolate during the last 14 days were excluded . Patients with proven colonization by P . aeruginosa were randomized ( D-1 ) and received either placebo or 300 mg per day iv azithromycin in a double blind fashion for a maximum of 20 days ( D1 to Dx ) . During the study , only the administration of antibiotics inactive against P . aeruginosa was allowed . Detailed information on study design is available in supporting information Protocol S1 and Checklist S1 . Patient characteristics and clinical outcome of the study are published elsewhere ( van Delden et al . , submitted ) . Starting the first day of proven colonization ( D-1 ) , we collected tracheal aspirates ( 0 . 3 to 5 ml ) and one P . aeruginoa isolate ( collection period: 3–20 days ) at 24 hours intervals . Samples were frozen at −80°C on site within 15 minutes , and sent on dry ice to the reference research laboratory at the University Hospital Geneva , where all analyses were performed in a blind fashion . The logit-transformed proportion of patients whose isolate was a lasR mutant was analyzed by logistic regression , with time ( a covariate ) , treatment ( placebo or azithromycin ) and the interaction fitted in GenStat v10 . Data were over dispersed , so a scaling factor to equalize the residual error and degrees of freedom was employed . From prospectively collected tracheal aspirates we extracted total genomic DNA and total RNA ( for details see supporting information Text S1 ) . We detected ( >104 genomic copies/g aspirate ) P . aeruginosa DNA in 98% of the aspirates , confirming the colonization by this organism . In the RNA extractions , we detected expression ( >5×104 copies/g aspirate ) of the rpsL housekeeping gene in 80% of the aspirates . This indicates that quality of both sample handling and RNA extractions were sufficient to detect bacterial gene expression in the majority of the tracheal aspirates . As a second control for the quality of the RNA extracts from clinical samples we plotted the amount of P . aeruginosa bacteria as determined by qRT-PCR from the genomic DNA extractions against the expression of the rpsL housekeeping gene . We observed a good correlation between these two variables ( r = 0 . 69 , P<0 . 001 ) . The number of P . aeruginosa in aspirates was determined by qRT-PCR of genomic DNA preparations . Aliquots of genomic DNA preparations were diluted 10 fold into H2O and 3 µl of this dilution were added to the PCR reaction mix containing 1× Quantitect Sybr Green Master Mix and 600 nM primers in a total volume of 15 µl . PCR conditions were as described below for cDNA analysis . A standard curve was obtained by addition of 10-fold dilutions of a P . aeruginosa culture to an aspirate collected from a patient not colonized by this organism . Genomic DNA was then isolated as described above and quantified by qRT-PCR . Under these conditions , we detected 104 CFU/g aspirate . Standard curves yielded reproducible values during the 3-month analysis period . P . aeruginosa was found in the aspirates at levels varying from 4×104–1 . 8×108 CFU/g . P . aeruginosa strain PAO1 was competed against a rare invading isogenic lasR knockout mutant [25] in 200 µl M9 minimal salts medium supplemented with 1% BSA [6] in 96-well plates , shaken at 200 rpm at 37°C in the presence or absence of azithromycin ( 5 and 10 mg/l ) for 72 hours . Six wells per environment were inoculated with 107 cells of overnight cultures ( grown in LB medium at 37°C ) , at a ratio of 1∶100 lasR mutant: wild type . Selection coefficients of the lasR mutants was calculated as the differences in malthusian parameters ( ln ( final density/starting density ) as previously described [26] , with cell counts determined by plating on LB agar and LB supplemented with 50 mg/l tetracycline . A selection coefficient of zero indicates that strains have equal fitness . Selection coefficients were regressed against azithromycin concentration . Densities ( colony forming units ) of pure cultures ( 6 replicates per treatment ) under the same conditions were determined at the same time . Densities were log10-transformed , to meet assumption of general linear models , and concentration ( a covariate ) , strain and the interaction fitted in GenStat . | With the rising development of antibiotic resistance and rapid spread of nosocomial pathogens , the search for new treatments beyond conventional antibiotics becomes a key aim of public health research . As such , anti-virulence therapies might be alternative antimicrobial strategies . However , consideration must be given to the potential evolutionary and environmental consequences of such interventions . Here we demonstrate a significant evolutionary impact of an anti-virulence intervention . Virulence and cooperative social behaviour of Pseudomonas aeruginosa rely on the quorum-sensing ( QS ) controlled production of extracellular products . In the absence of a specific intervention , non-cooperating ( and hence less virulent ) QS-mutants exploit and benefit from products provided by wild type isolates . As a consequence these less virulent mutants increase in frequency and provide a relative protection against infection to a colonized patient . In contrast , when QS-gene expression is reduced by the QS-inhibiting drug azithromycin , this advantage of QS-mutants is lost and virulent isolates predominate both in colonized patients and during in vitro experiments . These results suggest that QS-blockade may increase the prevalence of more virulent QS-responders among colonizing isolates in the hospital environment . More generally , the impact of anti-virulence interventions on the ecology and evolution of virulence of pathogenic bacteria needs to be assessed . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
]
| [
"evolutionary",
"biology/microbial",
"evolution",
"and",
"genomics",
"ecology/evolutionary",
"ecology",
"infectious",
"diseases/nosocomial",
"and",
"healthcare-associated",
"infections",
"microbiology/microbial",
"evolution",
"and",
"genomics",
"microbiology",
"infectious",
"diseases/respiratory",
"infections",
"infectious",
"diseases/bacterial",
"infections",
"ecology/population",
"ecology",
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance"
]
| 2010 | Quorum Sensing Inhibition Selects for Virulence and Cooperation in Pseudomonas aeruginosa |
Biomarker discovery aims to find small subsets of relevant variables in ‘omics data that correlate with the clinical syndromes of interest . Despite the fact that clinical phenotypes are usually characterized by a complex set of clinical parameters , current computational approaches assume univariate targets , e . g . diagnostic classes , against which associations are sought for . We propose an approach based on asymmetrical sparse canonical correlation analysis ( SCCA ) that finds multivariate correlations between the ‘omics measurements and the complex clinical phenotypes . We correlated plasma proteomics data to multivariate overlapping complex clinical phenotypes from tuberculosis and malaria datasets . We discovered relevant ‘omic biomarkers that have a high correlation to profiles of clinical measurements and are remarkably sparse , containing 1 . 5–3% of all ‘omic variables . We show that using clinical view projections we obtain remarkable improvements in diagnostic class prediction , up to 11% in tuberculosis and up to 5% in malaria . Our approach finds proteomic-biomarkers that correlate with complex combinations of clinical-biomarkers . Using the clinical-biomarkers improves the accuracy of diagnostic class prediction while not requiring the measurement plasma proteomic profiles of each subject . Our approach makes it feasible to use omics' data to build accurate diagnostic algorithms that can be deployed to community health centres lacking the expensive ‘omics measurement capabilities .
The aim of biomarker discovery is to find small subsets of measurements in ‘omics data that correlate with the clinical syndromes or phenotypes of interest . Despite the fact that most clinical phenotypes ( e . g . diseases ) are characterized by a complex set of clinical parameters , with a variable degree of overlap , current computational approaches do not take into consideration the multivariate nature of the phenotypes . The challenges arise both from the dependence of the diseases on several proteins and from the complexity of the symptoms . To overcome this limitation , in our framework , the data to be analysed is represented by two views , namely a plasma proteomics profile , and a set of clinical data composed of patient history , signs , symptoms and clinical laboratory measurements of the individuals with syndromes of interest . This type of problem can be described as multivariate in both the views , and the aim is to discover a sparse set of ‘omic variables ( proteomic-biomarkers ) that correlates with a combination of clinical variables ( clinical-biomarkers ) . Given the typically high number of variables and small number of patient samples in clinical ‘omic studies , dimensionality reduction techniques such as Principal component analysis ( PCA ) and Canonical Correlation Analysis ( CCA ) have become popular . PCA allows one to discover a set of latent variables in the data that explain most of the variance but they may not correlate with the clinical syndrome of interest . In contrast , CCA performs dimensionality reduction for two co-dependent datasets simultaneously so that the latent variables extracted from the two datasets are maximally correlated . Thus , the latent variables computed from one of the datasets can be used to predict the ones computed from the other , which is the basic goal in biomarker discovery . However , in both PCA and CCA , the latent variables depend on all variables and therefore hinder clinical interpretation and biomarker discovery and validation . To address these computational limitations sparse variants of PCA ( SPCA ) and CCA ( SCCA ) have been independently developed [1]–[3] . These methods use a L1-norm penalization to variable weights which favours models with a small set of variables having a non-negligible weight . Sparse approaches impose the penalisation in both views of the data , thus generating latent variables depending on small sets of variables [2] , [3] . Recently we have proposed an asymmetrical algorithm that imposes sparsity only in one of the data views [1] by penalizing dual variables related to the latent variables of the other dataset . This asymmetrical approach favours latent variables that relate to small clusters of data points . Here we use our asymmetrical SCCA algorithm for the unsupervised discovery of candidate biomarkers by correlating plasma proteomics data to multivariate clinical parameters from two human infectious diseases namely , tuberculosis and malaria , which present with overlapping complex clinical phenotypes in the affected host . Tuberculosis is the leading bacterial cause of death worldwide , with an estimated 8 . 8 million new cases of active disease and 1 . 6 million deaths per year [4] . The global burden of TB occurs in a background of complex disease phenotypes that range from the presence of latent TB to respiratory and constitutional symptoms overlapping with those of pulmonary active TB . Latent TB infection is thought to affect one third of the world's population and a higher proportion of the population of TB-endemic areas [5] . In this scenario the challenge is to distinguish symptomatic patients with active TB from those with latent disease but whose presenting symptomatology is attributable to some other infectious or inflammatory process [6] . Cerebral malaria ( CM ) and severe malarial anemia ( SMA ) are the major severe disease syndromes in African children with a high level of mortality in the under-five age group . The current WHO case definitions for severe malaria combine P . falciparum blood stage parasitemia with coma , severe anemia or respiratory distress [7] , and it is well documented that there is significant overlap across these syndromes [8] . To validate the biomarkers discovered by the SCCA approach , we study the prediction of diagnostic classes using the biomarkers . In particular we study a scenario were the expensive proteomics data is only available during the training time of the models , whilst in prediction time , clinical data and a previously learned biomarker model is available . In our belief this is a realistic setup considering possible real-world deployment of decision support systems into resource-poor health care centres .
The active TB dataset [6] consists of 412 patient data with three datasets: serum proteome profiles measured by SELDI-ToF mass-spectrometry [6] , [9] ( 270 variables ) , clinical data ( 19 variables ) and diagnostic classes ( Active TB , Symptomatic Control , Asymptomatic Control ) . The childhood severe malaria dataset consist of 944 patient data with three datasets: plasma proteome profiles measured by mass-spectrometry ( 774 variables ) , clinical data ( 57 variables ) and diagnostic classes ( Cerebral Malaria ( CM ) , Severe Malaria Anaemia ( SMA ) , Uncomplicated Malaria ( UM ) , Disease Control ( DC ) and Community Control ( CC ) ) . Plasma was subjected to high-throughput proteomic profiling by mass spectrometry as previously described [6] , [9] . The proteomics and clinical variables were standardized by subtracting the mean and dividing by standard deviation . Standardized proteomics and clinical profiles were converted to unit norm vectors by dividing by the Euclidean norm ( Figure 1 ) . Figure 1 shows a workflow of the algorithmic and experimental framework used for biomarker extraction . We assume data in two views , represented by da x m matrix Xa = ( xa ( 1 ) , … , xa ( m ) ) and db x m matrix Xb = ( xb ( 1 ) , … , xb ( m ) ) , where da and db are the dimensions of the feature vectors in the two views . The i'th example is thus given by a pair ( xa ( i ) , xb ( i ) ) . Canonical correlation analysis ( CCA ) is a family of statistical algorithms designed to situations where there are two available views or measurements of the same phenomenon and the goal is to find latent variables that explain both views ( ‘the generating model’ ) [10] . The CCA algorithm aims to find projection directions wa and wb that maximize the correlation of the projected data , the scores sa ( i ) = waTxa ( i ) and sb ( i ) = wbTxb ( i ) , in the two views:where Cab = XaXbT is the empirical covariance matrix of the views over the dataset . The first expression gives the correlation computed with explicit feature weights . The second expression gives the dual representation in terms kernel matrices Ka = XaT Xa and Kb = XbT Xb , as well as dual variables α = ( α1 , … , αm ) and β = ( β1 , … , βm ) giving weights to the examples in the two views . The corresponding projection directions are given in the dual representation by linear combinations of examples , wa = Xaα and wb = Xbβ . When CCA is performed in dual representation , it is commonly called Kernel Canonical Correlation Analysis ( KCCA ) . CCA can be applied to the data iteratively via deflation , that is , the projection of the data to the orthogonal complement of wa and wb , respectively , and performing the CCA analysis for the projected data . This process results in a sequence of projections ( wa ( 1 ) , wb ( 1 ) ) , ( wa ( 2 ) , wb ( 2 ) ) , … , ( wa ( r ) , wb ( r ) ) , where r denotes the minimum of the numerical ranks of Xa and Xb . The resulting projections are orthogonal to each other: wc ( i ) Twc ( j ) = 0 if i≠j , where by c we denote either of the two views a or b . We call the pair ( wa ( 1 ) , wb ( 1 ) ) , the leading pair of projections . By definition , the leading pair has the maximum canonical correlation in the sequence . The sequence of CCA projections is compactly represented by the matrices Wa = ( wa ( 1 ) , … , wa ( r ) ) and Wb = ( wb ( 1 ) , … , wb ( r ) ) . Sparse canonical correlation analysis differs from KCCA in that it aims to find sparse projection directions , those that only depend on a small number of variables . We use an asymmetric formulation of SCCA [1] , where sparse projection directions are aimed for in one view , while in the other view dual sparsity is aimed for , meaning that each projection direction can be expressed as a linear combination of a small number of examples . The optimization to be solved is the following:where w denotes the projection direction in the first view , and e denotes a vector of dual coefficients , where the coefficient ek , k = 1…m , denotes the weight of xb ( i ) . The first term of the objective aims to make the projection scores sa = waTXa and sb = Kbe to match , the second term penalizes the feature weights in the first view by 1-norm ( thus imposing sparsity ) , and the final term penalizes the dual variables by 1-norm ( thus imposing dual sparsity ) . The infinity norm in the constraint ensures that at least one example will have non-zero dual coefficient . This is achieved by fixing one example , indexed by k , to have coefficient ek = 1 and allowing the rest of the coefficients el , l≠k vary . The fixed example is called the seed example . The SCCA hyperparameters μ and γ control the balance between primal and dual sparsity , in the respective views . We used μ = 1 and γ = 1 in our experimental framework . SCCA can be applied iteratively to extract a series of projection directions by using one of two alternative approaches . In the deflation approach , like in CCA , deflation is used to arrive at a sequence ( w ( 1 ) , e ( 1 ) ) , ( w ( 2 ) , e ( 2 ) ) , … , ( w ( r ) , e ( r ) ) of r pairs of projection directions that are orthogonal to each other . In other approach , by selecting a set of different seed examples {s1 , … , sk} one obtains a set of ( w ( 1 ) , e ( 1 ) ) , ( w ( 2 ) , e ( 2 ) ) , … , ( w ( k ) , e ( k ) ) projection directions , which , in contrast to the deflation approach , are in general not orthogonal . The rationale of not requiring orthogonality of the components is two-fold: First , the deflation approach to generate orthogonal components is an order of magnitude slower method , due to the need to search for the best seed after each deflation operation . Secondly , when using the projection directions as input features to classification , there is no obvious benefit from the orthogonality . The non-deflating approach produces a set of latent variables that may have some redundancy due to near collinearity , however , machine learning methods such as SVM are not expected to be hampered by this . Indeed , in our tests , the SVM accuracy on SCCA features with and without deflation was very similar ( data not shown ) . Without loss of generality , we assume that the sequence is sorted so that the pair ( w ( 1 ) , e ( 1 ) ) has maximal canonical correlation in the sequence . Analogously to CCA , this pair is called the leading pair . In our experimental framework we selected seed examples by k-means clustering of the clinical data and choosing the cluster centers as the seeds . The value k = 3 was used for number of clusters in the biomarker extraction experiment and the value k = 30 in the diagnostic class prediction experiment . Figure 2 shows the workflow in diagnostic class prediction . We assume as input clinical data and the biomarker model built by canonical correlation analysis . In the experiments we use a model with 60 projection directions , 30 from SCCA and 30 projection directions from KCCA . The clinical data is projected onto the biomarker model Wb to obtain biomarker scores zb ( i ) = WbTxb ( i ) for each example . The Radial Basis Function ( RBF ) kernelis used as a non-linear transformation of the biomarker scores . We trained a Support Vector Machine ( SVM ) to predict diagnostic classes by using the RBF kernel as input and the predefined diagnostic classes as the output . As the SVM requires a binary classification problem , we use a one-against-one scheme to train a separate classifier for each pair of diagnostic classes . For evaluation , 10-fold cross validation was used . The RBF kernel parameters and the margin softness of SVM were tuned by internal 10-fold cross-validation in each training fold . We compared the CCA-based methodology to using raw clinical data as the input to SVM without making use of the biomarker model and using proteomics data processed with principal component analysis . We used 30 first principal components as the input to SVM . Statistical significance of the results was estimated using randomization tests . In randomization , a background data distribution consistent with the null hypothesis is generated by simulation , where the statistical connection to be tested has been broken , but the data distribution is otherwise kept close to the original data . In the case of canonical correlation analysis , we want to test if the correlation of the two views is significant , we use the null hypothesis H0: The data Xa and Xb are not correlated” . To generate a sample of canonical correlation values consistent with the null hypothesis , we generated randomized versions of the data by permuting the rows of the data matrix Xb , and performed canonical correlation analysis for the randomized data . This process was repeated 500 times to generate a set of canonical correlation values representing the distribution under the null hypothesis . The significance level of the test is given by the fraction of the distribution that is above the canonical correlation value obtained on the original data . We note that the randomization setup used here also automatically corrects for a possible multiple testing bias .
The leading pair of projection directions extracted from the proteomics ( view a ) and clinical data ( view b ) by sparse canonical correlation analysis showed a statistically significant correlation coefficient of 0 . 79 ( 0 . 01% significance level ) . Figure 3 depicts the correlation of the data in the directions corresponding to the leading pair of SCCA projections . The class clusters are relatively tight and do not overlap significantly . The leading pair of projection directions can be seen to pick up the properties in the data that separate active TB from symptomatic and asymptomatic controls . The proteomic profile , given by the weights in the leading projection direction wa ( 1 ) ( Figure 4 ) included 8 proteomic variables with non-negligible coefficients , selected out of 271 ( 3% ) , whilst in the clinical profile , given by the weights in wb ( 1 ) , 13 out of 18 variables had non-negligible coefficients . C-reactive protein ( CRP ) , has by far the largest positive weight in the clinical profile , persistent cough ( Cought7 ) , serum amyloid protein ( SAA ) are other discernible variables with positive weights , indicating that high values for them associate more frequently with active TB cases than the other two groups . BCG vaccination , PPD skin test , interferon-gamma level ( IFG ) , and previously case of TB ( PastTB ) have discernible negative weights , and thus intuitively correspond to variables less often found among active TB cases than the other two groups . Body mass index ( BMI ) and height appear with opposite signs , consistent with the definition of BMI . In the plasma proteomic profile the proteins at mass peaks m/z 11 , 475 , m/z 20 , 884 and m/z 43 , 541 have the largest positive weights ( presence makes the proteomic score higher ) . The protein at mass peak m/z 8 , 908 has the largest negative weight ( absence makes the proteomic score higher ) . Figure 5 depicts the projection of the malaria data onto the leading pair of SCCA directions . The canonical correlation showed a statistically significant correlation coefficient of 0 . 75 , ( 0 . 01% significance level ) . The data projection clearly clustered non-malaria community controls ( CC ) apart from malaria cases UM , SMA and CM . Figure 6 shows the weights of the proteomic and clinical variables in the leading SCCA projection directions . Eleven proteomic variables out of 774 ( 1 . 4% ) were present in the proteomic profile wa ( 1 ) , whilst all variables were present in the clinical profile wb ( 1 ) . The highest weights in the clinical profile , were temperature , and O-positive blood type ( BloodOPOS ) . In addition , 10 other clinical variables had positive weights , roughly indicating more abundant among the community controls than the other classes . The highest negative weights are with presence of malaria parasites ( MP ) and the presence of convulsions ( HowManyConvulsions ) . In addition , 42 other clinical variables had negative weights , thus intuitively less abundant among the community controls . In the proteomic profile , the proteins at mass peaks m/z Q10_1600_1 , 528 , Q10_2800_8 , 547 and H50_1800_8 , 765 had the highest positive weights and in addition three other variables had positive weights ( presence makes proteomic score higher ) . The proteins at mass peaks m/z Q10_2800_4 , 614 and H50_3000_20 , 188 had the highest negative weights ( absence makes proteomic score higher ) . Table 1 depicts the results of classification of the tuberculosis data in five classes defined by the latency and the presence of symptoms . We compared three different data: using proteomics data processed with principal component analysis ( PCA-Proteomics ) , using clinical data only ( Raw-Clinical ) and using the clinical projections extracted by canonical correlation analysis ( K+SCCA-Clinical ) . First , we observed that K+SCCA-Clinical had higher accuracy than Raw-Clinical in distinguishing non-TB subjects from symptomatic cases with no latent TB ( 85% vs . 77% ) and from symptomatic cases with latent TB ( 90% vs . 79% ) and in distinguishing non-TB subjects from non-symptomatic latent TB cases ( 85% vs . 78% accuracy ) . In the other classification tasks , the K+SCCA approach generally achieved similar level of accuracy as using the clinical data alone . In six of the experiments PCA-proteomics was the most accurate method , achieving 98% accuracy or higher . In contrast , in the other four experiments the accuracy of PCA-Proteomics ranges between 52%–77% and it is the least accurate method . Table 2 shows similar experiments for the malaria data . Here , we considered one-against-one classification within five classes ( SMA , CM , UM , DC , and CC ) . We first notice that Raw-Clinical obtained significantly higher accuracy than PCA-Proteomics alone for all class pairs . The K+SCCA-Clinical approach achieves higher accuracy than Raw-Clinical in two important cases: distinguishing Severe Malaria Anaemia from Uncomplicated Malaria ( 90% vs . 85% ) and Severe Malaria Anaemia from Disease Control ( 97% vs . 92% ) . For the class pairs CC vs . SMA , CC vs . UM and DC vs UM , Raw-Clinical and the K+SCCA-Clinical achieve 99%–100% accuracy . For the pairs CC vs . CM , CC vs . DC , and CM vs . UM , the K+SCCA-Clinical approach is slightly inferior , and for the pairs CM vs . DC and CM vs . SMA significantly inferior to Raw-Clinical . We also noticed that the K+SCCA-Clinical approach was able to improve on the proteomics data in all cases .
In this paper , we have put forward an approach for discovering biomarkers from plasma proteomic data by canonical correlation to clinical data collected from the same subjects . We have also shown an approach to predict diagnostic classes based on the selected biomarkers . We analysed a set of data consisting of plasma proteome and clinical profiles . Sparse canonical correlation analysis was shown to be effective in extracting small sets of proteomic variables , each representing a plasma protein , that correlate with clusters of similar clinical phenotypes in statistically significant manner ( p-value 0 . 01 ) . Sparsity of the extracted biomarker models is shown by the fact that 1 . 5% and 3% of the proteomics variables had non-negligible coefficients in the malaria and TB models , respectively . The sparsity of the ‘omic view of the SCCA model is deemed to be crucial for interpretation by human experts , as the set of proteomic variables to be studied remains tractable . Unlike SCCA , the KCCA method does not impose sparsity , so the KCCA is less amenable to human analysis . In diagnostic class prediction , we show that via canonical correlation analysis , it is possible to make use of proteomic data in order to improve on the diagnostic classification , even if no proteomics data is available at the time of prediction , only at the time of training the model . This is a close match to a real-world scenario of deploying a diagnostic tool to health care centres without expensive ‘omics measurement capabilities . In our experiments , the proposed approach appears to be advantageous ( a ) when proteomics data contains a strong signal predictive of the classification ( i . e . PCA-Proteomics accuracy higher than Raw-Clinical ) that can be mediated by the K+SCCA model , or ( b ) when proteomics data alone does not predict well ( i . e . PCA-Proteomics accuracy lower than Raw-Clinical ) but there is a synergistic latent signal between the proteomic and clinical profiles that K+SCCA can pick up . In the case of TB , a strong proteomics signal ( case a ) is found in six of the ten comparisons . In four of those six cases , K+SCCA matches the performance of Raw-Clinical , improves on Raw-Clinical on two of the cases and marginally loses in one of the cases . Evidence of a synergistic signal ( case b ) is found in four of the ten comparisons , in three of which K+SCCA matches Raw-Clinical and in one it exceeds the accuracy of Raw-Clinical: determining the presence of latent TB when there are no symptoms ( 85% accuracy versus 77% with clinical data alone ) . In malaria , first we note that the clinical data is very strong , there are no comparisons where PCA-proteomics exceeds the accuracy of Raw-Clinical ( i . e . case a ) . Thus , in this data set , the K+SCCA is required to pick out a synergistic latent signal ( case b ) between the proteomic and clinical variables , in order to improve on the predictions from clinical data alone . This appears to take place in two experiments: in the separation of severe malaria anemia from both uncomplicated malaria and from disease controls . In the comparisons involving cerebral malaria ( CM ) , we note that proteomics data seems to be weak in three of the four cases , and its seems that K+SCCA is hampered by this: Although it significantly improves over PCA-proteomics , it loses out to Raw-Clinical . Another observation of the experiments is that K+SCCA benefits the prediction of the difficult class pairs more than the easier ones: in all of the comparisons where Raw-Clinical accuracy is below 85% , K+SCCA improves on the Raw-Clinical model . Finally , we note that the canonical correlation analysis can also be used in the opposite way , namely to use the clinical data to extract more predictive biomarkers from proteomics data , and thus enhance the understanding of the systems biology underlying the complex phenotypes . Although this particular application was not the main focus in the present work , in our experiments the K+SCCA method often improved over the accuracy obtained with proteomics data alone . | Many infectious diseases such as tuberculosis and malaria are challenging both for scientists trying to understand the biochemical basis of the diseases and for medical doctors making diagnosis . The challenges arise both from the dependence of the diseases on sets of proteins and from the complexity of the symptoms . Biomarkers denote small sets of measurements that correlate with the phenotype of interest . They have potential use both in advancing the basic biomedical research of infectious diseases and in facilitating predictive diagnostic tools . We propose a new method for biomarker discovery that works by finding canonical correlations between two sets of data , the plasma proteomic profiles and clinical profiles of the subjects . We show that the method is able to find candidate proteomic biomarkers that correlate with combinations of clinical variables , called the clinical biomarkers . Using the clinical biomarkers improves the accuracy of diagnostic class prediction while not requiring the expensive plasma proteomic profiles to be measured for each subject . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"algorithms",
"systems",
"biology",
"medicine",
"infectious",
"diseases",
"computer",
"science",
"bacterial",
"diseases",
"mycobacterium",
"diagnostic",
"medicine",
"tuberculosis",
"global",
"health",
"biology",
"computational",
"biology",
"malaria",
"parasitic",
"diseases",
"hematology"
]
| 2013 | Biomarker Discovery by Sparse Canonical Correlation Analysis of Complex Clinical Phenotypes of Tuberculosis and Malaria |
Cortisol , secreted in the adrenal cortex in response to stress , is an informative biomarker that distinguishes anxiety disorders such as major depression and post-traumatic stress disorder ( PTSD ) from normal subjects . Yehuda et al . proposed a hypothesis that , in humans , the hypersensitive hypothalamus-pituitary-adrenal ( HPA ) axis is responsible for the occurrence of differing levels of cortisol in anxiety disorders . Specifically , PTSD subjects have lower cortisol levels during the late subjective night in comparison to normal subjects , and this was assumed to occur due to strong negative feedback loops in the HPA axis . In the present work , to address this hypothesis , we modeled the cortisol dynamics using nonlinear ordinary differential equations and estimated the kinetic parameters of the model to fit the experimental data of three categories , namely , normal , depressed , and PTSD human subjects . We concatenated the subjects ( n = 3 ) in each category and created a model subject ( n = 1 ) without considering the patient-to-patient variability in each case . The parameters of the model for the three categories were simultaneously obtained through global optimization . Bifurcation analysis carried out with the optimized parameters exhibited two supercritical Hopf points and , for the choice of parameters , the oscillations were found to be circadian in nature . The fitted kinetic parameters indicate that PTSD subjects have a strong negative feedback loop and , as a result , the predicted oscillating cortisol levels are extremely low at the nadir in contrast to normal subjects , albeit within the endocrinologic range . We also simulated the phenotypes for each of the categories and , as observed in the clinical data of PTSD patients , the simulated cortisol levels are consistently low at the nadir , and correspondingly the negative feedback was found to be extremely strong . These results from the model support the hypothesis that high stress intensity and strong negative feedback loop may cause hypersensitive neuro-endocrine axis that results in hypocortisolemia in PTSD .
PTSD is an anxiety disorder that results from exposure to traumatic events . According to the Diagnostic and Statistical Manual of Mental Disorders ( DSM-IV ) , the core symptoms are impaired concentration , emotional numbing , recurrent flashes of traumatic memories , social withdrawal , and hyperarousal [1] . Neuroendocrine studies identified the HPA axis as the site of action that brought about biochemical changes in response to severe stress and , in particular , cortisol variations were found to differ in PTSD compared to normal and depressed subjects . For the past 20 years , there were many contradictory reports about the findings of cortisol levels in PTSD that ranged from hypocortisolemia [2]–[6] to hypercortisolemia [7]–[9] to no change [10] , [11] , but all the reports indicated that the cortisol level was in the endocrinologic range with only subtle changes observed in neuro-psychiatric disorders . Before discussing the problem and findings of our modeling work , we provide a brief overview of various neuroendocrine findings of cortisol dynamics in PTSD patients based on the work of Yehuda [12] , and discuss the corresponding hypothesis that was generated about the HPA mechanism based on the clinical data . The stress responsive HPA axis belongs to the neuro-endocrine system that regulates cortisol through feedforward and feedback loops . Cortisol , also known as glucocorticoid , exhibits both ultradian and circadian patterns , which when disrupted result in various metabolic and psychiatric disorders including depression and PTSD . Stress induces the release of corticotrophin-releasing hormone ( CRH ) from the hypothalamus and activates adreno corticotrophic hormone ( ACTH ) in the anterior pituitary . ACTH moves to the adrenal cortex and stimulates the production of cortisol . Cortisol has a stronger affinity for mineralocorticoid receptors than for the glucocorticoid receptors ( G ) and forms a complex with these receptors . Both glucocorticoid and mineralocorticoid complexes undergo homo-dimerization to increase the activity of the complex [13] , and GR-cortisol complex in turn binds to CRH and ACTH to down regulate the production of cortisol ( see Figure 1 ) . This negative feedback loop from cortisol is vital in maintaining the homeostasis of the system during stress . When disrupted , it results in the loss of sensitivity of the HPA axis to the negative feedback regulation , based on which the popular “glucocorticoid cascade hypothesis” was formulated [14] . In contrast to the hypercortisolism observed in depressed patients , hypocortisolism was observed in PTSD subjects . Two hypotheses were proposed for hypocortisolism in PTSD subjects; ( i ) enhanced negative feedback of the HPA axis , and ( ii ) adrenal insufficiency . The enhanced negative feedback inhibition hypothesis , first proposed by Yehuda [15] , [16] , explained the hypocortisolism in PTSD subjects due to hypersensitive HPA axis with a strong negative feedback loop , while in the case of subjects with major depression , the weakened negative feedback loop resulted in hypercortisolism . To support this hypothesis , Yehuda et al . [17] evaluated the cortisol patterns collected from human subjects screened for PTSD and major depressive disorder by carrying out chronobiological analysis that fitted the cortisol data by standard cosinor and multicosinor models , and showed that only the difference in cortisol levels at nadir was statistically significant . Even though few earlier reports contradicted this hypothesis , and no statistically significant difference in the cortisol levels was observed between normal and PTSD subjects [7]–[9] , it is now widely accepted that in PTSD , the blunted cortisol secretion is due to its strong negative feedback loop . A survey of various neuro-endocrine findings on PTSD that predominately concerned cortisol dynamics was made in [18] . There are several reasons attributed to the discrepancy in cortisol levels in PTSD that include the heterogeneity of the epidemiological samples [19] , [20] , the methodologies used to determine cortisol levels , and the type of neuroendocrine challenges like DEX or Metyrapone tests that were used to probe the role of negative feedback loops contributed strongly to the distinction of various neuro-psychiatric disorders ( see [12] and the references therein for details of the test and its outcome ) . However , when all these protocols were tightly maintained , it was found that in PTSD , the cortisol level during the night was much lower than normal , and this is attributed to the strong negative feedback inhibition . All these observations from negative feedback loop tests resulted in ruling out the adrenal insufficiency hypothesis . The motivation for the present work , therefore , is to address the hypothesis proposed by Yehuda [15] , since the cosinor [17] models do not have the provision to show that the differing dynamics in normal , depressed and PTSD subjects were due to the varying strengths of the negative feedback loop in the HPA axis . To show the role of the negative feedback loop requires a mechanistic model with bio-chemical kinetics , motivating the development of an endocrine model of the HPA axis . Further , mechanism based models provide insight into the role of feedback loops and the functioning of networks on the whole [21] . Specifically , the options of drug targeting and its efficacy in the treatment of these disorders can be explored , since it is known that negative feedback loops play a vital role in dampening the effects of drugs [22] , [23] . Therefore , we constructed a nonlinear ordinary differential equations ( ODE ) model of the HPA axis to capture the 24 h cortisol dynamics of normal , depressed and PTSD subjects , and used the same data fitted by Yehuda et al . [17] to estimate the kinetic parameters of the ODE model through global optimization . To verify the hypothesis , two tunable kinetic parameters , namely , the strength of stress and the inhibition constant that determines the strength of the negative feedback loop were used in the model . Model construction , parameter estimation , bifurcation analysis , verification of the effect of negative feedback loops on pathology , simulation of phenotypes , summary and conclusion are elaborately explained in the following sections .
In developing the mathematical model of the cortisol molecular network depicted in Figure 1 , two assumptions were made: ( i ) the first order dilution rate due to the transport of hormones and autonomous degradation are considered together . Apart from dilution/autonomous degradation , Michaelis-Menten kinetics are separately considered for the degradation of the hormones and hormone complexes within each specific region of the brain ( hypothalamus , pituitary and adrenal ) ; and ( ii ) a sufficient number of molecules is present for the reactions to take place using continuum kinetics so that stochastic fluctuations ( internal noise ) are eliminated . The nonlinear ODE equations for the cortisol network are framed as follows: ( 1 ) ( 2 ) ( 3 ) ( 4 ) ( 5 ) CRH , ACTH , CORT , GR , and G are the corticotrophin-releasing hormone , adreno-corticotrophin hormone , cortisol , glucocorticoid receptor complex , and total glucocorticoid receptor , respectively . k is the bifurcation parameter that drives the system by initiating the CRH production , K is the inhibition constant that regulates the strength of the negative feedback loop , V are the rates at which the hormones CRH , ACTH , and CORT are degraded enzymatically through saturation kinetics , K are the Michaelis constants , K are the autonomous degradation constants , K are the rates of production of ACTH , CORT , and GR respectively , n1 , n2 are the Hill constants , and K1 is the activation constant . The negative feedback regulation by glucocorticoid receptor complex is described by Hill kinetics , the dilution/autonomous degradation rate by a first order reaction , and the enzymatic degradation by Michaelis-Menten kinetics . The glucocorticoid receptor dimerization reaction ( Eqn ( 4 ) ) that increases its own activity is expressed by Hill kinetics and this generates bistability in the model . The introduction of homo dimerization is akin to the model of Gupta et al . [24] that generates a kind of positive feedback in the model . Therefore , the present model has both a positive and a negative feedback loop , with the positive feedback loop giving rise to bistability , and the negative feedback loop giving rise to oscillations for the appropriate choice of parameters . Earlier models that described the HPA axis have either captured oscillations ( both ultradian or circadian ) or bistability ( [25] and references therein ) , but the present model is capable of exhibiting both of these dynamics . A detailed summary of the earlier work on mathematical modeling of cortisol dynamics was recently provided by Vinthers et al . [25] that classified cortisol mathematical models into two types; one for the ultradian dynamics that were endogenous , and the other for circadian oscillations driven by the external light input . The mathematical models for cortisol dynamics were based on either ordinary differential equations or delay differential equations , and these models primarily explained the occurrence of ultradian rhythms [26]–[37] . In describing the dynamics of these models , two further classifications were made regarding the origin of the ultradian rhythms that occur though ( i ) bursting , or ( ii ) fixed points; i . e , through Hopf bifurcation . However , Vinthers et al . [25] showed that their minimal ODE model of the HPA axis was incapable of exhibiting neither ultradian nor circadian dynamics for the imposed physiological constraints on the model . Similarly , when delay was suitably introduced in their ODE model , ultradian oscillations were observed , but the experimental evidence to account for such a long time delay was absent . Together , they suggested that the present mechanisms were inadequate to model the ultradian cortisol dynamics due to some missing links , and these were suggested to be the secretion of hormones in bursts by exocytosis , or the oscillatory elimination rates . Also , various ODE and DDE models discussed in [35] indicated that autonomous oscillations were absent . The present model consists of four dynamical variables that are inclusive of receptor dynamics . The model was formulated to exhibit only circadian oscillations that can fit the data of Yehuda et al . [17] for psychiatric disorders such as depression and PTSD . The main new aspect introduced in the present work is the elimination of the hormones in their respective brain regions using Michaelis-Menten kinetics that introduced implicit delay in the model , and therefore , can exhibit both ultradian and circadian oscillations . The model also has a lot of similarities to the circadian oscillatory model of Goldbeter [38] , and the model of Bliss et al [39] . For the choice of parameters , the model can also exhibit ultradian oscillations . We have not explicitly introduced a sinusoidal term that denotes the hypothalamic drive that induces the circadian oscillations of cortisol in HPA axis , and the driving term in the present model is the constant bifurcation parameter k . Similar to the model of Gupta et al . [24] , we neglected the circadian drive . Apart from enzymatic degradation , autonomous degradation/dilution was also considered , and modeled using first order kinetics . Modeling the receptor complex formation , and the dimerization reaction was similar to the work of Ferrell and Xiang [40] , [41] . The dimerization of receptor complex provides the positive feedback , and the dimer-cortisol complex , that binds the ACTH and CRH hormones , provides the negative feedback loop that down regulates the cortisol . The positive feedback leads to bistability , whereas the negative feedback is expected to lead to circadian oscillations for the choice of parameters . We used the software program XPPAUT for numerical integration and to generate all the bifurcation diagrams [42] . We provide the ordinary differential equation ( ODE ) file of XPPAUT as a separate supplementary file . We transported the data from XPPAUT to MATLAB to plot the figures . To estimate the kinetic parameters , and to perform the sensitivity and correlation analysis , we used the toolbox SensSB [43] . Clinical data for 24 h cortisol regulation in normal , PTSD , and depressed subjects were extracted from Yehuda et al . [17] using the software Labnotes ( V 1 . 0 ) [44] . The cortisol data from the blood was sampled at approximately 30 minutes from 10 . 00 AM to the next day at 10 . 00 AM . All the male PTSD subjects were Vietnam veterans , and the depressed subjects were outpatients with no PTSD . A distinct spike at 6 . 00 PM in all the three categories was observed due to the consumption of a meal , but the spike was pronounced in both normal and depressed subjects . The spiked data point was included in our analysis . Triplicate sets of cortisol data were available for the normal as well as for each of the pathological categories , and in some categories , the missing data were filled by imputation as follows; if the data of one of the three subject at one particular time point was missing , then the average value of the available data from the other two available subjects at that missing time point was taken . In order to estimate the kinetic parameters of the model , the time series of the three subjects were concatenated by pooling them together , and assumed to be homologous ( see Text S1 for further analysis on concatenation to explain the homogeneity of individual time series ) . The time series were concatenated by taking the average of the last point of one subject and the first point of the next subject ( Figure 2B ) . We concatenated the subjects ( n = 3 ) in each category and created a model subject ( n = 1 ) without considering the patient-to-patient variability within each case . This is done to present a model for a single patient in each group , and the variability within each of the groups is not considered due to lack of sufficient data . The signal to noise ratio of the mean subtracted concatenated time series ( to remove the linear trends ) was determined from the power spectrum of the time series to assess the quality of the signals . This is based on the method of Gang et al . [45] , and also to verify the circadian nature of the concatenated time series ( Figure 2C ) . The signal-to-noise ratio ( SNR ) was determined using the expression h , where h is the ratio of the height of the amplitude spectrum ( h ) to the background noise ( h ) , which is is the sum of the background noise obtained between 0–1 Hz ( half of the sampling frequency , 2 Hz ) , is the frequency of the spectrum corresponding to the maximum amplitude and is the frequency difference at full width half maximum ( see Figure 2C ) . The ratio is known as the quality/regularity factor q of the signal . A large SNR value indicates a strong signal-to-noise ratio [45] , and this was used to determine the degree of coherence of the signal in the presence of noise . The frequency of normal , PTSD , and depressed subjects' concatenated time series was found to be 0 . 0417 Hz , which was close to the circadian period . The SNR for normal , PTSD , and depressed subjects were 5 . 3 , 11 . 7 , and 15 . 5 , respectively . These values indicated that PTSD signal has a good SNR , while the concatenated time series of depressed subjects has a lower SNR than PTSD . This is in accordance to the trend observed by Yehuda et al . [17] that in the case of depressed patients , the weakened negative feedback loop in the HPA axis resulted in a noisy time series with poor SNR . Also , it is well known that a strong negative feedback loop provides good homeostasis , and low fluctuations in the system [46] , and the SNR values indirectly support the view that a strong fluctuation in depressed patients may be due to a weakened negative feedback loop . In the next section , it is shown through model simulation that depressed subjects indeed have a weakened negative feedback loop .
The nonlinearities of the model led to a multimodal parameter estimation problem , and therefore required global optimization methods to avoid convergence to local solutions [47] . In this work we used the algorithm SSm , available in the SensSB toolbox , that has been shown to be a powerful metaheuristic for kinetic parameter estimation of biological ODE models [48] . Although more parameters could be different between the three groups , according to the hypothesis , only two kinetic parameters , namely and , are considered to be significantly different in the three pathological cases . Therefore , the model calibration was performed simultaneously for the three time series , allowing and to differ for all the three cases , and forcing the remaining 18 parameters to be the same . The cost function was defined as the weighted least squares function ( J ) resulting from the sum of the squared distances between the experimental and predicted values of cortisol at each of the sampling points for normal , depressed , and PTSD , and it is given as follows: ( 6 ) where is the vector of parameters , is the number of measures for the experiment ( = 1 , 2 , 3 for Normal , Depressed and PTSD , respectively ) , is the experimental value of the cortisol for the experiment at the sampling point , and is the corresponding value predicted by the model . Since both the peak and the nadir of the cortisol concentration are critical in our study , the weights we chose were for the 10% highest and 10% lowest experimental data , and for the rest of them . Moreover , a death penalty constraint enforces the system to oscillate with a period close to 24 hours in order to agree with the circadian oscillations of the experimental cortisol data , and the kinetic parameters that led to stable steady states were discarded . For each iteration of the optimization method , the model was simulated until sustained limit cycle oscillations were achieved after discarding the transients , and the maximum value of the oscillation was taken as the first point to fit the experimental data for all the three categories of the concatenated time series . A good agreement between the experimental and the model predicted time series for all the three subjects was obtained ( Figure 3 ) although the fit for the depressed ( = 69 . 2 ) and the PTSD patients ( = 36 . 9 ) was much better than for the normal patients ( = 82 . 9 ) . Importantly , the model captured the significantly reduced levels of basal cortisol for PTSD subjects . Further , the free glucocorticoid-receptor concentration ( see Figure 4 ) was found to be higher for the PTSD than for the normal subjects , which agrees with Yehuda's findings [16] , [49] . Even though glucocorticoid-receptor concentration was not used in the model fitting , its elevated simulated level in PTSD reinforces the predictive power of the present model . The estimated strength of the stress determined by the parameter was higher for PTSD than for both depressed and normal subjects ( see Table 1 ) , and this was expected for an anxiety disorder precipitated by extreme stress . On the other hand , the estimated inhibition constant value was much smaller for PTSD than for both normal and depressed subjects , indicating the presence of stronger negative feedback loop . These values support Yehuda's hypothesis suggesting that an enhanced negative feedback of the HPA axis is responsible for the hypocortisolism observed in PTSD subjects , and this result explains the apparent contradiction of low cortisol levels in a disorder precipitated by extreme stress . In order to assess the statistical significance of this finding , the parameters k and K were reestimated 50 times keeping all the other parameters constant . Due to the high correlation between these two parameters and the noise in the measurements , different values were obtained in each optimization , even though a global optimization method was used . The different k and K values for 50 optimizations are shown in Figure 5A–B , and the Kolmogorov-Smirnoff ( KS ) test indicates a significant difference between normal ( N ) , PTSD ( P ) , and depressed ( D ) sets ( p-value0 . 05 ) . It is interesting to note that the feedback strength for PTSD was consistently much stronger ( lower K values ) than for depression ( larger K values ) as expected from the hypothesis . The peak and nadir cortisol concentration for the 50 sets of parameters were also determined as shown in Figure 6A–B . KS-test indicates a significant group difference between normal ( N ) , depression ( D ) , and PTSD ( P ) ( p-values0 . 05 ) . In tandem with the clinical data [17] , the nadir values were significantly different between N-P , N-D , and D-P . Also , it was observed that the simulated nadir for PTSD had very little variations within the different sets of parameters found , while simulations corresponding to depressed subjects showed very large variations in comparison to normal subjects , and this was due to the difference in the strength of the negative feedback loops . The only discrepancy from the clinical data was that , the simulated peak values for normal ( N ) and PTSD ( P ) were found to be significantly different ( p-values0 . 05 ) . Despite this discrepancy , the simulation of the reestimated parameter sets supports the hypothesis that PTSD arises due to the strong stress and strong negative feedback that results in hypocortisolemia from early night through late morning , whereas in depressed patients , the negative feedback is completely weakened and results in hypercortisolemia in comparison to normal subjects . To gain further insight into the cortisol dynamics predicted by the model , a local sensitivity analysis was performed for the optimal set of parameters given in the Table 1 . Parametric sensitivity analysis aims to investigate the effect of parameter changes on the model output , in this case , the concentration of cortisol . In this study we used relative sensitivity indices , computed by multiplying the partial derivative ( the absolute sensitivity function ) by the nominal value of the input and dividing by the output value . The relative sensitivity index ( SI ) of the model output to variations in the parameter evaluated for the optimal set of parameters is given by: ( 7 ) Robustness and sensitivity are “two sides of the same coin” [50]–[52] . The cortisol concentration for depressed subjects was found to be the most sensitive to changes in the inhibition constant among the three groups ( see Figure 7 ) . In contrast , due to the stronger negative feedback loop , the cortisol sensitivity to for PTSD subjects was much smaller than for normal and depressed subjects reinforcing the notion that strong negative feedback loops confer robustness to biological systems . The correlation between parameters was studied by computing the correlation between the dynamic sensitivities as described in [53] , [54] . The correlation matrix ( see Figure 8 ) showed a strong positive correlation between the sensitivities of the parameters and . This can be interpreted as the “compensatory effect of feedback to stress” with a low value of needed to maintain cortisol levels at an endocrinologic range ( although lower than the normal as reported for the majority of PTSD subjects [12] ) when the stress is high . The autonomous degradation constant for the cortisol ( ) was found to be two orders of magnitude higher than the rate for enzymatic degradation indicating that the autonomous degradation of cortisol dominates the degradation process . Accordingly , the sensitivity analysis confirmed that the model is almost irresponsive to changes on parameters and , but highly sensitive to changes in . In the case of CRH , the opposite effect was found; the autonomous degradation constant is smaller than the enzymatic degradation rate leading to a very low sensitivity for and a high sensitivity for both and . This suggests that the enzymatic degradation plays a more prominent role than the autonomous degradation for CRH . Similar behavior was found for the degradation of ACTH . The low sensitivity of the degradation rates of CRH and ACTH ( and ) can also explain that the estimated values are far from the values reported in the literature since changes in these values would not make a big difference in the cortisol dynamics . However , a very high correlation was observed between the pairs of parameters - and - , suggesting that other phenotypes with high autonomous degradation and low enzymatic degradation leading to similar cortisol dynamics could possibly be found . Bifurcation analyses were carried out by varying k , the intensity of stress , as the bifurcation parameter and cortisol as the dynamical variable . The estimated parameters for normal , PTSD , and depression were used to perform the analyses , except for k that was tuned to determine the occurrence of Hopf points . Bifurcation diagrams for normal , PTSD and depressed subjects were separately constructed as shown in Figure 9 , and all the three categories exhibit both bistability and Hopf bifurcation . As stress was increased from a low value , a z-shaped saddle-node bifurcation was observed . This can be interpreted as the time immediately aftermath of the trauma during which the cortisol level decreases as stress increases with a transition from hyper to hypocortisolemia . Hypocortisolemia is captured by the z-shaped bistablity in the model . This we attribute due to “peri/acute-trauma” . Evidences from the motor vehicle accident survivors suggest that their cortisol level immediately following the accident was significantly lower , and these survivors met the criteria for PTSD [55] , [56] . This was followed by a supercritical Hopf bifurcation with an unstable steady state surrounded by a stable limit cycle [57] . There are two such Hopf bifurcations found in the system , and both are supercritical in nature . The first Hopf point , HB1 , was found at a lower k value , and the other Hopf point , HB2 , terminated at a high k value . The bifurcation diagram indicates that for the choice of parameters obtained from the fitted data , the Hopf points are different for each of the categories . The birth of the Hopf points are advanced in the depressed patients , and delayed in the case of PTSD in comparison to the normal patients as shown in Figure 9 ( E–G ) . Similarly the Hopf point terminates at a much lower k value in the depressed subjects , and at an extremely high value in PTSD subjects in comparison to the normal subjects ( H–J ) . All the parameters in the simulation of bifurcations were the same except for the negative feedback strength , K , that was different in all the three categories . This indicated clearly that the stronger feedback in PTSD delays the Hopf bifurcation , but provides a very robust and wide bifurcation regime . Two parameter bifurcation analysis was also carried out in the k- K parameter plane to map out the oscillatory regimes for all the three different subjects . The wide oscillatory regime found in the k-K plane indicates the robustness of the oscillations to the k-K parameter changes . From the estimated parameters , the depressed subjects ( shown as ) were found close to the boundary between oscillations and steady state , PTSD ( ) were found inside the oscillatory region , and the normal ( ) were found below the depressed category ( Figure 9D ) . The period of the oscillations for the parameter values that were obtained from the estimation is close to circadian , and a very wide variation of the period to changes in k and K indicates the robustness of the oscillatory period to parameter changes ( Figure 10 ) . Normal , depressed and PTSD patients are also clearly distinguished in the stress-feedback-period parameter space ( see Figure 11A ) and in the stress-feedback-cortisol parameter space ( see Figure 11B ) . PTSD is a co-morbid disorder that may occur along with depression . There are two possible ways for the transition from normality to a disease state can take place: ( i ) normal PTSD depression , ( ii ) normal PTSD and the simulation predicts these transitions . There are also intermittent or borderline cases where the normal subjects may possibly transit to PTSD , which , when left untreated , may further degenerate to depression . Large population of individuals with PTSD share also major depression , and it has been observed that PTSD and depression share 10 out of 17 symptoms on the Hamilton rate of depression [58] , [59] . All these possibilities were captured in the model as the variations in the stress-feedback and strength-period parameter space . As indicated by the hypothesis , in the parameter space , PTSD has a low nadir cortisol value ( hypocortisolemia ) due to the strong negative feedback loop ( lower K lower in the feedback loop ) , while depression has a elevated nadir value ( hypercortisolemia ) due to a weakened negative feedback loop .
PTSD is a neuro-psychiatric disorder that requires prevention and intervention , for which identifying biomarkers is of paramount importance . The question that cortisol can be considered as a suitable biomarker for the psychiatric disorders is debatable , since the cortisol levels of PTSD and depressed patients are present within the normal endocrinological range . The problem is further compounded by the fact that most of these disorders are co-morbid in nature and cannot be clearly distinguished whether the subjects are both PTSD and depressed or either one of them . However , according to the hypothesis , the strength of stress and the negative feedback loops in the HPA axis are different for PTSD and depression , and capturing this difference through the model constitutes a substantial part of the present work . Mathematical modeling of cortisol dynamics have been carried out in previous studies , yet new models have been formulated and refined based on the information obtained from recent molecular biological techniques that further provided insight about the regulation and functioning of the network . There are multiple models describing the cortisol ultradian dynamics , but hardly any model was built towards understanding of hypocortisolemia in psychiatric disorders except that of Gupta et al . [24] , who modeled the chronic fatigue syndrome that resulted in hypocortisolemia . Their dynamical model distinguished the normal and pathological states as two stable steady states ( bistable , saddle-node bifurcation ) separated by an unstable steady state , and the two stable steady states can be traversed through by varying the strength of stress and initial conditions . The bistable dynamics in their model were due to the homo-dimerization of the glucocorticoid receptor-cortisol complexes , but the effect of the negative feedback loop , which was conspicuous in the HPA axis , its role in hypocortisolemia , and the generation of circadian/ultradian oscillations were not fully explored . The cortisol was not modeled for its oscillatory dynamics , and the rise and fall of cortisol in their model was obtained by modulating the stress bifurcation parameter as square wave dynamics that behaved more like an excitable system . The present model which exhibits both bistability and oscillations , emphasizes the importance of negative regulatory feedback loops that strongly up or down regulate the cortisol production in a circadian fashion in various neuro-psychiatric disorders . The model was built specifically to explain the role of negative feedback loops that can distinguish depressed and PTSD from normal subjects . In contrast to the models discussed in [35] , the present model also indicates that strong nonlinearity in the form of Hill's function and the implicit delay introduced through Michaelis-Menten kinetics generates intrinsic oscillations in the HPA axis without an external periodic forcing . This is one aspect that is introduced in the present work but requires further investigation . However , one drawback of the model is that the concentration of both CRH and ACTH exceeds the endocrinologic range as in most of the earlier published models for ultradian oscillations [30] . This happens despite the fact that enzymatic degradation was considered along with autonomous degradation as indicated in [25] . Also , while fitting the data for cortisol , corresponding ACTH and CRH circadian oscillatory data were not available , and therefore , the range for ACTH and CRH concentrations could not be exactly calibrated . Importantly , the limitation of the present study is that we have concatenated the subjects ( n = 3 ) in each of the category to single model subject ( n = 1 ) by assuming the homogeneity of subjects , and neglecting the patient-to-patient variability . So this can only be considered as an initial theoretical study , and the requires more data to validate these findings . The mechanistic clue that negative feedback loop in cortisol may determine different psychiatric disorders was first proposed by Yehuda , and this modeling study validates that hypothesis in a quantitative framework . However , the model contradicts the view of Yehuda that cortisol levels are insignificantly different at the peak , and this may be possibly due to the simplification , or some missing links in the model as suggested by Vinthers et al . [25] . Finally , the model is able to predict the transition between different disorders and it proposes that the transitions can be controlled by reducing the stress and regulating the strength of the feedback loop . This result is in tandem with an earlier study in which it was showed that a mild augmented cortisol treatment regimen followed for the PTSD patients with hypocortisolemia have reduced retrieval of excessive traumatic memories [60] . However , the exact mechanistic details about the role of cortisol negative feedback loop in the treatment were not clear . Therefore , it may be interesting to analyze the strength of negative feedback loop in this regimen by performing DEX or Metyrapone tests [12] . In summary , our model supports the view that the hypocortisolemia in PTSD is due to a strong negative feedback loop as hypothesized , and that the weak negative feedback loop is responsible for depression . Hypocortisolemia in the model was observed at all the time , and predicted statistically significant differences in cortisol levels even during the day ( i . e . , peak levels ) among normal , PTSD and depressed subjects for the concatenated patients . Again , more data is needed to validate this conclusion that is arrived with a very sparse amount of data . Importantly , the model predicted that the transitions from normal to PTSD , and vice-versa , PTSD to depression , and vice-versa , were due to a disrupted negative feedback loop , which when suitably regulated , can treat the disease . | PTSD is an anxiety disorder that occurs among persons exposed to a traumatic event involving life threat and injury . This is a co-morbid psychiatric disorder that occurs along with depression . Cortisol is an informative endocrine biomarker that can distinguish PTSD from other co-morbid disorders . In comparison to normal subjects , hypocortisolemia was observed during the night in PTSD , while hypercortisolemia was observed in depressed subjects . From analyzing the clinical data , Yehuda et al . hypothesized that hypocortisolemia in PTSD was due to the strong negative feedback loop operating in the neuroendocrine axis under severe stress . We complemented this hypothesis by constructing a mathematical model for cortisol dynamics in HPA axis and estimated the kinetic parameters that fitted the cortisol time series obtained from the clinical data of normal , depressed and PTSD patients . The parameters obtained from the simulated phenotypes also strongly support the hypothesis that , due to disruptive negative feedback loops , cortisol levels are different in normal , PTSD and depressed subjects during the night . Importantly , the model predicted the transitions from normal to various diseased states , and these transitions were shown to occur due to changes in the strength of the negative feedback loop and the stress intensity in the neuro-endocrine axis . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"systems",
"biology",
"biology",
"genetics",
"and",
"genomics"
]
| 2012 | Modeling Cortisol Dynamics in the Neuro-endocrine Axis Distinguishes Normal, Depression, and Post-traumatic Stress Disorder (PTSD) in Humans |
Despite the overall decrease in visceral leishmaniasis ( VL ) incidence on the Indian subcontinent , there remain spatiotemporal clusters or ‘hotspots’ of new cases . The characteristics of these hotspots , underlying transmission dynamics , and their importance for shaping control strategies are not yet fully understood and are investigated in this study for a VL endemic area of ~100 , 000 inhabitants in Bihar , India between 2007–2015 . VL incidence ( cases/10 , 000/year ) dropped from 12 . 3 in 2007 to 0 . 9 in 2015 , which is just below the World Health Organizations’ threshold for elimination as a public health problem . Clustering of VL was assessed between subvillages ( hamlets ) , using multiple geospatial and ( spatio ) temporal autocorrelation and hotspot analyses . One to three hotspots were identified each year , often persisting for 1–5 successive years with a modal radius of ~500m . The relative risk of having VL was 5–86 times higher for inhabitants of hotspots , compared to those living outside hotspots . Hotspots harbour significantly more households from the two lowest asset quintiles ( as proxy for socio-economic status ) . Overall , children and young adelescents ( 5–14 years ) have the highest risk for VL , but within hotspots and at the start of outbreaks , older age groups ( 35+ years ) show a comparable high risk . This study demonstrates significant spatiotemporal heterogeneity in VL incidence at subdistrict level . The association between poverty and hotspots confirms that VL is a disease of ‘the poorest of the poor’ and age patterns suggest a potential role of waning immunity as underlying driver of hotspots . The recommended insecticide spraying radius of 500m around detected VL cases corresponds to the modal hotspot radius found in this study . Additional data on immunity and asymptomatic infection , and the development of spatiotemporally explicit transmission models that simulate hotspot dynamics and predict the impact of interventions at the smaller geographical scale will be crucial tools in sustaining elimination .
Visceral leishmaniasis ( VL ) –also known as kala-azar—is the deadliest vector-borne parasitic disease after malaria worldwide and is transmitted by female sand flies [1–3] . On the Indian subcontinent ( ISC ) , VL is caused by Leishmania donovani and humans are considered the only host . Here , most of the infected individuals remain asymptomatic carriers of the parasite [4] . When symptoms do develop , they include prolonged fever , enlarged spleen and liver , anaemia and anorexia , leading to death when left untreated [5] . Globally , an estimated 15 , 000 to 65 , 000 cases occur each year [3 , 6 , 7] . VL is one of the neglected tropical diseases that has been targeted for elimination and control by the World Health Organization ( WHO ) [8] . For VL , the elimination target is defined as an annual VL incidence of <1 case per 10 , 000 capita at ( sub ) district level by 2020 on the ISC [8–10] . In the rest of the world , where the disease is also zoonotic , the WHO target for VL control is 100% detection and treatment of VL cases [11] . It has been estimated that the global health and economic gains from reaching these WHO targets would be enormous [12] . Interventions against VL on the ISC are based on early detection and treatment of cases and vector control , mainly through indoor-residual spraying of insecticide ( IRS ) [5 , 11] . In 2005 , the governments of India , Nepal and Bangladesh committed to large-scale VL elimination programs [13] . Over the past decade , VL incidence rates have decreased at subcontinent , country and regional levels , which may be attributable to these efforts or to overall socio-economic improvements [14–18] . Bihar state in northern India is the area most affected by VL , and accounts for approximately 80% of all reported VL cases on the ISC . Despite the overall decrease in VL incidence in Bihar , which in many places has reached or is nearing the elimination target , there remain hotspots of infection [19–24] . The drivers underlying VL hotspots and the duration and magnitude of the flare-ups at small-scales are not yet fully understood . In the past seven years different mathematical models that capture the transmission dynamics of VL , have been developed by Stauch et al . [13 , 25] and research groups of the NTD Modelling Consortium ( http://www . ntdmodelling . org/ ) to better understand VL transmission dynamics and to predict the impact of control , in particular the feasibility of achieving the WHO elimination target for the ISC with current control strategies [11 , 13 , 24–27] . All models have been based on data from VL endemic areas in India , Nepal and Bangladesh [11 , 26] , and currently do not include any spatiotemporal dynamics , besides the seasonal sand fly patterns . These models assume that without interventions , an endemic equilibrium is reached with homogenous distribution of cases , and currently do not simulate the presence of hotspots . The aim of this study is to gain insight into the spatiotemporal patterns and hotspots of VL incidence at hamlet level ( subunit of a village ) and their underlying drivers—using longitudinal case and population data from a study area in Muzaffarpur , Bihar , India [22 , 28] . At the local level , VL dynamics are known to be influenced by socio-economic status and immunity among other factors [1 , 28 , 29] . This has been confirmed for the study area during previous analyses , where belonging to the wealthiest quintile has proven to be protective . Inhabitants from this quintile have an odds ratio ( OR ) of 0 . 5 ( 95% confidence interval ( CI ) 0 . 3; 1 . 0 ) of acquiring VL , compared to the poorest asset quintile [1 , 28] . We hypothesise that—even in this era of decreasing VL incidence—VL outbreaks continue to arise in areas where relatively many people from the lower asset quintiles live . In addition , children and young adolescents ( 5–14 years ) are known to have the highest risk of developing clinical VL compared to the ( 0–4 years ) reference category , due to either more exposure to the parasite or a lack of immunity ( OR 2 . 5 , 95% CI 1 . 5; 4 . 0 ) [28] . According to previous studies , long-lasting immunity might be acquired after exposure to the parasite [11 , 30] . It is hypothesized that all people living in VL endemic areas like rural Muzaffarpur , at least in pre-control settings , are exposed to L . donovani at some point in their lives . Most people remain asymptomatic , just a small fraction of people develop clinical symptoms [4] . It is assumed that hamlets and villages get saturated over time and only susceptible inhabitants cause new flare-ups or ‘hotspots’ due to new-borns , migration , or waning immunity of those with past ( asymptomatic ) infection [25 , 29] . Both factors , poverty and age , are analysed in this study as potential drivers of hotspots . A better understanding about the presence of spatiotemporal clustering of VL at hamlet level and the underlying characteristics of such hotspots can aid in further shaping the intervention strategies towards 2020 and can be used to inform mathematical transmission models , subsequently improving their predictions regarding the feasibility of achieving and sustaining the VL elimination targets at a smaller geographical scale .
We exploited the georeferenced data on VL cases as well as demographic data on all households obtained in the “Muzaffarpur—TMRC Health and Demographic Surveillance System” for the period between 2007 and 2015 [31] . This study site is located in a densely populated rural area of the Kanti block ( or subdistrict ) in Muzaffarpur district , Bihar state , India ( Fig 1 ) , a region marked as a hyper-endemic for VL [22 , 28] . The study area covers approximately 85 km2 lying between 26 . 13054°N– 26 . 23554°N latitude and 85 . 16642°E– 85 . 27294°E longitude and includes 2% of the total population of Muzaffarpur district ( Census 2011 ) . The study area comprises 50 villages that can be further subdivided into 276 hamlets , also known as tola . Almost 100 , 000 inhabitants from ~13 , 500 households agreed to be enrolled in the study ( 90% of all households in the area ) and were followed annually . Cases were identified through active door-to-door surveys and a VL case was defined as “the combination of a clinical history typical for VL ( fever of >2 weeks’ duration with lack of response to antimalarial drug treatment ) ; a positive result in the rK39 rapid diagnostic test; and a good response to VL treatment” [28] . The documented year of detection of the case was based on the date of diagnosis and onset of treatment . Indoor residual spraying of houses in the region started in 2005 and has been ongoing ever since , but coverage has been poor [28] . Muzaffarpur has a humid tropical climate with temperatures ranging from 18 . 5 to 31 . 9 °C and an average rainfall around 1 , 300 mm annually . Bihar belongs to the five poorest states in India , and the inhabitants of Muzaffarpur district belong to the low to middle income segments of the Bihar wealth distribution [1 , 28 , 32] . The asset index of every household and the age of all inhabitants were assessed at the start of the study and were available for all VL cases detected between 2007 and 2015 . The asset index was based on the household wealth and living standards of Bihar and was determined by the type of housing and assets owned by the household like tables , chairs , bicycles and motorcycles . The asset index was assumed to remain stable throughout the study [1 , 28] . The variable “asset index” was transformed into quintiles ( with 20% of individuals belonging to category ‘1’ the poorest quintile of the Kanti block study population to 20% in ‘5’ the wealthiest population quintile within the area , measured at household level ) . An overview of the asset quintile distributions with the exact calculations is provided by Boelaert et al ( 2009 ) [1] . We calculated and plotted the overall annual VL incidence rates per 10 , 000 capita for each year of the study period and annual incidence rates for each hamlet . The annual incidence rates of hamlets with non-zero case counts are shown in a histogram . The geospatial heterogeneity in annual VL incidence was visualized with ArcGIS Pro version 2 . 1 . 0 ( www . esri . com ) . Hamlets were grouped into four incidence categories using the same cut-off values for all nine years . All hamlets with an annual VL incidence of zero VL cases per 10 , 000 capita were grouped into the same category . Based on the histogram , the remaining three categories were as follows: greater than 0 to 20 , greater than 20 to 46 and greater than 46 VL cases per 10 , 000 capita per year . As with other infectious diseases , chances that high VL incidence rates are present within a hamlet in one year , are expected to be higher for hamlets with high VL incidence rates in the previous year , compared to hamlets that had no cases in the previous year [33 , 34] . Patterns and autocorrelation over time in VL incidence within hamlets were quantified with generalized linear mixed models . We compared the predictive performance of negative binomial and Poisson models with or witout random intercepts and slopes over time using k-fold cross valiation ( k = 10 ) . In addition to random intercepts and slopes we tested if adding a gaussian process for temporal autocorrelation ( periods of consequtive years with outbreaks ) improved the models’ predictiove performance . Models were fitted using the rstan and brms packages in R version 3 . 3 . 3 . Spatial autocorrelation or clustering in interpolated annual VL incidence was assessed using Moran’s I index [35] . The I-statistic produces values range from -1 to +1 . An ‘I’ value around 0 suggests features with similar values are randomly distributed in space . An ‘I’ value significantly different from 0 suggests a pattern . Values close to +1 suggest positive autocorrelation , whereas a value close to -1 suggests negative autocorrelation . Significant ( p<0 . 05 ) positive autocorrelation indicates that features ( hamlets in this study ) are surrounded by features with similar values , which means the feature is part of a cluster [23 , 36] . We used a Poisson model through Kulldorff spatial scan statistics ( www . satscan . org ) to detect spatial clusters of high VL incidence or ‘hotspots’ for each year and spatiotemporal hotspots for the overall study period ( 2007–2015 ) . The model provides a series of Poisson-based draws against which annual VL incidence rates are compared . Under the null hypothesis , stating that cases are randomly dispersed in space , the expected number of cases is proportional to the population size of the area [37] . The analysis requires case and population counts for a set of data locations , as well as the geographical coordinates for each of the locations . We constructed a model with the following conditions: searching for the most likely hotspots without geographical overlap of hotspots within the same time frame and maximum radius of a cluster equal to average maximum semivariogram range . The semivariogram range represents a distance beyond which there is little or no spatial autocorrelation among variables and was measured by fitting semivariograms of the incidence data at hamlet level per 3-year time frame ( 2007–2009 , 2010–2012 and 2013–2015 ) . For each detected hotspot , the number of hamlets within the hotspot , radius ( km ) , population at risk within the hotspot and the observed number of cases versus the expected number of cases detected within the hotspot are given . The risk of having VL that is associated with an identified hotspot is presented as the relative risk ( RR ) , i . e . the ratio of estimated VL risk inside relative to outside a hotspot . Only significant ( p<0 . 05 ) hotspots were included in our further analyses . Incidence ( cases/10 , 000/year ) of each significant spatiotemporal hotspot were plotted over time to quantify the duration of VL outbreaks , hotspots were defined as ‘outbreaks’ when the incidence within a hotspot rose above 15 cases per 10 , 000 per year . Socio-economic status and age were analysed as potential drivers of VL hotspots , using multiple approaches . Two different methodologies were used to investigate a potential association between VL hotspots and socio-economic status ( measured by asset index ) . First , we compared the fraction of households in the different asset index quintiles of the study population , inside and outside of annual hotspots , using a Chi-squared test . A p-value of <0 . 01 was taken as the level of significance because of the large sample size of almost 100 , 000 inhabitants from ~13 , 500 households . We calculated the percentage difference and the 95% CI around the perectage difference of households within the poorest two asset quintiles inside and outside of hotspots , and the fractions were considered significantly different if zero did not fall into the CIs . Second , we investigated the potential association between VL incidence and poverty at the hamlet level by means of Poisson regression , with poverty being defined as the percentage of households within the two poorest asset quintiles of a hamlet . VL immunity is known to be linked to age and we used different methodologies to study the association between VL and age in relation to detected hotspots . First , we calculated the age distribution of inhabitants and VL cases living inside and outside of hotspots . We compared the age distribution of VL cases ( fraction 5–14 years versus all other ages ) using a Chi-squared test , a p-value of <0 . 05 was taken as the level of significance , because of the relatively small sample size of 329 VL cases . The age distribution of VL cases living within hotspots that were identified for the overall study period was investigated , by looking at the time-frame of outbreaks and splitting the outbreak into a first phase ( up to the peak ) and a later phase . A Chi-squared test was used to test if the age distributions ( 5–14 years versus the rest ) were significantly different when comparing the early with the later stage of an outbreak . Further , a binomial logistic regression model was created with age ( 0–4 , 5–14 , … , 35–44 and 45+ years ) and asset index as categorical predictors and hamlet as a random effect . The groups known to have the highest odds for VL in rural Muzaffarpur [28] , age 5–14 years and asset quintile 1 , were used as the reference categories for calculating the adjusted odds ratios . Whether an individual lived inside a detected hotspot or not , was added as a dichotomous predictor ( inside vs . outside ) , allowing for potential interaction with age and asset index . The Akaike Information Criterion ( AIC ) was used to determine whether adding of predictors and interactions between predictors improved the model fit . The difference in age distribution of VL cases over time was assessed by splitting outbreaks into an upward phase ( up to the peak ) and downward phase .
A total of 329 VL cases were detected between 2007 and 2015 , the average incidence was 4 . 2 cases per 10 , 000 per year within the study area . The highest annual VL incidence of 12 . 3 cases per 10 , 000 per year was documented in 2007 ( when the study started ) , followed by a steady decrease up to 2010 . In 2012 a small peak was observed . Annual incidence rates further dropped in later years of the study period and reached the elimination threshold in 2015 , when the annual VL incidence was 0 . 9 cases per 10 , 000 per year . Overall annual VL incidence rates between 2007 and 2015 and the 2020 elimination target are illustrated in Fig 2 . Annual VL incidence rates vary highly across the hamlets ( mean incidence 4 . 7 cases per 10 , 000 per year , range 0 . 0–555 . 6 cases per 10 , 000 per year ) . Around 40% of all hamlets ( 114 out of 276 ) captured all reported VL cases between 2007 and 2015 . Of all 2 , 484 hamlet-years ( 276 hamlets times nine years ) , there were 205 hamlet-years with VL case counts , and 57% of these hamlet-years with case counts occurred during the first three years of the study ( 2007–2009 ) . Almost half ( 48% ) of these hamlet-years had at least one neighouring hamlet ( within a range of 500 meters ) with reported VL cases during the same year . Furthermore , 65% of hamlets with reported VL cas es arose in the periphery around hamlets with reported VL cases in the previous year; 59% in or around hamlets with reported VL cases two years from that year and 53% in or around hamlets with reported VL cases three years from that year . The histogram ( top right panel of Fig 2 ) shows the incidence ( cases/10 , 000/year ) at hamlet level for all hamlet-years with VL case counts in that year . The median age of cases was 20 years , the youngest detected case was 2 years old and the oldest was 70 years . Generalized linear mixed effects regression for VL incidence over time within a hamlet revealed that VL incidence rates are temporally clustered for 1 . 6 years ( 95%-BCI 1 . 046–2 . 555 ) . The full model output is provided in S1 Table . Mapping the geographical distribution of VL illustrates the spatial heterogeneity in endemicity among hamlets in rural Muzaffarpur ( Fig 3 ) . Moran’s I index ( S2 Table ) showed that VL incidence rates per 10 , 000 per year were significantly spatially clustered at hamlet level in 2007 to 2012 . P-values below <0 . 05 or <0 . 01 suggest that the likelihood of this spatial pattern being random , i . e . a result of chance , is less than 5 or 1 percent respectively . In 2013–2015 , Moran’s I values approach zero and were not statistically significant , suggesting that no spatial pattern could be identified among hamlets with low and high incidence rates per 10 , 000 capita during these years . The combined semivariogram of the incidence data at hamlet level ( S1 Fig ) demonstrated that the maximum distance beyond which there is little or no autocorrelation among hamlets over 2 km . One to three significant spatial hotspots of VL were detected in each year . The hotspots include two to 40 hamlets and have a radius between 85 meters and 1 . 9 km ( modal radius of 500 meters ) . The largest hotspot covers a surface of 11 km2 . A total of 82 out of 276 hamlets were located within a hotspot during at least one year . The relative risk of having VL associated to living inside a hotspot is 5 to 86 higher relative to outside a hotspot . A full overview of the hotspots detected with spatial scan statistics is displayed in S2 Table . Three significant spatiotemporal hotspots that were detected over the overall study period ( Fig 4 ) . The hotspots have a duration of two to three years ( Fig 4 , panel B ) . A detailed overview of the characteristics of the spatial and spatiotemporal hotspots is given in S3 Table . Fig 5 shows the asset index distribution among households located inside ( panel A ) and outside hotspots ( panel B ) . This distribution differed significantly between households inside and outside hotspots ( χ2 = 54 . 7 , degrees of freedom ( df ) = 2 , p<0 . 001 ) , with hotspots holding 5% ( 95% CI 4 . 94%; 5 . 06% ) fewer wealthy ( 12% vs . 17% ) and 6% ( 95% CI 3 . 31%; 8 . 69% ) more poor households ( 51% vs . 45% ) . VL incidence and poverty at the hamlet level are significantly associated ( p<0 . 001 ) : hamlets with a higher VL incidence harbour relatively more households from the two poorest asset quintiles ( Fig 6 and Box 1 ) . The age distributions among VL cases that were detected within and outside of the hotspots are illustrated in Fig 7 . In the VL hotspots , 9% ( 95% CI -0 . 08%; 0 . 26% ) less cases were between 4 and 15 years old ( 34% vs . 43% ) , whereas the fraction of the entire population in the same age group was very similar inside and outside hotpots ( 27% vs . 26% , S3 Fig ) . The fraction of all VL cases belonging to the 5–14 years age group is not significantly different inside relative to outside of hotspots ( χ2 = 2 . 4 , df = 1 , p-value = 0 . 123 ) . S4 Fig shows there is no significant association between the average age of VL cases and annual incidence at the hamlet level for both hamlets within hotspots and outside of hotspots . The results presented in Table 1 illustrate that outside of hotspots , the risk of having VL is lower for all age groups compared to the reference group ( 5–14 years ) , whereas inside of hotspots , the older age groups ( 35–44 and 45+ years ) have a higher risk for VL compared to the reference group , also after adjusting for asset index . Comparable to the findings by Hasker et al . [28] , people from the highest asset quintile had the lowest risk of having VL , compared to the poorest asset reference group ( Box 1 ) . However , there was no difference in the association between asset index distribution and VL when comparing the people living inside to those outside of hotspots; adding an interaction term between asset index and hotspot did not improve the model fit ( AIC = 3794 . 6 vs . 3789 . 9 ) . We further investigated the potential role of progressively developing herd immunity at this small area level by comparing the age distribution of VL cases between the upward and the downward trend of small-scale VL outbreaks . For this , we used hotspots 1 and 2 ( Fig 4 ) as these hotspots showed clear peaks in VL incidence ( ≥15 cases/10 , 000/year ) and may therefore be considered as local outbreaks . Among the VL cases detected in the downward phase of a local ‘outbreak’ ( hotspot 1 and 2 combined ) , there was a shift towards younger ages , compared to cases detected during the upward phase of an outbreak ( hotspot 2 ) ( S5 Fig ) .
This data collected over a 9-year period in the heartland of the VL epidemic in India provided a unique opportunity to analyse changing patterns in VL epidemiology . The study demonstrates significant spatiotemporal heterogeneity in VL incidence at the subdistrict level in Bihar , India , and is the first study to investigate the drivers underlying hotspots at this geographical scale . The overall decrease in VL incidence observed in the area is comparable to the steep case declines observed in many other VL endemic areas on the ISC during the past ten years [17 , 38] . Despite the overall decrease , hotspots were identified during each year of the study period . The geospatial and ( spatio ) temporal clustering patterns found in this study ( Box 2 ) are comparable to the findings from other recent studies conducted in India and Bangladesh [39–41] . A recent study by Mandal et al . ( 2017 ) discovered that 93% of newly endemic villages in the Indian Vaishali district , a district located at the border of Muzaffarpur , occurred on the peripheries of previous year endemic villages , suggesting clustering in both space an time . We found similar patterns in Muzaffarpur: 65% of the hamlets with VL cases arised in the peripheries ( within a range of 500 meters ) of hamlets with VL in the previous year [40] . The researchers did not use statistical alayses to further explore the observed VL patterns . Earlier , spatial clustering of VL incidence was identified in Vaishali by Bhunia et al . ( 2013 ) . However , they used aggregated subdistrict level data to identify clustering , which differs from our geocoordinate based subvillage level approach . Moreover , only Moran’s I index and Getis-Ord Gi were applied . Both analyses separate clusters of high-high and low-low incidence levels , but do not specifically detect hotspots , like spatial scan statistics allows for [39] . Dewan et al . ( 2017 ) identified VL hotspots using the same analysis methods ( Moran’s I followed by Poisson spatial scan statistics ) at the mauza ( cluster of several villages ) level in an endemic area of Bangladesh [41] . All four identified hotspots had a ~2 km radius and inhabitants of hotspots had a 4 to 11 times higher relative risk of having VL , compared to people living outside of hotspots . In our study , hotspots had a 85 meters to 2 km radius and inhabitants of hotspots had a 5 to 86 times higher relative risk of having VL . So , at this smaller geographical scale , hotspots are smaller and have a higher relative risk , compared to the findings by Dewan et al . ( 2017 ) . These findings might suggest that the smaller the geographical scale of analysis , the stronger clustering of VL is present . Our findings suggest that hotspots are , at least partially , driven by poverty , since hotspots arise more frequently in hamlets with the highest proportion of ‘poorest of the poor’ . We hypothesised that hotspots might develop in areas with a relatively large susceptible population , leading to relatively older VL cases within hotspots , most clearly during the early stage of outbreaks . Our finding that outside hotspots the risk of VL was significantly lower among people of age 45+ years of age compared to people of age 5–14 years , but was comparable between these two age categories inside hotspots , suggests a potential role of long-lasting immunity as an underlying driver of heterogeneity in VL incidence . However , additional data on immune status are needed to further explore this hypothesis . The steep decrease in overall VL incidence rates observed over the 9-year study period may be the result of intervention strategies in place—even though IRS was mentioned to be suboptimal within the study area—making it challenging to interpret the temporal trends in VL incidence [28] . Furthermore , possible natural cycles or development of herd immunity might have played a role in the observed downward trend [20 , 42–44] . In contrast to the research by Hasker et al . [26] , other studies pointed out that being 15 years or older came with an increased risk of developing VL [45 , 46] . Outbreaks at this small geographical scale might be explained by specific risk factors increasing susceptibility to L . donovani infection , like migration and immunosuppression ( mainly due to HIV-co-infection ) [19 , 20 , 30 , 46] . The risk related to proximity of VL cases , could also be another underlying driver of development of VL , as was explored by Chapman et al . [47] and Hasker et al . [48] . The size of hotspots make it is most likely that both human movement and the movement of sand flies play an important role in the spread of infection [49–51] . The role of population density , access to health services , vector characteristics and vector control strategies , which have shown to be potent drivers of other vector-borne infections , would pose interesting additional factors to explore in future analyses as potential drivers of VL hotspots [23 , 52–56] . In current mathematical transmission models for VL , a homogeneous seasonal equilibrium of incidence is assumed at subdistrict level when no interventions are in place . Heterogeneity is included at individual level regarding age-dependent sand fly exposure [26] . To include spatiotemporal clustering of VL cases ( hotspots ) as an additional source of heterogeneity in these models , together with a human migration pattern , would represent a more realistic framework when simulating infection at the smaller geographical scale . This becomes especially relevant close to and in the first years after reaching the elimination target [11 , 25 , 26] , in which case a stochastic individual-based transmission model would be a suitable tool for predictions . Since poverty partially explains the spatial heterogeneity—as families belonging to the same sociocultural group ( caste ) often live in close proximity—stratifying the population by socio-economic status in the current transmission models would probably better portray the spatial clustering of outbreaks in these populations . Current models assume that immunity lasts only for 1 or 2 years after infection ( asymptomatic and/or symptomatic ) . Based on our findings , a longer immunity ( i . e . ~15 to 30 years ) would seem more realistic , and is also congruent with clinical observations , at least for clinical VL cases . The identification of focal areas that are at greater risk for VL may help define priority areas of specific interventions when nearing the 2020 elimination target . Following our findings , active case detection might be focussed on areas where the ‘poorest of the poor’ are located when resources are limited . Based on our estimates , the current IRS policy of spraying all households within a radius of 500 meters from a detected case in less endemic areas , seems reasonable . However , the true impact of IRS on the number of VL cases remains uncertain [57] . Improving socio-economic conditions among the poorest of the poor households might be another effective control measure when targeting for VL elimination , but remains challenging . Symptomatic VL cases are only ‘the tip of the iceberg’ and there are still many unknowns regarding the roles of asymptomatic individuals and people with post-kala azar-dermal leishmaniasis ( PKDL ) ( Box 3 ) . Previous studies have shown that “asymptomatic infection also tends to show clustering around VL cases” [29] , and that having a new asymptomatic infection in a household puts the relatives at risk for developing VL or asymptomatic infection in the future [30] . Therefore , future research on L . donovani transmission in the post-elimination era should include asymptomatic individuals , by measuring Direct Agglutination Test ( DAT ) titres and rK39 antibody levels [58] . The ratio symptomatic: asymptomatic is estimated to be between 1:6 to 1:17 in India [4 , 59 , 60] , this suggests that around the 329 clinical VL cases ~2 , 000 to ~6000 individuals were infected without showing any signs or symptoms . Though the ratio of incident asymptomatic infections to incident clinical cases seems to increase with decreasing transmission intensity [61] , the overall prevalence of asymptomatic infections remains relatively low . Even when life-long immunity would be generated after an asymptomatic infection—what is not certain at this point—not all inhabitants of endemic hamlets will develop VL immunity during their lifetimes with the current incidence rates . The increasing pool of susceptible individuals that forms as VL incidence decreases may be a source of new large-scale epidemics of clinical cases—and this leads to the question whether elimination is desirable if one does not aim for the zero transmission goal [27] . Several study limitations could have affected our results . This study is , to our knowledge , the first study to assess temporal clustering of VL in endemic settings and to explore underlying drivers of spatial and spatiotemporal VL hotspots . Ideally , we would have used one model to simultaniously assess spatial and temporal autocorrelation . For example , by using a spatiotemporal model through Integrated Nested Laplace Approximation ( INLA ) [62] . However , these more advanced types of analyses are not easy to interpret and the applied methods in this study were sufficient in identifying spatial and temporal autocorrelation . Moreover , Muzaffarpur has always been one of the higher incidence districts , and it remains challenging to generalise our findings to low endemic settings . Heterogeneity is likely more evident in highly endemic settings , and these areas therefore best allow for investigating potential drivers of VL transmission dynamics . With deceasing overall VL incidence levels , larger study populations might be needed in the future . In conclusion , spatiotemporal heterogeneity in VL incidence is evident at subdistrict level in Bihar , India . This heterogeneous and spatiotemporally clustered distribution of VL at hamlet level can be a useful feature to include in the next generation of mathematical transmission models . Hotspots are to some extent driven by poverty , illustrating VL as a disease of ‘the poorest of the poor’ . The modal hotspot radius of 500 meters , and the average duration of hotspots of 1–5 years could be relevant for planning and targeting vector control and active case detection strategies . Data on asymptomatic infection , migration patterns , sand fly distribution are required to further understand the drivers and transmission dynamics underlying VL hotspots , as urgent questions need to be addressed with regard to the future investment in VL elimination . Should we try to sustain the current status of low transmission intensity or push transmission further down to zero ? | Visceral leishmaniasis ( VL ) is the deadliest vector-borne parasitic disease after malaria worldwide and is one of the neglected tropical diseases targeted for elimination and control by the World Health Organization . Despite the overall decrease in VL incidence in Bihar , India , a region previously highly endemic for VL , there remain hotspots of cases for reasons that are poorly understood . This study demonstrates clear spatiotemporal heterogeneity of VL in an area of ~100 , 000 population in Muzaffarpur , Bihar , India between 2007–2015 . Identified hotspots occurred almost always in the poorest communities , confirming that VL is a disease of ‘the poorest of the poor’ . Our finding that older age groups ( 35+ years ) are at a comparable high risk for VL within hotspots , similar to the 5–14 years highest risk group , but not outside hotspots , suggests a potential role of waning immunity as driver of hotspots . To further understand the drivers and transmission dynamics underlying VL hotspots , and ultimately inform policy , additional longitudinal data are needed to understand more about the role of asymptomatic infections , human movement , sand fly distribution and immunity . Including hotspots in mathematical transmission models is relevant for predicting the feasibility of reaching and sustaining VL elimination at a smaller geographical scale . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
]
| [
"kala-azar",
"medicine",
"and",
"health",
"sciences",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"india",
"age",
"distribution",
"sand",
"flies",
"parasitic",
"diseases",
"geoinformatics",
"age",
"groups",
"neglected",
"tropical",
"diseases",
"population",
"biology",
"insect",
"vectors",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"geography",
"zoonoses",
"spatial",
"autocorrelation",
"protozoan",
"infections",
"disease",
"vectors",
"people",
"and",
"places",
"population",
"metrics",
"asia",
"immunity",
"earth",
"sciences",
"leishmaniasis",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"species",
"interactions"
]
| 2018 | Visceral leishmaniasis: Spatiotemporal heterogeneity and drivers underlying the hotspots in Muzaffarpur, Bihar, India |
Two classes of antiviral drugs , neuraminidase inhibitors and adamantanes , are approved for prophylaxis and therapy against influenza virus infections . A major concern is that antiviral resistant viruses emerge and spread in the human population . The 2009 pandemic H1N1 virus is already resistant to adamantanes . Recently , a novel neuraminidase inhibitor resistance mutation I223R was identified in the neuraminidase of this subtype . To understand the resistance mechanism of this mutation , the enzymatic properties of the I223R mutant , together with the most frequently observed resistance mutation , H275Y , and the double mutant I223R/H275Y were compared . Relative to wild type , KM values for MUNANA increased only 2-fold for the single I223R mutant and up to 8-fold for the double mutant . Oseltamivir inhibition constants ( KI ) increased 48-fold in the single I223R mutant and 7500-fold in the double mutant . In both cases the change was largely accounted for by an increased dissociation rate constant for oseltamivir , but the inhibition constants for zanamivir were less increased . We have used X-ray crystallography to better understand the effect of mutation I223R on drug binding . We find that there is shrinkage of a hydrophobic pocket in the active site as a result of the I223R change . Furthermore , R223 interacts with S247 which changes the rotamer it adopts and , consequently , binding of the pentoxyl substituent of oseltamivir is not as favorable as in the wild type . However , the polar glycerol substituent present in zanamivir , which mimics the natural substrate , is accommodated in the I223R mutant structure in a similar way to wild type , thus explaining the kinetic data . Our structural data also show that , in contrast to a recently reported structure , the active site of 2009 pandemic neuraminidase can adopt an open conformation .
Strategies to combat the burden of disease caused by influenza mainly rely on vaccination [1] . However , in cases when vaccine efficacy is low or vaccine is unavailable , for instance during the first months of a pandemic , antiviral drugs are an important line of defense . For individual patients , especially when the immune system is compromised , antiviral therapy may be life saving [2] . Traditionally , both the adamantane and neuraminidase inhibitor class of drugs have been available for prophylaxis and therapy . However , the majority of recently circulating influenza viruses are resistant to the adamantanes , including the A/H1N1 2009 pandemic influenza virus [3] . This leaves neuraminidase inhibitors ( NAIs ) as the only option . Stimulated by the first neuraminidase ( NA ) crystal structures , the NAIs were designed to interact with the active site of all NA types [4] , [5] . Because the NA active site is highly conserved , resistant mutants with amino acid changes in the proximity of this active site , were considered likely to be enzymatically compromised and thus predicted to be of marginal epidemiological and clinical significance [6] . Indeed , the first viruses identified as harboring NAI resistant mutations were compromised in their replicative capacity and transmissibility [7]–[9] . Nevertheless , a novel oseltamivir-resistant influenza A/H1N1 variant emerged in the 2007–2008 influenza season and became the dominant virus in some parts of the world [10] , [11] . Resistance to oseltamivir was caused by a previously identified , and frequently observed , histidine to tyrosine change in the NA at position 275 ( H275Y ) . In contrast to earlier observations on H275Y mutants , this change did not compromise viral replication or transmissibility in the background of the 2007–2008 H1N1 virus [12]–[14] . Additional mutations in the NA were identified that explained why this resistant virus was able to emerge and become widespread [15] , [16] . With the 2009 pandemic , a novel influenza A/H1N1 virus was introduced into the human population [17] . Given that the appearance of NAI drug resistant mutations varies with the type and structure of the neuraminidase that the virus carries , we were interested to identify and characterize novel patterns of resistance in this virus . We identified a novel isoleucine to arginine change at position 223 ( I223R ) in the NA of a virus isolated from an immune suppressed child on prolonged oseltamivir and zanamivir therapy [18] . In contrast to the frequently observed H275Y change , which causes selective resistance to oseltamivir , the I223R mutation conferred a resistance phenotype against both oseltamivir and zanamivir . Soon after the identification of this first clinical case , the I223R change was found as a single change or in combination with H275Y , in a number of other cases [19] , [20] . The I223R/H275Y double mutant proved to have high levels of resistance to oseltamivir . Both in vitro studies and in animal models , viruses with I223R and the combination of I223R with H275Y were found not to be compromised in their replication capacity or transmissibility [21] , [22] . The structural basis of resistance conferred by the H275Y has been described previously [23] . Here we address the role of the I223R in NAI resistance alone and in combination with H275Y . Both I223R and H275Y changes are near the active site of A/H1N1 pandemic NA . Previously , it was shown that different NA sub types ( N1–N9 ) can be grouped into two genetically distinct groups [24] . The active sites of group-1 NAs ( N1 , N4 , N5 , N8 ) have an extra cavity when crystallized in the absence of inhibitor , because of the ‘open’ conformation of an active site loop , the 150-loop . This loop , containing residues 148–151 , closes a cavity adjacent to the active site upon drug binding ( 150-cavity ) . In contrast , in group-2 ( N2 , N3 , N6 , N7 , N9 ) NA , this loop adopts the closed conformation in the presence or absence of active site ligands . Recently however , a crystal structure for a ligand-free form of the H1N1 pandemic NA was reported , which showed the active site in a closed conformation [25] . More recently , NMR studies on the pandemic neuraminidase have considered this difference and suggested that the neuraminidase prefers to adopt an open conformation [26] . Our results support this suggestion by indicating that a ligand-free form of the pandemic neuraminidase adopts an open conformation . Together with this structure we present the structures of the I223R mutant neuraminidase in complex with oseltamivir and zanamivir to investigate the resistance mechanism of the I223R mutant neuraminidase . We interpret our results of enzyme kinetic measurements in relation to the structures of the complexes .
To study the effects of the single mutations I223R and H275Y , and the double mutant on the enzymatic properties of NA , Michaelis-Menten ( KM ) constants were determined using MUNANA as a substrate . Relative to wild type NA , KM values increased marginally for the I223R and H275Y single mutants . There was a maximum 8-fold increase for the double mutant ( Table 1 ) . In the 2007–2008 influenza H1N1 season , naturally occurring H275Y variants had emerged , which were not attenuated in virus growth or transmissibility . Our observations suggest that , as was observed with the non-attenuated H275Y variants in the 2007–2008 season , the occurrence of viable NAI resistant pandemic influenza A/H1N1 viruses is not an unlikely event . The sensitivity to oseltamivir and zanamivir was determined for the single and double mutants . By comparison to the oseltamivir inhibitor constant ( KI ) of wild type NA , the KI of the I223R mutant increased 48 times and the I223R/H275Y double mutant KI more than 7500 times . The change in the KI for I223R ( 48-fold ) was largely accounted for by an increased dissociation rate constant for the enzyme-inhibitor complex ( 15-fold ) rather than by a reduced association rate constant ( kon ) ( 3-fold ) . In the case of zanamivir , the KI increased 9-fold for the single and 22-fold for the double mutant relative to wild type . The change in the KI for I223R ( 9-fold ) was in this case accounted for by a reduced association rate constant ( 4-fold ) and a slightly increased dissociation rate constant ( 2-fold ) . Thus the I223R mutation has a markedly greater effect on NA inhibition by oseltamivir than by zanamivir . Although the KM-values for MUNANA are increased about equally for both the I223R and H275Y single mutants , the change in oseltamivir inhibitor constant for the H275Y mutant relative to wild type , is approximately 10 times greater than for the I223R mutant ( Table 1 ) . This may explain why the H275Y change may be the more likely resistance change in oseltamivir monotherapy and of course , to date , the H275Y mutation is the most frequently detected oseltamivir resistance mutation . Crystals of wild type and mutant neuraminidases were grown in the absence or presence of the oseltamivir and zanamivir neuraminidase inhibitors . The crystals yielded high-resolution diffraction data and the structures were solved by molecular replacement and refined by standard procedures ( Table 2 . ) The first striking feature from these data is that the 150-loop in the ligand-free I223R NA adopts an open conformation , as seen in the first reported N1 structure [24] , but in contrast to the closed conformation of this loop seen in the recently reported crystal structure of the 2009 pandemic A/H1N1 NA ( Figure 1A ) [25] . Recent NMR experiments likely clarify the apparent discrepancy in these different studies by showing that the 150-loop is flexible and that an equilibrium between the open and closed conformations exists in solution [26] . Inspection of our current structure shows that a phosphate ion , from the crystallization buffer , interacts with lysine 150 to stabilize the open conformation . We speculate that the different crystallization conditions used in the previous pandemic N1 study led to the stabilization of the closed form [25] . The exact function of the open 150-loop in group-1 neuraminidases is unknown , but opening and closing of the loop may have evolved for natural sialo-glycan substrates to fit into the active site . The open conformation of the 150-loop ( residues 148–151 ) generates an additional cavity at the edge of the active site that we have previously called the 150-cavity ( Figure 1A ) , which we have suggested to be an additional target site for neuraminidase inhibitor design [24] , [27] . In the wild type N1 neuraminidase structure , a small pocket , the “223-pocket” ( Figure 1C ) , also adjacent to the active site , but distinct from the 150-cavity , contains two water molecules . In the I223R neuraminidase , this pocket is occupied by the side chain of R223 . In displacing the two water molecules seen in the wild-type structure , the arginine makes a hydrogen bond with the side chain of serine 247 ( S247 ) . As a result , the side chain of S247 adopts a different rotamer and is oriented towards the active site which enables it to make a second hydrogen bond with glutamic acid 277 ( E277 ) . These changes in the structure of the I223R neuraminidase result in shrinkage at the edge of the active site pocket that accommodates the hydrophobic pentoxyl group of oseltamivir . Thus , oseltamivir binding to the active site of I223R neuraminidase requires the reorientation of the S247 and R223 side chains . Moreover , the reorientation of the side-chain of E277 , which is required for oseltamivir binding to wild-type neuraminidase , is presumably made even less energetically favorable in the I223R mutant by the need to disrupt the hydrogen bond between S247 and E277 . Thus the binding of oseltamivir to the active site of the I223R mutant results not only in changes in the conformation of the side chain of S247 but also in the side-chain of R223 being translated about 1 . 1 Å out of the active site compared to the mutant ligand-free structure ( Figure 2A ) . S247 therefore plays an important role in the decreased sensitivity of the I223R neuraminidase to inhibitors . Interestingly , a serine to asparagine change at position 247 ( S247N ) has also been linked to oseltamivir resistance [28] . Like the I223R change , the S247N mutation also causes enhanced resistance to oseltamivir when accompanied by the H275Y change . Both the single and the S247N/H275Y double mutant viruses were found not to be compromised in a ferret model suggesting that the mutant neuraminidases retain most of their normal activity [29] . An asparagine at position 247 , an amino acid with a larger polar side chain , can also affect the size of the hydrophobic pocket . Therefore , although there is no structural information in that case we speculate that , like the I223R change , the resistance mechanism of the S247N change also may involve shrinkage of the hydrophobic pocket . The mechanism whereby the H275Y mutant neuraminidase binds substantially more weakly to oseltamivir than wild type has been described previously [23] . In brief , the introduction of the bulkier tyrosine residue perturbs the position and the reorientation of the acidic side chain of E277 required for oseltamivir binding , such that the hydrophobic pentoxyl substituent of oseltamivir is translated out of the active site towards the pocket occupied in the I223R structure by the side chain of the arginine residue ( Figure 2B ) . The I223R mutation , therefore , appears to restrict the alternative position that the pentoxyl substituent adopts in the H275Y mutant . Furthermore , the binding of oseltamivir by the I223R/H275Y double mutant becomes even less energetically favorable than by either of the two single mutant proteins . Our structural data therefore provide an explanation for the 1750-fold poorer binding of oseltamivir by the I223R/H275Y double mutant by comparison with the wild-type neuraminidase ( Table 1 ) . Our structural studies help to explain how the single I223R and I223R/H275Y double mutant neuraminidases bind zanamivir almost as well as wild type . Instead of the hydrophobic pentoxyl substituent present in oseltamivir , zanamivir possesses the same polar glycerol substituent as sialic acid ( Figures 2 and 3 ) . Consequently , binding of zanamivir to the active site of neuraminidase does not require the reorientation of the side chain of E277 . In the case of the wild type neuraminidase , the glycerol moiety of zanamivir forms two hydrogen bonds with E277 . This interaction is unaffected in the H275Y mutant ( Figure 3 ) . In the I223R single or double mutant neuraminidases , somewhat different interactions are made by the glycerol moiety: one hydrogen bond is formed with E277 and another hydrogen bond is formed with S247 ( Figure 3A ) . To facilitate zanamivir binding in this way R223 is reoriented only marginally and it still retains the R223 hydrogen bond with S247 ( Figure 3 ) . The plasticity of the pandemic neuraminidase active site , especially around the hydrophobic pocket , allows the emergence of drug resistant influenza variants . In addition to the I223R change , isoleucine 223 changes to valine ( I223V ) and lysine ( I223K ) accompanying the H275Y change have also been reported [20] . This further indicates that this residue plays an important role in neuraminidase inhibitor resistance and stresses the importance of a detailed understanding of the mechanism of resistance . Since the single I223R change described here resulted in only a 9-fold increase in the KI for zanamivir , it is interesting to consider why the I223R mutant virus persisted in the IV zanamivir treated immunocompromised patient . As previously suggested [30] , suboptimal drug levels in the patient's respiratory tract , due to the route of administration ( IV versus inhaled ) as well as the immune status of the patient may have contributed to the emergence and persistence of the I223R mutant virus .
Hoffmann-La Roche Ltd . ( Switzerland ) kindly provided oseltamivir carboxylate . Zanamivir was kindly provided by GlaxoSmithKline ( the Netherlands ) . Recombinant influenza viruses were generated by reverse genetics essentially as previously described [31] . Each virus contained seven segments of A/WSN/33 and the neuraminidase gene of either A/Netherlands/2631_1202/2010 ( NL2631 ) , A/California/07/09 ( Cal07 ) or A/Vietnam/1203/04 ( VN04 ) . Recombinant viruses were propagated in embryonated chicken eggs . Allantoic fluids were harvested after 48 hrs and cleared from debris by centrifugation at 3000 rpm for 15 minutes . They were then used directly to measure enzyme kinetics or further processed to purify the neuraminidase for crystallography . Michaelis-Menten constants ( KM ) were determined at 37°C as previously described [23] , using standard initial rate measurements with virus diluted in 32 . 5 mM MES buffer ( pH 6 . 4 ) , 5 mM CaCl2 and fluorescent substrate 2′- ( 4-methylumbelliferyl1 ) -α-D-N-acetylneuraminic acid ( MUNANA ) concentrations ranging from 2 . 0 to 200 µM . Inhibition constants ( KI ) were determined by following the reduction in MUNANA hydrolysis rate following addition of inhibitor to a virus/MUNANA mixture approximately 100 s after initiation of the reaction . KI was then calculated using equation ( 1 ) [32]: ( 1 ) where VI is the new steady state hydrolysis rate in the presence of inhibitor at concentration [I] , V0 is the rate in the absence of inhibitor , [S] is the MUNANA concentration , KM is the Michaelis-Menten constant and KI is the dissociation constant for inhibitor binding . The kinetic parameters for inhibitor binding were determined by analyzing the exponential approach to VI using the following equation [33]: ( 2 ) where F0 and Ft are the fluorescence signals at time zero and time t , and kOBS is the first-order rate constant . Association ( kon ) and dissociation ( koff ) rate constants for inhibitor binding were then determined by plotting kOBS versus [I] using the following equation [34]: ( 3 ) In most cases the kinetically determined KI values ( koff/kon ) agree well with those determined using equation ( 2 ) . Measurements were performed in triplicate using three different inhibitor concentrations . Recombinant virus was harvested from clarified allantoic fluid by centrifugation at 6000 rpm overnight at 4°C and resuspended in 10 mM TRIS buffer , ( pH 8 . 0 ) , 150 mM NaCl ( TRIS-buffer ) . Virus was then layered over a continuous sucrose gradient ( 15–40% sucrose in TRIS-buffer ) and centrifuged at 25 . 000 rpm for 45 minutes at 4°C . The virus-containing fraction was collected and virus was pelleted by dilution of the remaining sucrose with Tris-buffer and centrifugation at 27 . 000 rpm for 90 minutes at 4°C . The purified virus was resuspended in Tris-buffer . Next , glycoproteins were released from the virus by bromelain ( Sigma-Aldrich ) digestion for 1 hour at 37°C and digested viruses were pelleted by centrifugation at 55 . 000 rpm for 10 minutes . A protease inhibitor tablet was added to the NA containing supernatant to prevent further protein degradation ( Roche ) . NA was purified by layering the supernatant over a sucrose gradient ( 5–25% sucrose in 10 mM TRIS buffer ( pH 8 . 0 ) and centrifugation at 38000 rpm for 18 hours at 4°C . Subsequently , NA-containing fractions were pooled and further purified using an Q-15 anion-exchange column ( Sartorius ) . Finally , buffer was changed by overnight dialysis against Tris-buffer supplemented with 5 mM CaCl2 and neuraminidase protein was concentrated 6 mg/ml using a 50 kD vivaspin column ( Sartorius ) for crystallization experiments . Neuraminidase crystals were obtained by vapor diffusion from sitting drops dispensed with an Oryx 8 robot ( Douglas Instruments ) . The drops consisted of 100 nl of protein in the absence or presence of 1 mM inhibitor ( oseltamivir or zanamivir ) mixed with 100 nl of reservoir solution . The reservoir solution of the I223R ligand-free crystal consisted of 20% PEG1000 , 0 . 6 M Ammonium Phosphate and 0 . 1 M Sodium Acetate ( pH 4 . 6 ) . The reservoir solution of the I223R crystal in complex with zanamivir consisted of 18% PEG3350 , 0 . 2 M Sodium Fluoride and 0 . 1 M bis-TRIS Propane buffer ( pH 6 . 5 ) . The reservoir solutions of the other crystals consisted of 15% PEG3350 , 0 . 1 M bis-TRIS propane and 0 . 1 M sodium acetate buffer ( pH 4 . 6 ) . Crystals were transferred into a cryoprotectant that consisted of reservoir solution supplemented with 20% ( v/v ) ethylene glycol before flash freezing in liquid nitrogen . Data sets were recorded on an ADSC Q315 CCD , Pilatus 6M-F and SLS/Dectris Pilatus miniCBF detectors at the Diamond light source ( Oxford , UK ) . Diffraction images were integrated using iMOSLFM [35] or DENZO and scaled with SCALA [36] or SCALEPACK for the ligand-free I223R structure . Neuraminidase structures were solved by molecular replacement with PHASER [37] using the wild type structure ( protein databank ( PDB ) identification ( ID ) code 2HU4 ) as the initial search model . Refinement was performed using Refmac5 [38] or PHENIX Refine [39] . Manual model building was done using Coot [40] , structure validation was assessed with MOLPROBIDITY [41] and figures were created using Pymol ( http://pymol . sourceforge . net/ ) . | Recently , a pandemic A/H1N1 influenza virus was isolated from an immune compromised patient with a novel antiviral resistance pattern to the neuraminidase inhibitor class of drugs . This virus had an amino acid change in the viral neuraminidase enzyme; an isoleucine at position 223 was substituted by an arginine ( I223R ) . Patients infected with such a virus leave physicians with reduced antiviral treatment options , since pandemic viruses are naturally resistant to the other class of antivirals , the adamantanes . Previously , we have shown that this mutant virus retains its potential to cause disease and may still be able to spread in the human population . Here we used enzyme kinetic measurements and crystal structures of the I223R mutant neuraminidase to determine the resistance mechanism of this amino acid change . We found that the I223R change results in shrinkage of the active site of the enzyme . As a result , binding of the neuraminidase inhibitors is affected . In addition , we found that the active site of our pandemic neuraminidase structure , crystallized in the absence of inhibitor , has an extra cavity ( 150-cavity ) adjacent to the active site . Our study could aid in the development of novel inhibitors designed to target the 150-cavity and active site of the enzyme . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
]
| [
"medicine",
"infectious",
"diseases",
"protein",
"chemistry",
"drugs",
"and",
"devices",
"influenza",
"pharmacology",
"biology",
"viral",
"diseases",
"drug",
"interactions",
"biophysics"
]
| 2012 | H1N1 2009 Pandemic Influenza Virus: Resistance of the I223R Neuraminidase Mutant Explained by Kinetic and Structural Analysis |
The balance between pro-inflammatory and regulatory immune responses in determining optimal T cell activation is vital for the successful resolution of microbial infections . This balance is maintained in part by the negative regulators of T cell activation , CTLA-4 and PD-1/PD-L , which dampen effector responses during chronic infections . However , their role in acute infections , such as malaria , remains less clear . In this study , we determined the contribution of CTLA-4 and PD-1/PD-L to the regulation of T cell responses during Plasmodium berghei ANKA ( PbA ) -induced experimental cerebral malaria ( ECM ) in susceptible ( C57BL/6 ) and resistant ( BALB/c ) mice . We found that the expression of CTLA-4 and PD-1 on T cells correlates with the extent of pro-inflammatory responses induced during PbA infection , being higher in C57BL/6 than in BALB/c mice . Thus , ECM develops despite high levels of expression of these inhibitory receptors . However , antibody-mediated blockade of either the CTLA-4 or PD-1/PD-L1 , but not the PD-1/PD-L2 , pathways during PbA-infection in ECM-resistant BALB/c mice resulted in higher levels of T cell activation , enhanced IFN-γ production , increased intravascular arrest of both parasitised erythrocytes and CD8+ T cells to the brain , and augmented incidence of ECM . Thus , in ECM-resistant BALB/c mice , CTLA-4 and PD-1/PD-L1 represent essential , independent and non-redundant pathways for maintaining T cell homeostasis during a virulent malaria infection . Moreover , neutralisation of IFN-γ or depletion of CD8+ T cells during PbA infection was shown to reverse the pathologic effects of regulatory pathway blockade , highlighting that the aetiology of ECM in the BALB/c mice is similar to that in C57BL/6 mice . In summary , our results underscore the differential and complex regulation that governs immune responses to malaria parasites .
The outcome of microbial infections is dependent on the balance between pro-inflammatory and regulatory immune responses . Priming of naïve T cells and their differentiation into effector cells needs to be balanced by switching off these cells , at an appropriate stage of infection , in order to prevent tissue damage ( immune pathology ) . Two important T cell inhibitory pathways involve signalling through members of the CD28:B7 superfamily of costimulatory molecules , namely cytotoxic T lymphocyte antigen-4 ( CTLA-4; CD152 ) and programmed death-1 ( PD-1; CD279 ) . While CTLA-4 is expressed on activated T cells including regulatory T cells [1]–[3] , PD-1 is broadly expressed on activated T cells , regulatory T cells and other haematopoietic cells [4] . T cell activation through the T cell receptor ( TCR ) and the costimulatory molecule CD28 results in increased expression of CTLA-4 [1] . Since both CD28 and CTLA-4 bind to B7-1 ( CD80 ) and B7-2 ( CD86 ) on antigen-presenting cells [5]–[7] , sequential expression of CD28 and then CTLA-4 allows T cells to be intrinsically self-regulating . CTLA-4 has higher affinity to the B7 molecules than CD28 . Similarly , PD-1 binds to PD ligand 1 ( PD-L1; CD274 ) and 2 ( PD-L1; CD273 ) which are upregulated on activated macrophages and dendritic cells ( DCs ) [4] . In addition , PD-L1 , which is also expressed on activated T cells , has recently been shown to bind B7-1 [8] , suggesting that there may be opportunities for cross-talk between the CTLA-4/B7 and PD-1/PD-L1 pathways . Consistent with their roles in the physiological regulation of cellular immune responses , CTLA-4−/− and PD-1−/− mice develop spontaneous autoimmune diseases; CTLA-4−/− mice die 2–3 weeks after birth from systemic lymphoproliferation [9] , [10] while PD-1−/− mice develop lupus-like glomerulonephritis and destructive arthritis [11] , [12] . Accumulating data suggest that PD-1/PD-L1 signalling , and in some cases CTLA-4 signalling , is implicated in the T cell exhaustion that is seen in many chronic infections [13]–[15] . For example , expression of PD-1 is associated with progressive loss of CD8+ T cell effector function during persistent lymphocytic choriomeningitis virus ( LCMV ) infection in mice; blockade of PD-1/PD-L1 interactions but not the CTLA-4 pathway augmented T cell function and allowed the virus to be controlled [16] . Similarly , in humans , PD-1 is upregulated on both CD4+ and CD8+ T cells during human immunodeficiency virus ( HIV ) infection and on CD8+ T cells during hepatitis C virus ( HCV ) infections and is associated with functional impairment of T cells and disease progression [17]–[19] . In vitro blockade of PD-1/PD-L1 pathway significantly increases CD8+ and CD4+ T cell function during HIV infection [20] . There is limited information available on the role of the PD-1/PD-L2 in chronic infections . In vitro blockade of PD-1/PD-L1 and to a lesser extent PD-1/PD-L2 resulted in reversal of immune dysfunction in HCV [21] . PD-L2 expression on dendritic cells is correlated to morbidity in experimental chronic schistosomiasis [22] . High levels of CTLA-4 expression are found on HIV-specific CD4+ T cells , but not on CD8+ T cells , and in vitro blockade of CTLA-4 enhances HIV-specific CD4+ T cell function [17] . Likewise , CTLA-4 blockade augments T cell responses to , and resolution of chronic infections such as Helicobacter pylori [23] , Leishmania donovani [24] , Leishmania chagasi [25] and Trypanosoma cruzi [26] in mice . With regard to acute infections , CTLA-4 blockade during Nippostrongylus brasiliensis [27] and Listeria monocytogenes [28] infection greatly enhanced T cell responses , resulting in more effective infection control . However , although CTLA-4 blockade enhanced T cell responses during Mycobacterium bovis infection , this did not have any effect on bacterial clearance [29] . PD-L1−/− mice are markedly more resistant to rabies virus [30] and Histoplasma capsulatum infection [31] than are wild-type mice . While these studies clearly indicate a role for the PD-1/PD-L1 pathway in dampening T cell responses , there is , rather confusingly , some evidence that this pathway is important in promoting CD8+ T cell responses in murine influenza virus [32] and Listeria monocytogenes [33] infections , suggesting that the outcome of PD-1/PD-L1 interactions might be modified by other regulatory pathways . Moreover , in Plasmodium yoelii malaria infections , CTLA-4 blockade increased T cell activation and IFN-γ production leading to early resolution of infections with the non-lethal 17X strain , but to increased severity of infections with the highly virulent 17XL strain of the parasite [34] , suggesting that enhancing T cell activation can be beneficial in relatively mild infections but can exacerbate virulent infections . Limited data are available for the PD-1/PD-L2 pathway during acute infections: PD-1/PD-L2 but not PD-1/PD-L1 blockade favours trypanosomatid growth in macrophages [35] and PD-L2 blockade enhances Th2 responses during Nippostrongylus infection [36] . Very few studies have directly contrasted the roles of CTLA-4 and PD-1 in the same infection , investigated the role of these pathways in determining susceptibility or resistance to infection in different mouse strains , or evaluated the extent to which they modulate immune pathology versus pathogen clearance . Here we have directly compared the roles of the CTLA-4 and PD-1 pathways in an acute malaria infection model in which resistance or susceptibility to immune-mediated pathology varies among strains of mice . P . berghei ANKA ( PbA ) infection of experimental cerebral malaria ( ECM ) -susceptible C57BL/6 mice reproduces the neurological signs associated with human cerebral malaria , the most severe complication of infection by the human parasite , P . falciparum [37] . Ante-mortem , the diagnostic neurological signs of ECM are ataxia and/or paralysis , which quickly leads to seizures , prostration and death within 10 days of infection . Histologically , CM is characterised by oedema , petechial haemorrhages and adherence of leucocytes and parasitised red blood cells to brain endothelium [37] . The essential triggers for ECM include systemic priming of CD4+ T cells by conventional DCs [37] , the production of pro-inflammatory cytokines such as IFN-γ [38] , [39] , the recruitment of effector CD8+ T cells to the brain [40] , [41] , and parasite accumulation in the cerebral microvasculature [42]–[44] . C57BL/6 mice infected with PbA develop a multi-organ disease as recently described [45] and – as in the brain - this is mediated by T cells and IFN-γ . However , this systemic disease , in the absence of cerebral involvement , does not appear to be fatal and mice will die at a later time point from hyperparasitaemia . The current best model of the pathogenesis of ECM is that CD8+ T cells damage cerebral vascular endothelial cells and the underlying basement membrane , thereby breaching the blood brain barrier , causing haemorrhage and oedema . Thus , pathology manifests initially in the brain because this organ is particularly vulnerable to the immediate consequences of endothelial damage . In contrast , the majority of PbA-infected BALB/c mice do not develop ECM but die from high parasitaemia and anaemia 2–3 weeks post-infection [46] . Hence , BALB/c mice are considered resistant to PbA-induced immune pathology . We hypothesised , therefore , that T cell-mediated inflammatory responses may be down-regulated in BALB/c mice ( preventing control of parasitaemia but also preventing accumulation of T cells in the brain ) , and that this might be due to differential regulation of CD4+ and/or CD8+ T cells by CTLA-4 and/or PD-1 . We found that in C57BL/6 mice , ECM develops despite high levels of expression of inhibitory receptors on CD4+ and CD8+ T cells . Conversely , we found that blockade of either CTLA-4 or PD-1/PD-L1 , but not PD-1/PD-L2 , during PbA-infection leads to the onset of ECM in normally resistant BALB/c mice and that this is accompanied by the characteristic features of T cell hyperactivity , raised IFN-γ levels and accumulation of CD8+ T cells and parasites in the brain . Thus , the CTLA-4 and PD-1/PD-L1 pathways seem to function very efficiently in BALB/c mice , maintaining the balance between immunity and immune pathology during the critical early stage of infection .
Animal experiments performed in the United Kingdom were approved by the LSHTM Animal Procedures and Ethics Committee and were performed under licence from the United Kingdom Home Office under the Animals ( Scientific Procedures ) Act 1986 . In Singapore , all experiments and procedures were approved by the Institutional Animal Care and Use Committee ( IACUC ) of A*STAR ( Biopolis , Singapore ) ( Authorization No IACUC 080321 ) in accordance with the guidelines of the Agri-Food and Veterinary Authority ( AVA ) and the National Advisory Committee for Laboratory Animal Research ( NACLAR ) of Singapore . Six- to twelve week old BALB/cAnNCrl mice and C57BL/6NCrl mice were purchased from Charles River UK Ltd and maintained under barrier conditions . PbA parasites , derived from the PbA clone 15cy1 , which had been genetically engineered to express green fluorescent protein ( PbAgfp [47] , referred to here as PbA ) were maintained by passage through naïve mice . For in vivo imaging experiments , seven- to eight weeks old BALB/cJ mice were bred in-house and kept under specific pathogen-free conditions . Transgenic P . berghei ANKA 231c1l parasites expressing luciferase under the control of the ef1-a promoter ( referred here as PbAluc ) were provided by Dr . Christian Engwerda ( Queensland Institute for Medical Research , Brisbane , Australia ) from a stock originally from Leiden , The Netherlands [48] , [49] . Experimental infections were initiated by i . v . inoculation of 104 PbA-parasitised red blood cells ( pRBC ) and infected mice were monitored for neurological symptoms ( paralysis , ataxia , convulsions , and coma occurring between day 6 and 10 post-infection ) as previously described [50] . All mice that developed signs of irreversible pathology were immediately humanely sacrificed and their brains examined for signs of ECM ( see Protocol S1 in Text S1 for additional details ) . Cumulative ECM incidence during the observation period was then reported . Parasitaemia was determined by examination of Giemsa-stained thin blood smears . On various days post-infection , mice were sacrificed and exsanguinated . Moreover , their spleens were removed , and single spleen cell suspensions were prepared by homogenisation through a 70 µm cell strainer ( BD Biosciences ) . CD4+ and CD8+ T cells were purified by magnetic bead sorting ( MACS , Miltenyi Biotec ) . Brain-sequestered leucocytes were isolated from perfused animals as described [51] . Live cells were counted by trypan blue exclusion . Heparinised plasma was stored at −70°C for cytokine quantification . Blocking antibodies to CTLA-4 [UC10-4F10-11] , PD-L1 [9G2] and PD-L2 [TY5] and neutralising antibodies to IFN-γ [XMG1 . 2] and TNF [XT3 . 11] were administered by intraperitoneal injection ( 0 . 4 mg/mouse ) on days −1 , 1 , 3 , 5 and 7 of infection . Depleting antibodies to CD4 [GK1 . 5] and CD8 [53 . 6 . 72] were administered by intraperitoneal injection ( 0 . 25 mg/mouse ) on days −1 , 1 , 4 and 6 ( or on days 4 and 6 ) of infection . All antibodies were rat-α-mouse IgG and were obtained from BioXCell; control rat IgG was obtained from Pierce . Antibodies [clones] for cell-surface staining were obtained from eBiosciences ( α-mouse CD4 [GK1 . 5] , CD8 [53 . 6-7] , CD11a [M17/4] , CD11c [N418] , CD44 [IM7] , CD62L [MEL-14] , CD71 [R17217] , CD273/PD-L2 [122] , CD274/PD-L1 [MIH5] , CD279/PD-1 [RMP1-30] ) , F4/80 [BM8] or BD Biosciences ( α-mouse CD3 [145-2C11] , CD4 [RM4-5] and CD8 [53-6 . 7] ) . Isolated leucocytes were directly stained according to standard protocols . Antibodies for intracellular staining were obtained from eBiosciences ( α-mouse CD152/CTLA4 [UC10-4B9] , FoxP3 [FJK-16s] , and IFN-γ [XMG1 . 2] ) . Intracellular staining was performed by permeabilising cells with 0 . 1% Saponin/PBS . Cells were analysed using a FACSCalibur or LSR II ( BD Immunocytometry Systems ) and FlowJo software ( TreeStar ) . Plasma cytokines were assayed by cytometric bead array ( mouse inflammation kit; BD Bioscience ) following the manufacturer's protocol . Intracellular IFN-γ was assayed by flow cytometry ( above ) following 5-hour culture of mixed spleen cells in the presence of PMA ( 50 ng/mL ) , ionomycin ( 1 µg/mL ) , and Brefeldin A ( 1 µg/mL ) . Secreted IFN-γ and IL-10 were assayed by conventional ELISA [52] in supernatants of purified CD4+ or CD8+ T cells cultured ( at 105 cells per well ) for 24 or 48 hours respectively in the presence of α-CD3 [clone 145-2C11 , 1 µg/mL] and α-CD28 [clone 37 . 51 , 1 µg/mL] antibodies ( eBioscience ) . Brain and liver tissues were fixed in 10% formaldehyde saline , paraffin-wax embedded , sectioned , stained with haematoxylin and eosin and examined by light microscopy at 20X magnification . Distribution of PbAluc parasites was monitored daily by in vivo imaging ( IVIS; Xenogen , Alameda , California ) . Infected mice were anaesthesised , injected s . c . with 100 µl of D-luciferin potassium salt ( Caliper Life Sciences ) ( 5mg/ml in PBS ) and , two minutes later , bioluminescence images were acquired , with medium binning factor and fields-of-view ( FOV ) of 21 . 7 and 4 cm for the whole body ( ventral ) and head ( dorsal ) , respectively . Imaging time was between 5 to 60 seconds per mouse . In terminal experiments , mice were given a second injection of luciferase substrate and , within 3 minutes , mice were sacrificed ( by cervical dislocation ) . Brains were removed and imaged with 10 cm FOV . To allow comparison of images from different days of the experiment , uninfected mice injected with luciferin were imaged for background subtraction . Bioluminescence in the brains was quantified using Living Imaging 3 . 0 software and expressed as average radiance units ( p/s/cm2/sr ) . Statistical analysis was performed in GraphPad Prism ( GraphPad Software Inc ) . Comparisons between two groups were made using the Mann Whitney test . For comparisons involving more than two groups , statistical significance was determined using the Kruksal Wallis test with Dunn's post-test for multiple comparisons with p<0 . 05 taken as evidence of a significant difference . Differences in survival curves between two groups were analysed using the Log-rank ( Mantel Cox ) test . Differences in cumulative ECM incidence between two groups was analysed using the Fisher's exact test . Bonferroni correction was used to adjust for multiple comparisons within the Log-rank ( Mantel Cox ) and Fisher's exact tests . The Bonferroni-corrected threshold for significance was calculated by dividing the conventionally set level of significance ( 0 . 05 ) by the number of comparisons .
The course of PbA infection was compared in C57BL/6 and BALB/c mice ( Figure 1A ) . PbA-infected C57BL/6 mice became moribund on day 7 post-infection after developing neurological signs . In contrast , the majority of PbA-infected BALB/c mice survived up to day 15 post-infection when they were euthanised due to the development of severe anaemia and moderate parasitaemia . Over the course of three experiments , 100% ( 12/12 ) of C57BL/6 mice but only 25% ( 4/16 ) of BALB/c mice developed signs of ECM ( Figure 1B ) . There was no difference between C57BL/6 and BALB/c mice in peripheral parasitaemia up to day 7 ( Figure 1C ) . To determine if T cell regulatory receptors could explain differences in susceptibility to ECM , intracellular expression of CTLA-4 and surface expression of PD-1 were compared on splenic T cells of BALB/c and C57BL/6 mice at different times after PbA infection ( Figure 2 ) . The proportions of CD4+ and CD8+ T cells expressing either CTLA-4 or PD-1 were similar in the two mouse strains during the first 6 days of infection , but surprisingly , on day 7 – i . e . at exactly the time when susceptible mice began to show signs of ECM – the proportions of both CD4+ and CD8+ T cells expressing CTLA-4 and PD-1 were significantly higher in ECM-susceptible C75BL/6 mice than in ECM-resistant BALB/c mice ( Figure 2A , B ) . Consistent with previous reports in other models that PD-L1 is also expressed on T cells and is further upregulated during activation [4] , PD-L1 was upregulated on virtually all CD4+ and CD8+ T cells by day 5 of PbA infection in both C57BL/6 and BALB/c mice ( Figure S1A in Text S1 ) . In addition , dendritic cells ( Figure S1B in Text S1 ) and macrophages ( Figure S1C in Text S1 ) from both groups of mice upregulated PD-L1 and PD-L2 during infection . The higher frequency of CTLA-4hi and PD-1hi cells in infected C57BL/6 mice correlated with a higher proportion of CD4+ and CD8+ splenic T cells able to secrete IFN-γ ( Figure 2C ) . Furthermore , T cells from C57BL/6 mice expressed significantly higher levels of effector and activation markers: CD11ahi , a surrogate marker for polyclonal , antigen-experienced CD8+ T cells [53] , [54] and CD62Llo , a commonly used marker of effector cells [53] , [55] ( Figure S2 in Text S1 ) . In both strains of infected mice , and in both CD4+ and CD8+ T cells , CTLA-4 and PD-1 seem to be co-expressed on the same cells ( Figure 2D , Figure S3 in Text S1 ) , indicating that the two receptors may function co-operatively . Consequently , CTLA-4hi/PD-1hi expression on T cells coincided with CD11ahi ( Figure S3B , E in Text S1 ) and CD62Llo ( Figure S3C , F in Text S1 ) suggesting that these are activated cells induced in response to infection . Moreover , the vast majority of CTLA-4hi and PD-1hi CD4+ T cells were FoxP3- , indicating that they are not classical regulatory T cells ( Figure 2E , F ) . Together , these data indicate that CTLA-4 , PD-1 and PD-L1 are upregulated on activated T cells during PbA infection in both ECM-susceptible and ECM-resistant mice . However , the very high degree of T cell activation in C57BL/6 mice may provide a situation where positive signals override physiological levels of immune regulation mediated by CTLA-4 and PD-1 such that they are unable to prevent immune-mediated pathology . To determine whether the CTLA-4 and PD-1 pathways modulate ECM pathogenesis in the ECM-resistant strain , the outcome of PbA infection was compared between control mice and mice treated with blocking antibodies to PD-L1 or CTLA-4 . We decided to focus on CTLA-4 and PD-1/PD-L1 particularly because both of these inhibitory pathways have been extensively studied in chronic infections but a comparison of the two pathways during an acute infection was lacking . BALB/c mice treated with either α-CTLA-4 ( 20/20; 100% ) or α-PD-L1 ( 20/22; 90 . 9% ) developed classical neurological signs of ECM and were euthanised on days 7–8 post-infection or days 8–10 post-infection , respectively , whereas control mice ( treated with rat IgG or PBS ) survived for up to two weeks and were euthanised due to severe anaemia ( Figure 3A , B ) . Importantly , antibody treatment had no effect on parasitaemia ( Figure 3C ) . The survival curves for α-CTLA-4 and α-PD-L1-treated mice differ significantly from the control mice . In addition , the survival curves for α-CTLA-4 and α-PD-L1-treated mice were also significantly different from each other . Since PD-1 also binds another ligand , PD-L2 , the outcome of PbA infection was compared between control mice and mice treated with α-PD-L2 . As shown in Figure S4 in Text S1 , α-PD-L2 treatment had no effect on the course of infection or the pathological outcome of PbA-infection in BALB/c mice . Consistent with their development of neurological signs of ECM , numbers of arrested CD8+ T cells were significantly higher in brain microvessels of α-CTLA-4-treated and α-PD-L1-treated BALB/c mice than in brains of control mice ( Figure 3D ) . Moreover , histological examination revealed more frequent petechial haemorrhages and a higher proportion of cerebral blood vessels plugged with parasitised red blood cells in brains of BALB/c mice treated with α-CTLA-4 or α-PD-L1 antibodies than in brains of control ( treated with rat IgG or PBS or left untreated ) mice ( Figure 3E , Table 1 ) as well as increased numbers of pigmented ( parasite-containing ) macrophages in their livers ( Table 1 ) . The accumulation of parasitised erythrocytes in the microvasculature of the brain is a cardinal feature of cerebral pathology in both human cerebral malaria [56] and ECM in mice [42] , [43] . To further quantify the effects of CTLA-4 and PD-L1 blockade on parasite accumulation , α-CTLA-4- and α-PD-L1 antibody-treated BALB/c mice ( and controls ) were infected with PbAluc ( transgenic PbA expressing luciferase ) and whole body , head and brain parasite burdens were quantified by bioluminescence at the onset of signs of ECM . The course of PbAluc infection in control , α-PD-L1- and α-CTLA-4-treated mice ( Figure 4A–C ) was similar to the course of PbA infection ( Figure 3A–C ) , confirming that insertion of the luciferase gene had not significantly altered the basic biology of the parasite , although the onset of ECM was slightly delayed ( α-CTLA-4-treated mice developed ECM on day 10 post infection; α-PD-L1-teated mice developed ECM on day 11 and control mice were euthanised on day 18 ) . Nevertheless , after day 7 of infection , whole body ( Figure 4D ) , head ( Figure 4E ) and isolated brain ( Figure 4F , G ) parasite burdens were significantly higher in α-CTLA-4- and α-PD-L1-treated mice than in control mice . It should be noted that the detection of a weak luminescence signal in the brains of control ( no ECM ) mice in Figure 4F simply reflects the presence of luminescent parasites in the circulating blood in all organs . These results show that , despite similar peripheral parasite burden in control and antibody-treated mice , overall parasite biomass is significantly increased when the CTLA-4/PD-1 regulatory pathways are blocked . Taken together , these data indicate that blockade of either the inhibitory receptor CTLA-4 or PD-L1 , leads to striking immune pathology , with all the phenotypic characteristics of ECM , in otherwise ECM-resistant mice . Since ECM is known to be a consequence of T cell-mediated inflammation in susceptible C57BL/6 mice[57] , the development of neurological signs of ECM - together with CD8+ T cell infiltration and parasite accumulation in the microvasculature of the brain - in PbA-infected BALB/c mice after α-CTLA-4 or α-PD-L1 antibody treatment was suggestive of an enhanced inflammatory T cell response . To explore this hypothesis , we assessed levels of splenic T cell activation on day 7 post-infection ( Figure S5 in Text S1 ) . The proportions of splenic CD4+ and CD8+ T cells expressing an activated phenotype [CD71+ ( transferrin receptor ) and CD4+ T cells expressing CD44+ ( Pgp-1 ) ] were generally higher in α-CTLA-4-treated than in control mice ( Figure S5 in Text S1 ) . Notably , the proportions of splenic CD8+ T cells expressing CD11a+ was generally higher in both α-CTLA-4 and α-PD-L1-treated mice than in control mice . A trend was observed for higher proportions of splenic CD8+ T cells to express CD62L− and CD11a+ in α-PD-L1-treated than in control mice . Thus , the slightly earlier onset of ECM in the α-CTLA-4-treated mice than in the α-PD-L1-treated mice ( Figures 3A and 4A ) correlates with the slightly higher levels ( statistically significant ) of T cell activation in the α-CTLA-4-treated mice . Plasma concentrations of cytokines and chemokines ( measures of systemic inflammation ) peaked on day 5 after infection ( Figure 5 ) . Plasma concentrations of IFN-γ , MCP-1 and IL-10 were significantly higher in α-CTLA-4- and α-PD-L1-treated mice than in untreated control mice . Concentrations of TNF and IL-6 were significantly higher in α-CTLA-4-treated mice than in either of the other two groups , and importantly , TNF and IL-6 levels did not differ between control mice and α-PD-L1-treated mice ( Figure 5A ) . To determine whether differences in plasma cytokine levels were related to differences in CD4+ and CD8+ T cell cytokine secretion , IFN-γ and IL-10 were quantified in culture supernatants of purified CD4+ and CD8+ T cells stimulated for 48 and 24 hours , respectively with α-CD3/α-CD28 antibodies . CD4+ T cells isolated on days 5 and 7 post-infection from α-CTLA-4-treated mice secreted significantly more IFN-γ and IL-10 than did CD4+ T cells from control mice ( Figure 5B ) . In addition , CD4+ T cells isolated from day 5 post-infection from α-PD-L1-treated mice secreted more IFN-γ than did CD4+ T cells from control mice . Furthermore , CD4+ T cells isolated from α-CTLA-4-treated mice secreted significantly more IFN-γ ( day 7 ) and IL-10 ( days 5 and 7 ) than did CD4+ T cells from α-PD-L1 treated mice . Similarly , CD8+ T cells isolated from α-CTLA-4 ( day 7 post-infection ) - and α-PD-L1 ( days 5 and 7 post-infection ) -treated mice secreted significantly more IFN-γ than did CD8+ T cells from control mice ( Figure 5C ) . CD8+ T cells isolated on day 5 post infection from α-PD-L1 -treated mice secreted significantly more IFN-γ than did CD8+ T cells from α-CTLA-4 treated mice . As further confirmation that the changes in IFN-γsecretion were due to changes in T cell function , spleen cells collected from treated mice on day 7 of infection were analysed by intracellular cytokine staining following short-term stimulation with PMA/ionomycin ( Figure 5D–G ) . Consistent with the secreted cytokine data ( above ) the proportions of IFN-γ+splenic CD4+ and CD8+ T cells were higher in α-CTLA-4- and α-PD-L1-treated mice than in control mice . Blockade of either CTLA-4 or PD-L1 renders normally resistant BALB/c mice fully susceptible to ECM , and this is associated with increased levels of activation and inflammatory cytokine secretion in both the CD4+ and CD8+ T cell populations , consistent with the hypothesis that signalling through both the CTLA-4 and PD-1 pathways is required to down-regulate T cell reactivity and thereby prevent ECM . To determine whether CD4+ or CD8+ T cell populations ( or both ) are the targets of CTLA-4 and PD-1 mediated regulation , CTLA-4 and PD-L1 blockade were combined with in vivo depletion of CD4+ or CD8+ cells ( Figure 6 ) . Depletion of CD8+ cells before and during PbA infection ( α-CD8 antibodies administered on days -1 , 0 , +4 and +6 of infection; full course ) or just prior to the expected onset of neurological signs ( α-CD8 antibodies administered on Days +4 and +6 of infection; late ) completely abrogated the development of ECM in both α-CTLA-4-treated ( Figure 6A ) and α-PD-L1-treated ( Figure 6B ) mice; instead , CD8-depleted mice developed severe anaemia and were euthanised significantly later than non-depleted mice . In contrast , although depletion of CD4+ cells throughout infection ( α-CD4 antibodies administered on days −1 , 0 , +4 and +6 of infection ) led to a delay in onset of ECM in α-CTLA-4 treated mice ( day 8–11 as compared to day 7–8 in α-CTLA-4 treated mice but were not given α-CD4 antibodies ) , CD4+ depletion later in infection ( α-CD4 antibodies administered on days +4 and +6 of infection ) in α-CTLA-4 treated mice had no effect on the development of ECM . These results are entirely consistent with data from C57BL/6 mice indicating that CD4+ T cells play an essential helper role in priming CD8+ T cells in the first 4 or 5 days of infection but after this time only CD8+ T cells are required for initiation of ECM [40] . Among α-PD-L1 treated mice , neither CD4+ depletion throughout infection nor CD4+ depletion later in infection had any significant effect on the development of ECM . These results reveal fundamental differences between the CTLA-4 and PD-1 pathways and suggest that the CD4+ T cells in this infection model are effectively regulated via the CTLA-4/B-7 pathway but are relatively unaffected to regulation via the PD-1/PD-L1 pathway . Moreover , additional pairwise analysis comparing the results of CTLA-4 blockade and PD-L1 blockade in Figure 6 indicate that the effects of CTLA-4 blockade and PD-L1 blockade are similar ( i . e . not significantly different ) for all treatment regimes with the exception that CTLA-4 blockade combined with late-stage CD4 depletion is significantly different than is PD-L1 blockade combined with late-stage CD4 depletion ( p = 0 . 0006; Bonferroni correction indicates that the threshold for significance is p = 0 . 006 ) . Both IFN-γ and TNF have been implicated in the development of ECM [38] , [58] . To determine their roles in the induction of ECM in α-CTLA-4- and α-PD-L1-treated PbA-infected BALB/c mice , in vivo treatment with neutralising α-IFN-γ or α-TNF antibodies throughout infection ( days −1 , 1 , 3 , 5 , 7 ) was performed ( Figure 6 C , D ) . Among α-CTLA-4 treated mice , neutralisation of IFN-γ abrogated the development of ECM , while neutralisation of TNF had a small but significant effect on the survival curve . Among the α-PD-L1 treated mice , neutralisation of IFN-γ also abrogated the development of ECM , but neutralisation of TNF had no significant effect .
Regulation of effector T cell function is crucial for immune homeostasis during infection . Immune homeostasis is maintained , in part , by the negative regulators of T cell activation , CTLA-4 and PD-1 . While PD-1/PD-L1 - and to some extent the CTLA-4/B-7 pathway - negatively regulate T cell responses during chronic infections , their roles in acute infections are much less clear . In addition , there is limited information available for the role of PD-1/PD-L2 during infection . The extent to which these pathways may regulate different aspects of the T cell response to acute infections in animals with differing susceptibility to immunopathology is not known . Whilst other pathways of immune regulation may contribute to the outcome of virulent malaria infections , only modest effects on PbA-induced ECM were observed after IL-10 neutralisation in BALB/c mice [59] . Similarly , depletion of regulatory T cells using α-CD25 antibodies did not increase mortality rates in BALB/c mice during primary infection with PbA [60] . Therefore , in this study , we directly compared the roles of the CTLA-4 and PD-1/PD-L pathways during acute infection with a virulent rodent malaria parasite , PbA in ECM-susceptible ( C57BL/6 ) and ECM-resistant ( BALB/c ) mice . We initially hypothesised that inadequate T cell expression of CTLA-4 and PD-1 during infection , leading to overproduction of Th-1 cytokines and migration of activated CD8+ T cells to the brain , could explain susceptibility to ECM . Interestingly , contrary to our expectations , proportions of CD4+ and CD8+ T cells expressing CTLA-4 and PD-1 were significantly higher in ECM-susceptible C57BL/6 mice than in ECM-resistant BALB/c mice , and CTLA-4 and PD-1 expression were positively correlated with IFN-γ secretion as well as with the CD11ahi and CD62Llo phenotype . Indeed , CTLA-4 and PD-1 expression coincided with expression of the activation markers CD11ahi and CD62Llo . The expression of both CD11ahi and CD62Llo were used as surrogate T cell activation markers due to a paucity of defined CD8+ or CD4+ T cell epitopes associated with pathological responses . Besides the expansion of parasite-specific T cells , there is likely to be non-specific bystander activation . However , it remains unclear whether T cells activated in a bystander manner can contribute to ECM pathogenesis . Studies with ovalbumin-expressing PbA and ovalbumin-specific transgenic T cells suggest that this is possible since transfer of antigen-specific effector memory cells into mice deficient of recombination-activating gene is not always sufficient to induce ECM [61] , [62] . Indeed , in their study , Miyakoda et al specifically analysed non-specific activation of CD8 T cells [62] and found that while non-specific activation occurred it was at a much lower level compared to specific activation . Our results raise the intriguing and important question of why the very efficient activation of CTLA-4 and PD-1 pathways in C57BL/6 mice fails to protect them from acute immune pathology . One interpretation is that T cell activation is so extensive in ECM-susceptible C75BL/6 mice that positive T cell-derived signals override physiological levels of immune inhibition mediated by CTLA-4 and PD-1 . Another interpretation of this data is that in ECM-susceptible mice , CTLA-4 and PD-1 are induced on highly activated T cells , but that their down-stream signalling is impaired . Further studies are required to understand why these potent immunoregulatory pathways are unable to control T cell activation in animals with severe malarial immunopathology and to determine whether these pathways play roles in humans with cerebral malaria . Strikingly , however , in vivo blockade of either CTLA-4 or PD-1/PD-L1 , but not PD-1/PD-L2 , rendered otherwise resistant BALB/c mice fully susceptible to ECM . Treated animals exhibited characteristic neurological signs , their brains revealed the cardinal features of ECM ( haemorrhages , CD8+ T cell arrest and parasite intravascular accumulation in the microvasculature ) and there was clear evidence of excessive systemic inflammation and T cell activation . Thus , CTLA-4 and PD-1/PD-L1 play essential , independent and non-redundant roles in preventing ECM in resistant animals; this reinforces the urgent need to compare the activation and down-stream effects of these pathways in humans with or without severe malaria pathology . It is noteworthy that in vivo blockade of either CTLA-4 did not affect the expression of PD-1 on T cells; in vivo blockade of PD-1/PD-L1 did not affect the expression of CTLA-4 on T cells ( data not shown ) . This study reveals – for the first time - evidence of subtle but important differences in the effects of CTLA-4-mediated and PD-1/PD-L1-mediated immune regulation during acute infections . Infected animals treated with α-CTLA-4 always succumb to infection significantly earlier than animals treated with α-PD-L1 . In addition , disruption of the CTLA-4 pathway led to higher levels of T cell activation , significantly higher levels of circulating TNF and IL-6 and , accordingly , earlier onset of ECM than did blockade of the PD-1/PD-L1 pathway . Furthermore , we observed that depletion of CD4+ T cells during PbA infection abrogated the effects of α-CTLA-4 treatment but had no effect on α-PD-L1 treatment . Thus , our study suggests that CTLA-4 may be much more effective than PD-1/PD-L1 at regulating CD4+ T cells particularly in this experimental model of cerebral malaria , although this hypothesis remains to be directly tested . Since CD4+ T cells are essential for activation of CD8+ T cells and for their arrest in the brain during ECM [40] , effective regulation of CD4+ T cells by CTLA-4 is likely to interrupt the chain of events leading to ECM at a much earlier stage of infection than is regulation of CD8+ T cell activity ( by either CTLA-4 or PD-1/PD-L1 ) . It should be noted however , that in other models - such as during M . tuberculosis infection - the PD-1/PD-L1 pathway has a direct effect on CD4+ T cells in preventing T cell-driven exacerbation of infection [63] . Our study thus raises the intriguing hypothesis that in the our malaria model , CTLA-4 may be the primary regulatory pathway for CD4+ T cells , thereby indirectly affecting CD8+ T cell responses , whereas PD-1/PD-L1 preferentially and directly fulfils this role for CD8+ T cells . This notion is consistent with published data suggesting that CD8+ T cells are less dependent upon costimulation through the CD28/CTLA-4/B7 axis than are CD4+ T cells [64]-[66] and that blockade of the PD-1/PD-L1 pathway restores the effector functions of CD8+ T cells in the absence of CD4+ T cell help [16] , [67] . Such a hypothesis is also consistent with associations between PD-1/PD-L1 expression and inability to control chronic viral infections such as LCMV [16] and rabies [30] , the failure of CTLA-4 blockade to ameliorate CD8+ T cell exhaustion during LCMV infection [16] , and the preferential expression of CTLA-4 on CD4+ rather than CD8+ cells during HIV infection [17] . In addition , in LCMV-infected mice that lacked CD4+ T cell-help , blockade of the PD-1/PD-L1 pathway reinvigorated the ‘helpless’ CD8+ T cells and allowed them to function as effector cells [16] . Our observation that CTLA-4 blockade enhances susceptibility to ECM in BALB/c mice but had no effect on the outcome of infection in C57BL/6 mice is somewhat at odds with a previous study in which CTLA-4 blockade was shown to increase susceptibility to ECM , but in C57BL/6 mice [68] . The apparent discrepancy may be explained by the unusually slow kinetics of PbA infection in the study of Jacobs et al and by the fact that the majority of the control mice survived for at least 20 days after infection and did not develop ECM [68]; in our hands , and those of most other investigators , PbA uniformly causes death from ECM within 10 days in all C57BL/6 mice . It is not possible to distinguish whether this reflects differences in susceptibility between colonies of C57BL/6 mice or differences in virulence of the parasite isolates ( the PbA parasites used in the study of Jacobs et al appear to be highly attenuated ) , but the message is clear: the virulence of mouse/parasite combinations that do not normally lead to ECM is significantly exacerbated by CTLA-4 blockade . In addition to the very important observations emanating from our direct side-by-side comparison of the roles of CTLA-4 and PD-1/PD-L1 , we have significantly extended our understanding of the role of T cell regulatory pathways during malaria infection by thoroughly characterising the effects of regulatory blockade on both CD4+ and CD8+ T cells . The onset of ECM in BALB/c mice following CTLA-4 or PD-1/PD-L1 blockade was accompanied by elevated production of pro-inflammatory cytokines and by increased migration of activated CD8+ T cells to the brain . Neutralisation of IFN-γ or depletion of CD8+ T cells during PbA infection was shown to reverse the pathologic effects of the inhibitory pathway blockade , confirming that the aetiology of ECM in the BALB/c mice is similar to that in susceptible C57BL/6 mice [40] . These experiments not only identify the primary targets of CTLA-4- and PD-1/PD-L1-mediated regulation as being pro-inflammatory T cells , but also re-emphasise that CD8+ T cells and IFN-γ are critical effectors of ECM . In summary , we have revealed essential , independent and non-redundant roles for CTLA-4/B-7 and PD-1/PD-L1 pathways in regulating T cell-mediated pathology and host resistance to PbA-induced ECM . Exploration of the relationship between T cell regulatory pathways and outcome of malaria infection in humans is clearly now a priority; such studies will need to go beyond simple characterisation of CTLA-4 or PD-1 expression [69] , [70] and consider the potential for genetic variation in the receptors , their ligands and downstream signalling molecules to affect the outcome of infection . | T cells are part of the body's defense system in response to infection . However , once the infection has been suitably controlled , these T cells must be switched off . Inhibitory pathways , such as CTLA-4 and PD-1 , are known to send the ‘turn off’ signal to T cells during chronic infections . However , their roles in acute infections , such as malaria , are unclear . We compared the function of these inhibitory pathways in mice that are either susceptible or resistant to severe malarial disease ( cerebral malaria ) . Strikingly , we found that receptors for CTLA-4 and PD-1 are more highly expressed in T cells from susceptible mice than from resistant mice . Therefore , cerebral malaria develops despite the high expression of these inhibitory receptors . Moreover , we demonstrated that blocking these inhibitory receptors in the resistant mice increased the function of T cells , which in turn led to the characteristic signs of cerebral malaria . Finally , reminiscent of what is known for the susceptible strain , we confirmed that certain T cells ( CD8+ ) and molecules ( IFN-γ ) are crucial to the development of cerebral malaria in the otherwise resistant mice . Thus , the CTLA-4 and PD-1 inhibitory pathways have essential , independent and non-redundant roles in regulating the body's complex response to malaria . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
]
| [
"medicine",
"infectious",
"diseases",
"immune",
"cells",
"immunity",
"immunology",
"biology",
"parasitic",
"diseases",
"immune",
"system"
]
| 2012 | The CTLA-4 and PD-1/PD-L1 Inhibitory Pathways Independently Regulate Host Resistance to Plasmodium-induced Acute Immune Pathology |
Passive acoustic sensing has emerged as a powerful tool for quantifying anthropogenic impacts on biodiversity , especially for echolocating bat species . To better assess bat population trends there is a critical need for accurate , reliable , and open source tools that allow the detection and classification of bat calls in large collections of audio recordings . The majority of existing tools are commercial or have focused on the species classification task , neglecting the important problem of first localizing echolocation calls in audio which is particularly problematic in noisy recordings . We developed a convolutional neural network based open-source pipeline for detecting ultrasonic , full-spectrum , search-phase calls produced by echolocating bats . Our deep learning algorithms were trained on full-spectrum ultrasonic audio collected along road-transects across Europe and labelled by citizen scientists from www . batdetective . org . When compared to other existing algorithms and commercial systems , we show significantly higher detection performance of search-phase echolocation calls with our test sets . As an example application , we ran our detection pipeline on bat monitoring data collected over five years from Jersey ( UK ) , and compared results to a widely-used commercial system . Our detection pipeline can be used for the automatic detection and monitoring of bat populations , and further facilitates their use as indicator species on a large scale . Our proposed pipeline makes only a small number of bat specific design decisions , and with appropriate training data it could be applied to detecting other species in audio . A crucial novelty of our work is showing that with careful , non-trivial , design and implementation considerations , state-of-the-art deep learning methods can be used for accurate and efficient monitoring in audio .
There is a critical need for robust and accurate tools to scale up biodiversity monitoring and to manage the impact of anthropogenic change [1 , 2] . Modern hardware for passive biodiversity sensing such as camera trapping and audio recording now enables the collection of vast quantities of data relatively inexpensively . In recent years , passive acoustic sensing has emerged as a powerful tool for understanding trends in biodiversity [3–6] . Monitoring of bat species and their population dynamics can act as an important indicator of ecosystem health as they are particularly sensitive to habitat conversion and climate change [7] . Close to 80% of bat species emit ultrasonic pulses , or echolocation calls , to search for prey , avoid obstacles , and to communicate [8] . Acoustic monitoring offers a passive , non-invasive , way to collect data about echolocating bat population dynamics and the occurrence of species , and it is increasingly being used to survey and monitor bat populations [7 , 9 , 10] . Despite the obvious advantages of passive acoustics for monitoring echolocating bat populations , its widespread use has been hampered by the challenges of robust identification of acoustic signals , generation of meaningful statistical population trends from acoustic activity , and engaging a wide audience to take part in monitoring programmes [11] . Recent developments in statistical methodologies for estimating abundance from acoustic activity [4 , 12 , 13] , and the growth of citizen science networks for bats [9 , 10] mean that efficient and robust audio signal processing tools are now a key priority . However , tool development is hampered by a lack of large scale species reference audio datasets , intraspecific variability of bat echolocation signals , and radically different recording devices being used to collect data [11] . To date , most full-spectrum acoustic identification tools for bats have focused on the problem of species classification from search-phase echolocation calls [11] . Existing methods typically extract a set of audio features ( such as call duration , mean frequency , and mean amplitude ) from high quality search-phase echolocation call reference libraries to train machine learning algorithms to classify unknown calls to species [11 , 14–19] . Instead of using manually defined features , another set of approaches attempt to learn representation directly from spectrograms [20 , 21] . Localising audio events in time ( defined here as ‘detection’ ) , is an important challenge in itself , and is often a necessary pre-processing step for species classification [22] . Additionally , understanding how calls are detected is critical to quantifying any biases which may impact estimates of species abundance or occupancy [12 , 23] . For example , high levels of background noise , often found in highly disturbed anthropogenic habitats such as cities , may have a significant impact on the ability to detect signals in recordings and lead to a bias in population estimates . Detecting search-phase calls by manual inspection of spectrograms tends to be subjective , highly dependent on individual experience , and its uncertainties are difficult to quantify [24] . There are a number of automatic detection tools now available which use a variety of methods , including amplitude threshold filtering , locating areas of smooth frequency change , detection of set search criteria , or based on a cross-correlation of signal spectrograms with a reference spectrogram [see review in 11] . While there are some studies that analyse the biases of automated detection ( and classification ) tools [25–30] , this is generally poorly quantified , and in particular , there is very little published data available on the accuracy of many existing closed source commercial systems . Despite this , commercial systems are commonly used in bat acoustic survey and monitoring studies , albeit often with additional manual inspection [9 , 10] . This reliance on poorly documented algorithms is scientifically undesirable , and manual detection of signals is clearly not scalable for national or regional survey and monitoring . In addition , there is the danger that manual detection and classification introduces a bias towards the less noisy and therefore more easily identifiable calls . To address these limitations , a freely available , transparent , fast , and accurate detection algorithm that can also be used alongside other classification algorithms is highly desirable . Here , we develop an open source system for automatic bat search-phase echolocation call detection ( i . e . localisation in time ) in noisy , real world , recordings . We use the latest developments in machine learning to directly learn features from the input audio data using supervised deep convolutional neural networks ( CNNs ) [31] . CNNs have been shown to be very successful for classification and detection of objects in images [32 , 33] . They have also been applied to various audio classification tasks [34–36] , along with human speech recognition [37 , 38] . Although CNNs are now starting to be used for bioacoustic signal detection and classification tasks in theoretical or small-scale contexts ( e . g . bird call detection ) [39] , to date there have been no application of CNN-based tools for bat monitoring . This is mainly due to a lack of sufficiently large labelled bat audio datasets for use as training data . To overcome this , we use data collected and annotated by thousands of citizen scientists as part of our Indicator Bats Programme [7] and Bat Detective ( www . batdetective . org ) . We validate our system on three different challenging test datasets from Europe which represent realistic use cases for bat surveys and monitoring programmes , and we present an example real-world application of our system on five years of monitoring data collected in Jersey ( UK ) .
We created a detection system to determine the temporal location of any search-phase bat echolocation calls present in ultrasonic audio recordings . Our detection pipeline consisted of four main steps ( Fig 1 ) as follows: ( 1 ) Fast Fourier Transform Analysis—Raw audio ( Fig 1A ) was converted into a log magnitude spectrogram ( FFT window size 2 . 3 milliseconds , overlap of 75% , with Hanning window ) , retaining the frequency bands between 5kHz and 135kHz ( Fig 1B ) . Recordings with a sampling rate of 44 . 1kHz , time expansion factor of 10 , and 2 . 3ms FFT window , resulted in a window size of 1 , 024 samples . We used spectrograms rather than raw audio for analysis , as it provides an efficient means of dealing with audio that has been recorded at different sampling rates . Provided the frequency and time bins of the spectrogram are of the same resolution , audio with different sampling rates can be input into the same network . ( 2 ) De-noising–We used the de-noising method of [40] to filter out background noise by removing the mean amplitude in each frequency band ( Fig 1C ) , as this significantly improved performance . ( 3 ) Convolutional Neural Network Detection–We created a convolutional neural network ( CNN ) that poses search-phase bat echolocation call detection as a binary classification problem . Our CNNFULL consisted of three convolution and max pooling layers , followed by one fully connected layer ( see Supplementary Information Methods for further details ) . We halved the size of the input spectrogram to reduce the input dimensionality to the CNN which resulted in an input array of size of 130 frequency bins by 20 time steps , corresponding to a fixed length , detection window size of 23ms . We applied the CNN in a sliding window fashion , to predict the presence of a search-phase bat call at every instance of time in the spectrogram ( Fig 1D ) . As passive acoustic monitoring can generate large quantities of data , we required a detection algorithm that would run faster than real time . While CNNs produce state of the art results for many tasks , naïve application of them for detection problems at test time can be extremely computationally inefficient [33] . So , to increase the speed of our system we also created a second , smaller CNN which included fewer model weights that can be run in a fully convolutional manner ( CNNFAST ) ( Supplementary Information Methods , Supplementary Information S1 Fig ) . ( 4 ) Call Detection Probabilities–The probabilistic predictions produced by the sliding window detector tended to be overly smooth in time ( Fig 1D ) . To localise the calls precisely , we converted the probabilistic predictions into individual detections using a non-maximum suppression to return the local maximum for each peak in the output prediction ( Fig 1E ) . These local maxima corresponded to the predicted locations of the start of each search-phase bat echolocation call , with associated probabilities , and were exported as text files . We trained our BatDetect CNNs using a subset of full-spectrum time-expanded ( TE ) ultrasonic acoustic data recorded between 2005–2011 along road-transects by citizen scientists as part of the Indicator Bats Programme ( iBats ) [7] ( see Supplementary Information Methods for detailed data collection protocols ) . During surveys , acoustic devices ( Tranquility Transect , Courtplan Design Ltd , UK ) were set to record using a TE factor of 10 , a sampling time of 320ms , and sensitivity set on maximum , giving a continuous sequence of ‘snapshots’ , consisting of 320ms of silence ( sensor listening ) and 3 . 2s of TE audio ( sensor playing back x 10 ) . As sensitivity was set at maximum , and no minimum amplitude trigger mechanism was used on the recording devices , our recorded audio data contained many instances of low amplitude and faint bat calls , as well as other night-time ‘background’ noises such as other biotic , abiotic , and anthropogenic sounds . We generated annotations of the start time of search-phase bat echolocation calls in the acoustic recordings by uploading the acoustic data to the Zooniverse citizen science platform ( www . zooniverse . org ) as part of the Bat Detective project ( www . batdetective . org ) , to enable public users to view and annotate them . The audio data were first split up into 3 . 84s long sound clips to include the 3 . 2s of TE audio and buffered by sensor-listening silence on either side . We then uploaded each sound clip as both a wav file and a magnitude spectrogram image ( represented as a 512x720 resolution image ) onto the Bat Detective project website . As the original recordings were time-expanded , therefore reducing the frequency , sounds in the files were in the audible spectrum and could be easily heard by users . Users were presented with a spectrogram and its corresponding audio file , and asked to annotate the presence of bat calls in each 3 . 84s clip ( corresponding to 320ms of real-time recordings ) ( Supplementary Information S2 Fig ) . After an initial tutorial ( S1 Video ) , users were instructed to draw bounding boxes around the locations of bat calls within call sequences and to annotate them as being either: ( 1 ) search-phase echolocation calls; ( 2 ) terminal feeding buzzes; or ( 3 ) social calls . Users were also encouraged to annotate the presence of insect vocalisations and non-biotic mechanical noises . Between Oct 2012 and Sept 2016 , 2 , 786 users ( including only the number of users which had registered with the site and performed more than five annotations ) listened to 127 , 451 unique clips and made 605 , 907 annotations . 14 , 339 of these clips were labelled as containing a bat call , with 10 , 272 identified as containing search-phase echolocation calls . Due to the inherent difficulty of identifying bat calls and the inexperience of some of our users , we observed a large number of errors in the annotations provided . How to best merge different annotations for multiple users is an open research question . Instead , we visually inspected a subset of the annotations from our most active user and found that they produced high quality annotations . This top user had viewed 46 , 508 unique sound clips and had labelled 3 , 364 clips as containing bat search-phase echolocation calls ( a representative sample is shown in Supplementary Information S3 Fig ) . From this we randomly selected a training set of 2 , 812 clips , consisting of 4 , 782 individual search-phase echolocation call annotations from Romania and Bulgaria , with which to train the CNNs ( corresponding to data from 347 road-transect sampling events of 137 different transects collected between 2006 and 2011 ) ( Fig 2A ) . Data were chosen from these countries as they contain the majority of the most commonly occurring bat species in Europe [41] . This training set was used for all experiments . The remaining annotated clips from the same user were used to create one of our test sets , iBats Romania and Bulgaria ( Fig 2A and see below ) . Occasionally , call harmonics and the associated main call were sometimes labelled with different start times in the same audio clip . To address this problem , we automatically merged annotations that occurred within 6 milliseconds of each other , making the assumption that they belonged to the same call . We measured the top user’s annotation accuracy on the test set from Romania and Bulgaria compared to the expert curated ground truth . This resulted in an average precision of 0 . 845 ( computed from 455 out of 500 test files this user had labelled ) . This is in contrast with the second most prolific annotator who had an average precision of 0 . 67 ( based on 311 out of 500 files ) . To evaluate the performance of the detection algorithms , we created three different test datasets of approximately the same size ( number and length of clips ) ( Fig 2A and 2B , Supplementary Information S1 Table ) . These datasets were chosen to represent three different realistic use cases commonly used for bat surveys and monitoring programmes and included data collected both along road-transects ( resulting in noisier audio ) , and using static ultrasonic detectors . The test sets were as follows: ( 1 ) iBats Romania and Bulgaria—audio recorded from the same region , by the same individuals , with the same equipment , and sampling protocols as the training set , corresponding to 161 sampling events of 81 different transect routes; ( 2 ) iBats UK—audio recorded from a different region ( corresponding to data from 176 sampling events of 111 different transects recorded between 2005–2011 in the United Kingdom , chosen randomly ) , by different individuals , using the same equipment type , and identical sampling protocols as part of the iBats programme [7] as the training set; and ( 3 ) Norfolk Bat Survey—audio recorded from a different region ( Norfolk , UK ) , by different individuals , using different equipment types ( SM2BAT+ Song Meter , Wildlife Acoustics ) and different protocols ( static devices from random sampling locations ) as part of the Norfolk Bat Survey [9] in 2015 . These data corresponded to 381 sampling events from 246 static recording locations ( 1km2 grid cells ) , randomly chosen . The start times of the search-phase echolocation calls in these three test sets were manually extracted . For ambiguous calls , we consulted two experts , each with over 10 years of experience with bat acoustics . As these data contained a significantly greater proportion of negative ( non-bat calls ) as compared to positive examples ( bat calls ) , standard error metrics used for classification such as overall accuracy were not suitable for evaluating detection . Instead , we report the interpolated average precision and recall of each method displayed as a precision-recall curve [42] . Precision was calculated as the number of true positives divided by the sum of both true and false positives . We consider a detection to be a true positive if it occurred within 10ms of the expert annotated start time of the search-phase echolocation call . Recall was measured as the overall fraction of calls that were present in the audio that were correctly detected . Curves were obtained by thresholding the detection probabilities from zero to one and recording the precision and recall at each threshold . Algorithms that did not produce a continuous output were represented as a single point on the precision-recall curves . We also report recall at 0 . 95 precision , a metric that measures the fraction of calls that were detected while accepting a false positive rate of 5% . Thus a detection algorithm gets a score of zero if it was not capable of retrieving any calls with a precision greater than 0 . 95 . We compared the performance of both BatDetect CNNs to three existing closed-source commercial detection systems: ( 1 ) SonoBat ( version 3 . 1 . 7p ) [43]; ( 2 ) SCAN’R version 1 . 7 . 7 . [44]; and ( 3 ) Kaleidoscope ( version 4 . 2 . 0 alpha4 ) [45] . For SonoBat , calls were extracted in batch mode . We set a maximum of 100 calls per file ( there are never more than 20 search-phase calls in a test file ) , and set ‘acceptable call quality’ and ‘skip calls below this quality’ parameters both to zero , and used an auto filter of 5KHz . For SCAN’R , we used standard settings as follows: setting minimum and maximum frequency cut off at 10 kHz and 125 kHz , respectively; minimum call duration at 0 . 5 ms; and minimum trigger level of 10 dB . We used Kaleidoscope in batch mode , setting ‘frequency range’ to 15-120kHz , ‘duration range’ to 0-500ms , ‘maximum inter-syllable’ to 0ms , and ‘minimum number of pulses’ to 0 . We also compared two other detection algorithms that we implemented ourselves , which are representative of typical approaches used for detection in audio files and in other bat acoustic classification studies: ( 4 ) Segmentation—an amplitude thresholding segmentation method [46] , this is related to the approach of [47]; and ( 5 ) Random Forest–a random forest-based classifier [48] . Where relevant , the algorithms for ( 4 ) and ( 5 ) used the same processing steps as the BatDetect CNNs . For the Segmentation method , we thresholded the amplitude of the input spectrogram resulting in a binary segmentation . Regions that were greater than the threshold St , and bigger than size Sr , were considered as positive instances . We chose the values of St and Sr on the iBats ( Romania and Bulgaria ) test dataset that gave the best test results to quantify its best case performance . For the Random Forest algorithm , as opposed to extracting low dimensional audio features we instead we used the raw amplitude values from the gradient magnitude of the log magnitude spectrogram as a higher dimensional candidate feature set . This enabled the Random Forest to learn features that it deemed useful for detecting calls . We compared the total processing time for each of our own algorithms , and timings were calculated on a desktop with an Intel i7 processor , 32Gb of RAM , and a Nvidia GTX 1080 GPU . With the exception of the BatDetect CNNFULL , which used a GPU at test time , all the other algorithms were run on the CPU . To demonstrate the performance of our method in a large-scale ecological monitoring application , we compared the number of bat search-phase echolocation calls found using our BatDetect CNNFAST algorithm to those produced from a commonly used commercial package using SonoBat ( version 3 . 1 . 7p ) [43] as a baseline , using monitoring data collected in iBats programme in Jersey , UK from 2011–2015 . Audio data was collected twice yearly ( July and August ) from 11 road-transect routes of approximately 40km by volunteers using the iBats protocols ( Supplementary Information , Supplementary Methods ) , corresponding to 5 . 7 days of continuous TE audio over five years ( or 13 . 75 hours of real-time data ) . For the BatDetect CNNFAST analysis , we ran the pipeline as described above , using a conservative probabilistic threshold of 0 . 90 ( so as to only include high precision predictions ) . Computational analysis timings for the CNNFAST for this dataset were calculated as before . For the comparison to SonoBat , we used the results from an existing real-world analysis in a recent monitoring report [49] , where the audio files were first split into 1 min recordings , and then SonoBat was used to detect search-phase calls and to fit a frequency-time trend line to the shape of the call [49] . All fitted lines were visually inspected and calls where the fitted line included background noise or echoes , were rejected . Typically , monitoring analyses group individual calls into sequences ( a bat pass ) before analysis . To replicate that here in both analyses , individual calls were assumed to be part of the same call sequence ( bat pass ) if they occurred within the same 3 . 84s sound clip and if the sequence continued into subsequent sound clips . We compared number of bat calls and passes detected per transect sampling event across the two analyses methods using generalized linear mixed models ( GLMM ) using lme4 [50] in R v . 3 . 3 . 3 [51] in order to control for the spatial and temporal non-independence of our survey data ( Poisson GLMM including analysis method as a fixed effect and sampling event , transect route and date as random effects ) .
Both versions of our BatDetect CNN algorithm outperformed all other algorithms and commercial systems tested , with consistently higher average precision scores and recall rates across the three different test datasets ( Table 1 , Fig 3A–3C ) . In particular , the CNNs detected a substantially higher proportion of search-phase calls at 0 . 95 precision ( maximum 5% false positives ) ( Table 1 ) . All the other algorithms underestimated the number of search-phase echolocation calls in each dataset , except Segmentation , which produced high recall rates but with low precision ( a high number of false positives ) . The CNNs relative improvement compared to other methods was higher on the road transect datasets ( iBats Romania & Bulgaria; iBats UK; Table 1 , Fig 3A and 3B ) . Overall the performance of CNNFAST was slightly worse than the larger CNNFULL across all test datasets , with the exception of improved recall at 0 . 95 precision in the static Norfolk Bat Survey dataset ( Fig 3C , Table 1 ) . Precision scores for all commercial systems ( SonoBat , SCAN’R and Kaleidoscope ) were reasonably good across all test datasets ( >0 . 7 ) ( Fig 3A–3C ) . However , this was at the expense of recall rates , which were consistently lower than for the CNNs and Random Forest , where the maximum recall rates were 44–60% of known calls detected ( Fig 3C ) . The recall rates fell to a maximum of 25% of known calls for the road transect datasets ( Fig 3A and 3B ) . The Random Forest baseline performed significantly better than the commercial systems on the two challenging roadside recorded datasets ( Fig 3A and 3B ) . This is a result of the training data and the underlying power of the model . However , unlike our CNNs , the simple tree based model is limited in the complexity of the representations it can learn , which results in worse performance . For the static Norfolk Bat Survey its performance more closely matches that of SonoBat , but with improved recall . CNNFULL , CNNFAST , Random Forest , and the Segmentation algorithms took 53 , 9 . 5 , 11 , and 17 seconds respectively , to run the full detection pipeline on the 3 . 2 minutes of full spectrum iBats Romania and Bulgaria test dataset . Compared to CNNFULL there was therefore a significant decrease in the time required to perform detection using CNNFAST , which was also the fastest of our methods overall . Notably , close to 50% of the CNN runtime was spent generating the spectrograms for detection , making this the most computationally expensive stage in the pipeline . Our BatDetect CNNFAST algorithm detected a significantly higher number of bat echolocation search-phase calls per transect sampling event , across 5 years of road transect data from iBats Jersey , compared to using SonoBat ( CNNFAST mean = 107 . 69 , sd = 48 . 01; SonoBat mean = 64 . 95 , sd = 28 . 53 , Poisson GLMM including sampling event , transect route and date as random effects p<2e-16 , n = 216 ) ( Fig 4 , Supplementary Information S2 Table ) . The differences between the two methods for bat passes was much smaller per sampling event , although CNNFAST still detected significantly more passes per transect recording ( CNNFAST mean = 29 . 57 , sd = 11 . 26; SonoBat mean = 27 . 27 , sd = 10 . 85; Poisson GLMM including sampling event , transect route and date as random effects p = 0 . 00143 , n = 216 ) ( Fig 4 , Supplementary Information S2 Table ) . Running only on the CPU , the CNNFAST algorithm took 24 seconds to process one hour of time-expanded audio .
The BatDetect deep learning algorithms show a higher detection performance ( average precision and recall ) for search-phase echolocation calls with the test sets , when compared to other existing algorithms and commercial systems . In particular , our algorithms were better at detecting calls in road-transect data , which tend to contain noisier recordings , suggesting that these are extremely useful tools for measuring bat abundance and occurrence in such datasets . Road-transect acoustic monitoring is a useful technique to assess bat populations over large areas and programmes have now been established by government and non-government agencies in many different countries [e . g . , 7 , 52 , 53–55] . Noisy sound environments are also likely to be a problem for other acoustic bat monitoring programmes . For example , with the falling cost and wider availability of full-spectrum audio equipment , the range of environments being acoustically monitored is increasing , including noisy urban situations [56 , 57] . Individual bats further from the microphone are less likely to be detected as their calls are fainter , and high ambient noise levels increase call masking and decrease call detectability . Additionally , a growth in open-source sensor equipment for bat acoustics using very cheap MEMs microphones [58] may also require algorithms able to detect bats in lower quality recordings , which may have a lower signal to noise ratio or a reduced call band-width due to frequency-dependent loss . Our open-source , well documented algorithms enable biases and errors to be directly incorporated into any acoustic analysis of bat populations and dynamics ( e . g . occupancy models [e . g . , 23] . The detections with BatDetect can be directly used as input for population monitoring programmes when species identification is difficult such as the tropics , or to other CNN systems to determine bat species identity when sound libraries are available . Our result that deep learning networks consistently outperformed other baselines , is consistent with the suggestion that CNNs offer substantially improved performance over other supervised learning methods for acoustic signal classification [39] . The major improvement of both CNNs over Random Forest and the three commercial systems was in terms of recall , i . e . increasing the proportion of detected bat calls in the test datasets . Although the precision of the commercial systems was often relatively high , the CNNs were able to detect much fainter and partially noise-masked bat calls that were missed by the other methods , with fewer false positives , and very quickly , particularly with CNNFAST . Previous applications of deep learning networks to bioacoustic and environmental sound recognition have used small and high-quality datasets [e . g . , 35 , 39] . However , our results show that , provided they are trained with suitably large and varied training data , deep learning networks have good potential for applied use in real-world heterogeneous datasets that are characteristic of acoustic wildlife monitoring ( involving considerable variability in both environmental noise and distance of animal from sensor ) . Our comparison of CNNFULL and CNNFAST detectors was favourable , although CNNFAST had a slightly poorer performance showing a trade-off between speed and accuracy . This suggests that CNNFAST could potentially be adapted to work well with on-board low power devices ( e . g . Intel’s Edison device ) to deliver real-time detections . Avoiding the spectrogram generation stage entirely and using the raw audio samples as input [59] , could also speed up performance of the system in the future , as currently over 50% of the CNN test time is taken up by computing spectrograms . While our results have been validated on European bats , no species or region-specific knowledge , or particular acoustic sensor system is directly encoded into our system , making it possible to easily generalise to other systems ( e . g . frequency division recordings ) , regions and species with additional training data . Despite this flexibility , this version of our deep network may be currently biased towards common species found along roads , although the algorithms did perform well on static recordings on a range of common and rare species in a range of habitats in the Norfolk Bat Survey [9] . Nevertheless , in future , extending the training dataset to include annotated bat calls from verified species-call databases to increase geographic and taxonomic coverage , will further improve the generality of our detection tool . Other improvements to the CNN detectors could also be made to lessen taxonomic bias . For example , some bat species have search phase calls longer than the fixed input time window of 23ms of both CNNs ( e . g . horseshoe bats ) . This may limit our ability currently to detect species with these types of calls . One future approach would be to resize the input window [33] , thus discarding some temporal information , or to use some form of recurrent neural network such as a Long Short-Term Memory ( LSTM ) [60] that can take a variable length sequence as input . There are many more unused annotations in the Bat Detective dataset that could potentially increase our training set size . However , we found some variability in the quality of the citizen science user annotations , as in other studies [61] . To make best use of these annotations , we need user models for understanding which annotations and users are reliable [62 , 63] . The Bat Detective dataset also includes annotations of particular acoustic behaviours ( feeding buzzes and social calls ) , which in future can be used to train detection algorithms for different acoustic behaviours [e . g . , 64] . Our evaluation on large-scale ecological monitoring data from Jersey [49] , demonstrated that our open-source BatDetect CNNFAST pipeline performs as well or better ( controlling for spatial and temporal non-independence ) compared with an existing widely-used commercial system ( SonoBat ) that had been manually filtered ( false positives were removed ) . Here we assume that the manually filtered data represents the ground truth , although it may slightly underrepresent the true number of calls due to missing detections on the part of SonoBat . Interestingly , although the CNNFAST consistently detected more of the faint and partially-masked calls , most bat passes are likely to still contain at least one call that is clearly-recorded enough to be detected by SonoBat , meaning that the total number of detected bat passes is similar across the two methods . No manual filtering is performed for CNNFAST , but the increase in detected calls mirrors the results observed in Fig 3 at high thresholds ( i . e . the left of the curves ) , where both CNNs detected 10–20% more calls than SonoBat on the related driving transect based test sets . Our system results in a large reduction processing time—several minutes for our automatic approach compared to several days split between automatic processing and manual filtering as reported by the authors of [49] . Specifically , it takes CNNFAST under 10 seconds to process the 500 files in the iBats Romania and Bulgaria test set compared to 30 minutes for SonoBat in batch mode . This increase in performance both in terms of speed and accuracy is crucial for future large scale monitoring programmes . The results in our monitoring application raises an interesting question—what is the value of the additional detected calls ? Fig 4 shows a large increase in the number of detected calls and a slight increase in the number of detected bat passes . It may be the case that our current heuristic for merging calls into passes is too aggressive and as a result under reports the true number of bats when there were multiple calling at the same time . Further improvements to our system may come from a better understanding of the patterns of search-phase calls within sequences [65] . Instead of the existing heuristic we would ideally also be able to learn the relationship between individual calls and passes from labelled training data . The current generation of algorithms for bat species classification that are based on extracting simple audio features may perhaps not be best suited to make use of the extra calls we detect . However , when large collections of diverse species data become available only relatively minor architectural changes will be required to our detection pipeline to adapt it for species classification ( e . g . changing the final layer of our CNNs ) . As we have already observed for detection , with enough data , representation learning based approaches can also be applied to the problem of species classification with the promise of large increases in accuracy . These extra calls will be invaluable to create more powerful models , enabling them to perform accurately in diverse and challenging real world situations . For some noisy and faint bat calls it may always be difficult to identify them to the species level , and as a result a coarser taxonomic prediction may have to suffice . Our BatDetect search-phase bat call detector significantly outperforms existing methods for localising the position of bat search-phase calls , particularly in noisy audio data . It could be combined with automatic bat species classification tools to scale up the monitoring of bat populations over large geographic regions . In addition to making our system available open source , we also provide three expertly annotated test sets that can be used to benchmark future detection algorithms .
All training and test data , including user and expert annotations , along with the code to train and evaluate our detection algorithms are available on our GitHub page ( https://github . com/macaodha/batdetect ) . | There is a critical need for robust and accurate tools to scale up biodiversity monitoring and to manage the impact of anthropogenic change . For example , the monitoring of bat species and their population dynamics can act as an important indicator of ecosystem health as they are particularly sensitive to habitat conversion and climate change . In this work we propose a fully automatic and efficient method for detecting bat echolocation calls in noisy audio recordings . We show that our approach is more accurate compared to existing algorithms and other commercial tools . Our method enables us to automatically estimate bat activity from multi-year , large-scale , audio monitoring programmes . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion",
"Data",
"reporting"
]
| [
"acoustics",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"engineering",
"and",
"technology",
"audio",
"signal",
"processing",
"applied",
"mathematics",
"signal",
"processing",
"vertebrates",
"neuroscience",
"animals",
"mammals",
"animal",
"signaling",
"and",
"communication",
"simulation",
"and",
"modeling",
"algorithms",
"animal",
"behavior",
"mathematics",
"zoology",
"echolocation",
"research",
"and",
"analysis",
"methods",
"sensory",
"physiology",
"computer",
"and",
"information",
"sciences",
"behavior",
"bioacoustics",
"physics",
"auditory",
"system",
"animal",
"migration",
"eukaryota",
"animal",
"navigation",
"physiology",
"bats",
"biology",
"and",
"life",
"sciences",
"sensory",
"systems",
"physical",
"sciences",
"amniotes",
"organisms",
"acoustic",
"signals"
]
| 2018 | Bat detective—Deep learning tools for bat acoustic signal detection |
Human red blood cells ( RBCs ) lose ∼30% of their volume and ∼20% of their hemoglobin ( Hb ) content during their ∼100-day lifespan in the bloodstream . These observations are well-documented , but the mechanisms for these volume and hemoglobin loss events are not clear . RBCs shed hemoglobin-containing vesicles during their life in the circulation , and this process is thought to dominate the changes in the RBC physical characteristics occurring during maturation . We combine theory with single-cell measurements to investigate the impact of vesiculation on the reduction in volume , Hb mass , and membrane . We show that vesicle shedding alone is sufficient to explain membrane losses but not volume or Hb losses . We use dry mass measurements of human RBCs to validate the models and to propose that additional unknown mechanisms control volume and Hb reduction and are responsible for ∼90% of the observed reduction . RBC population characteristics are used in the clinic to monitor and diagnose a wide range of conditions including malnutrition , inflammation , and cancer . Quantitative characterization of cellular maturation processes may help in the early detection of clinical conditions where maturation patterns are altered .
A typical red blood cell loses 30% of its volume and 20% of its intracellular hemoglobin ( Hb ) over the course of its 100-day lifespan in the circulation [1]–[6] . These physical characteristics control the RBC's ability to fulfill its main function of delivering oxygen to the tissues . With steady state production and recycling rates of about 2 . 5 cells per second in an adult human , these processes must be tightly regulated , but despite RBC availability and many years of research [7] the mechanisms for these volume and hemoglobin loss events are not clear [8] . The complete blood count includes RBC population characteristics and is used to diagnose and monitor almost all diseases and medical conditions . These RBC characteristics are determined in large part by maturation events . Quantifying RBC maturation would help diagnose and monitor many pathologic conditions [9] , [10] . Nutritional deficiencies , inflammation , and cancer often lead to changes in the circulating population of RBCs . RBCs shed vesicles as they circulate , and this process is thought to dominate the changes in the RBC physical characteristics occurring during maturation ( see e . g . , [11]–[14] ) . Because vesicle formation can not be monitored directly in vivo , theoretical studies are essential . To investigate the role of vesiculation in RBC maturation we take advantage of powerful new single-cell measurements ( see e . g . , [15]–[18] ) . Despite the large number of techniques measuring RBC single-cell characteristics , there are not many theoretical studies integrating these data . We develop a model combining biophysical considerations with measurements of volume and hemoglobin content ( as in Figure 1 ) . The model predicts RBC dry mass and dry density characteristics ( i . e . , mass and density of non-water cellular content ) , and we use measurements of those quantities to validate the models and to define the physical requirements for RBC maturation processes . Figure 1 shows the variation in volume and Hb mass for some of the youngest cells ( <2 days old ) called reticulocytes , identified by RNA staining [19] , and the total RBC population from one healthy human . In Table 1 we summarize our estimates for volume and Hb loss in the transition from reticulocyte to mature cell . We first propose a theoretical framework to investigate the effects of vesiculation on biophysical properties of RBCs as they age . In order to apply the theory we estimate the vesiculation rate from existing empirical data . We then describe the requirements for the volume reduction process , for the hemoglobin reduction process , and for the surface area reduction process . We show that vesiculation cannot account for all the volume and Hb lost by the RBCs , as the sum of volume lost or Hb mass lost in the vesicles is considerably smaller than what is lost by the cells . We find that vesiculation can explain the surface area loss and that surface area loss must be coupled to the hemoglobin loss in order to explain the observed dry density profile of red blood cells .
Here we propose a family of stochastic processes that describes how the biophysical properties of a cell change as it ages and sheds vesicles in a leak-less process where all mass lost from the cell is assumed to be in the vesicles . The model describes how a property e . g . , volume , mass , etc . , changes when the quantity is lost in a single vesicle . Thus , ( at age ) is the difference between the initial value and the total quantity lost in vesicles , ( 1 ) is a Poisson process with constant rate , counting the number of vesicles lost by age . We assume that is independent of age and of , as indicated by analysis of vesicles from stored blood [20] . Thus , for simplicity our first approximation of the reduction term is a compound Poisson stochastic process [21] . Using these assumptions we can calculate the expectation with respect to at age : ( 2 ) Since the population of RBCs has mixed ages , we treat the age as a random variable , and treat as the conditional expectation of given that the cell age is : . Taking a second expectation , now with respect to age , with and assumed to be independent of we get: ( 3 ) This analysis allows inference even from crude measurements , as it only requires the mean of the initial population and that of the general population , to calculate the difference . A more detailed analysis and simulations ( using Poisson sample paths , see e . g . , [22] ) are discussed in the materials and methods . In order to apply the model we need to estimate the characteristics of vesiculation . Our knowledge of the vesiculation process is summarized by two parameters: the vesicle size distribution ( vesicle size is quantified here either by volume , , or by radius , since we assume it is a sphere ) , which has been measured by atomic force microscopy ( AFM ) [23]–[25] and recently also by a micro-nuclear magnetic resonance ( NMR ) system [20] , and the vesiculation rate , . Direct estimation of vesiculation rate is not feasible in physiological conditions because currently a specific RBC cannot be monitored repeatedly in vivo . Here we approach the rate estimation by modeling dynamics and analyzing steady states . The vesicle count in a blood sample reflects the balance between production and clearance , formulated as ( 4 ) is the red blood cell concentration ( cells/ ) and is the vesicle concentration ( vesicles/ ) . The model parameters are the average vesiculation rate ( vesicles/cells/day ) and the clearance rate ( 1/day ) . If we assume steady state ( ) , then ( 5 ) is measured routinely in the clinic , and there are values of in the literature . The clearance rate can be estimated by combining vesicle labeling studies [26] and Eq . ( 4 ) ( see Table 2 and materials and methods ) . In Table 2 we collect estimates for obtained from measurements of vesicle and RBC counts and Eq . ( 5 ) . We find that estimated can span two orders of magnitude ( per day ) in healthy humans . In this section , we use the estimated average vesiculation rates ( in Table 2 ) and empirical vesicle sizes ( see materials and methods ) to show that most ( >80% ) of the volume reduction during cell aging cannot be explained by vesiculation . We model the cell volume reduction using Eq . ( 1 ) , replacing with : ( 6 ) is the volume of a cell at age , and is the volume of the ith vesicle ( approximated as a ball with radius ) , leading , as in Eq . ( 3 ) , to the following relationship between and vesicle size: ( 7 ) We estimate from the young reticulocyte population ( red dots in Figure 1 ) , assumed to be at age , ( using the Advia reticulocyte dye intensity , we take cells associated with the highest third ) . We estimate from the total RBC population . The maximal estimate for is ( vesicles/cell/day ) and would require an average vesicle radius >185 nm to account for the total volume lost ( as reported in Table 3 ) . That vesicle radius is twice as large as any reported average vesicle measurement ( see materials and methods ) . The reported average vesicle radii are 53–93 nm , and would require >75 , which is larger than the largest empirical estimate of . In Figure 2 we can see the large difference between the model predictions ( log of Eq . ( 7 ) in blue ) and the range of experimental observations based on blood cell measurements from 21 healthy human adults ( solid lines mark the means , and the accompanying shaded regions mark the ranges ) . The large difference between the range of observed and values and the values predicted by the model indicates that vesiculation can explain only a small fraction ( <20% ) of the observed volume loss ( see Table 3 ) . In the materials and methods we conclude the same using independent data and a geometric argument . We now utilize Eq . ( 1 ) to describe the Hb mass reduction . The Hb mass in a cell of age ( ) is the difference between the initial Hb mass ( ) and the total mass lost in vesicles ( is the Hb mass in a single vesicle ) . The relation between mass lost and vesiculation parameters , as in Eq . ( 3 ) , is given by ( 8 ) where is the Hb concentration , [Hb] in the vesicle . Combining blood measurements of cellular Hb mass ( as in Figure 1 ) with Eq . ( 8 ) we obtain the predicted relation between and vesicle size in Figure 2 ( black line for the mean prediction ) . The vesiculation model can explain only of the total Hb lost , and at most under extreme assumptions ( see Table 3 ) . Thus we conclude now that processes other than vesiculation must be involved in RBC maturation . In our model the Hb mass in the vesicles is proportional to the mean Hb concentration , , in the shedding cells ( 30–35 g/dL ) . If the vesicle [Hb] were at least five times larger , on average , the vesiculation could explain the total Hb loss , but such Hb concentration is not physically possible and would also require an unknown process to pack the Hb into the vesicles . Alternatively , measurements may underestimate the small vesicle fraction , but we show later that an increased small vesicle fraction is inconsistent with dry mass measurements , suggesting that an additional mechanism is indeed involved in the Hb reduction . The evolution of the surface area , and the relation between area lost and vesiculation parameters , as in Eq . ( 3 ) , is given by ( 9 ) is the membrane surface area of a sphere-shaped vesicle . We use the values from Table 4 to estimate the area loss and Eq . ( 9 ) for the predictions in Figure 2 ( dash-red line ) . We find that in the case of surface area loss , vesiculation alone is a sufficient mechanism for the surface area loss , given the constraints on vesicle sizes ( average radii are 53–93 nm ) and ( 4–13 vesicles/day ) . We can decouple the hemoglobin mass dynamics from the volume dynamics by measuring cellular dry mass ( non-water mass ) and cellular dry density ( density of non-water constituents ) . We now integrate models of Hb reduction and surface area reduction ( Eq . ( 8 ) and Eq . ( 9 ) ) resulting in Eq . ( 15 ) to study the changes in the cell's dry mass and density during maturation . A cell's Hb mass is estimated to be about 95–97% of its dry mass [27] . We use newly available measurements of single-cell dry mass and dry density ( referred to as SMR ) [16] . Figure 3 shows cellular dry mass and density ( purple dots ) for an RBC population obtained via the SMR . Dry density shows very little variation with dry mass , despite the large change in dry mass over the cell's life , ( dry mass has a coefficient of variation of 16 . 9% versus a 0 . 3% for the dry density ) . We assume the following: I . RBC dry mass consists of membrane and Hb . II . The Hb and membrane density and are constant during the cell life , and ( see Table 5 ) . III . The membrane mass and volume are linearly related to the membrane surface area , see Eq . ( 14 ) . Rewriting Eq . ( 8 ) , using the number of vesicles : ( 10 ) shows that the lost Hb mass is determined by the number of vesicles , their Hb concentration , and their volume . Details on the simulations are given in the materials and methods . Briefly , we simulate the evolution of the cell's dry mass-density , using Eq . ( 15 ) , Advia data ( see Figure 1 ) , and a random sample of vesicle sizes ( see Materials and Methods ) under two sets of assumptions . Model 1: Fix the vesicle's Hb mass by setting the vesicle [Hb] equal to that of the shedding cell . Model 2: Fix the number of vesicles at the highest level consistent with empirical evidence ( ) . We compare the simulation results to the SMR dry mass-density measurements . Figure 3a shows that when young cells ( red dots ) evolve according to model 1 their dry density increases ( see blue dots ) , inconsistent with the experimental observation ( purple dots ) , and determines an N ( ) that is inconsistent with empirical estimates ( ) . We therefore exclude the possibility that we are underestimating the number of small vesicles . When young cells evolve according to model 2 ( Figure 3b ) their dry density trend fits the experimental data , demonstrating that the rate of Hb loss matches the rate of membrane loss during aging . However , this model requires the vesicle Hb content to be physically unrealistic , based on the values in Table 5 and [26] . The consistency of the model with the data implies that some of the Hb is lost via a mechanism other than vesiculation , and that mechanism is synchronized with vesiculation . One mechanism that can generate such synchronization between the surface loss and hemoglobin loss is a leaky vesiculation , in which some hemoglobin is lost to the surroundings during the process of vesicle release .
We establish that while vesiculation alone can explain the observed membrane lost during RBC maturation , it cannot explain all the Hb or volume lost . There must be an additional process to explain the remaining 60–90 of the volume and Hb reduction occurring during RBC maturation . Dry density is determined by the ratio of Hb mass and cell membrane surface area . Because we see very small changes in cell dry density despite large changes in dry mass , we conclude that changes in Hb mass must be coupled to changes in cell membrane surface area . We suggest that the unknown process ( es ) responsible for up to 90% of the Hb mass and volume reduction are therefore physically linked to the vesiculation-based changes in surface area . It is possible that the unknown processes involve leaky vesiculation , intracellular degradation and excretion and/or interactions with white blood cells or other cells . RBC indices are used in the clinic to monitor and diagnose a wide range of conditions . Our work helps focus future investigation of molecular mechanisms of RBC maturation whose characterization may help in the early detection of clinical conditions where the maturation pattern is altered .
We estimate the vesicle clearance rate from data reported in [26] , where vesicle labeling studies were performed in rats . This data is equivalent to a trajectory of the system in Eq . ( 4 ) with , allowing us to get a physiological estimate of . In those experiments the RBC vesicles were reported to be highly enriched relative to platelet vesicles ( 16 . 7∶1 ) . Thus , the vesicle fraction data appearing in Figure 4 ( blue dots ) is the sum of RBC vesicles ( ) with some small initial fraction of platelet vesicles ( ) . The results of the experiments describe the total vesicle concentration over time: . We assume here a linear ODE model for the clearance of vesicles , without interaction , namely , taking Eq . ( 4 ) for twice , replacing and with and to form the clearance model for platelets . The sum of the solutions of these two models gives the total vesicle concentration over time , The data reported in [26] is the remaining fraction of vesicles in the circulation and not absolute counts . Thus , we derive the clearance parameter from the remaining fraction , where . The reported purity of RBC derived vesicles versus total vesicles is ( ) and similarly for the platelet derived vesicles . Thus , ( 11 ) and we have two parameters to estimate , and . The result of the fitting of those two parameters appears in Figure 4 ( red/dashed curve ) . Ignoring the reported purity , we fit 3 parameters using , ( 12 ) and improve the fit , as expected , ( see black/solid curve in Figure 4 ) . The fitted value of raises the hypothesis that the reported protocol leads to 80% enrichment of RBC vesicles . Modeling note: in a linear model , different initial loads behave in the same way . The current data does not allow us to probe this question . Future vesicle labeling studies should include different doses of vesicles to test whether the response is sensitive to the initial load or not . Here we investigate the second vesicle characteristic , the vesiculation rate , based on the assumption that a sphere-shaped red cell will be cleared as it cannot deform . Given an initial cell with a maximum volume of 110 fl ( Table 6 ) which has surface to volume ratio of 1 . 7 [3] the initial surface area will be 187 . Assuming that cells cannot get smaller than fl and that those cells are spheres leads to an estimate of the total area lost during the cell life in circulation to be 130 . If this area is distributed to a random sample of vesicles ( drawn from an empirical distribution [20] with averaged radius ( ) of 83 . 5 nm ) over the course of 120 days ( Age ) , the cell is estimated to lose 12 vesicles per day: The above values , in particular and the use of radii sampled only from the PDF of NMR ( larger vesicles ) , yield a conservative vesiculation rate estimate . If we choose m2 , , and vesicles radii sampled from the PDFs measured with both AFM and NMR ( see Figure 5 ) , sampled with equal weight , we obtain an upper bound on the vesiculation rate of 27 vesicles per day . The average rates are lower . This analysis thus support the vesiculation rate estimates in the main text using independent data and reasoning . We argue here that it is geometrically impossible to explain the volume lost only by vesiculation , supporting the conclusions obtained in the main text using our stochastic processes theory , with independent data . Previous measurements [3] comparing young RBCs to the total population found that the percent change in area is similar to the percent change in volume . Formulating mathematically , ( 13 ) where and are the total surface and total volume lost . Using empirical values from Table 4 the equality in ( 13 ) holds with less than 1% error . If vesiculation is responsible for all volume and surface area loss , then and , where N is the total number of vesicles ( assumed to be of fixed size ) . The above relationship requires that the surface to volume ratio remain constant for the entire RBC lifetime ( as in Table I of [3] ) . However , this relationship implies that the surface to volume ratio of the cell ( in initial state ) is identical to that of the vesicle ( ) . Given a cell with volume V and surface area S , we define as the solution of . From the isoperimetric inequality in [28] , we know that given a volume , the minimal surface area bounding is a sphere . Hence , in general , for a given volume , ( with equality for a sphere ) . For a vesicle , it is reasonable to assume a spherical shape so since . For example , if , then . Estimates of for RBCs are around , more than an order of magnitude smaller , and thus inconsistent with vesiculation as the only volume loss mechanism . Here we report the findings of others regarding vesicle sizes , which together with the vesiculation rate complete the current characterization of the vesiculation process required in order to apply the theory in the main text . We reproduce the vesicle size distribution from [25] figure 1D ( referred to as AFM ) and the vesicle distribution from blood stored in blood bank conditions [20] ( referred to as NMR ) . Notice that here size is radius ( to be consistent with the mathematical analysis ) rather than diameter as is used more commonly in the literature , and specifically in [20] , [25] . The vesicles in [25] were induced using Ca++/ionophore A23187 ( Sigma , St Louis , MO ) . Smaller vesicles are referred to as ‘nano-vesicles’ and the larger ones are referred to as ‘micro-vesicles’ . We use their counts and merge the two types to form a vesicle size probability density function , pdf , ( Figure 5 , left , blue line ) and cumulative distribution function , cdf , ( Figure 5 , right , blue line ) . The AFM cdf shows that the ( for the NMR , it is ) which is the mean vesicle size we used in some of the simulations ( it is above the 95% CI of the mean even for the NMR data ) . Even if the vesicle size is dependent on the cell age , we are using a realistic sample size or a larger-than-average vesicle size , and thus we expect that a simple age dependence would not allow explanation of the results based on vesiculation alone . Note that a small number of very large vesicles during the RBC life with large volume ( e . g . , total of 15 fl ) might explain the volume loss but would be inconsistent with measured steady state vesicle concentrations . The cell dry density calculations require the use of the membrane volume and mass , which are assumed to follow: ( 14 ) is the membrane area . This relation is assumed to hold independently of cell age , using the values of and from Table 5 . The following conditions are used in both model 1 and 2 . We estimate the Hb mass lost from measurements as , that is , the difference between the mean Hb mass of the young RBCs and that of the total RBC population ( Advia data as in Figure 1 ) . Then we generate a Poisson sample path [22] and use a random sample of vesicle radii from the distribution in Figure 5 . With the assumption that the vesicles are spherical , we can infer the vesicle surface area and volume . The simulations require initial values of Hb mass and surface area . Hb mass is obtained per sample from the Advia clinical analyzer , and the initial surface area is volume-matched to the data in [29] . By ‘volume-matched’ we mean that for each young cell with known Hb mass and volume , we find the cell with the closest volume in the volume and surface area data from [29] , and use that surface area to form the Hb mass-surface area pair . We use Eq . ( 8 ) and Eq . ( 9 ) for the evolution of the Hb mass and surface area , their sum for the dry mass , and Eq . ( 15 ) for the dry density . The trend in the dry mass and dry density is predicted by combining the data from the Advia clinical blood analyzer and volume-matched surface area from [29] . The SMR data is a third data source , and each data set may have a fixed offset depending on the particular device calibration ( see Table 7 for summary on data sets used ) . The data in Figure 3 is obtained by matching the means of the dry mass and dry density of the initial data used in the simulation to that of the top 20% of the measured SMR dry mass ( offset of 14 . 6% for the mass and 0 . 9% for the density ) . The offset does not affect the relation between the dry mass and dry density , as we are just adding a constant . Note that the assumption of a ball-shaped vesicle is conservative , as it leads to a lower bound estimate on the actual increase in density . Any other shape will require more vesicle surface area to contain the same amount of volume , requiring the shedding cell's density to increase more than in the case of a ball-shaped vesicle . The following calculations show under assumptions in Eq . ( 8 ) and Eq . ( 9 ) , and in particular , that the shedding of each vesicle by a cell requires that the cell's dry density must increase . Using Eq . ( 10 ) we calculate the cell dry density: ( 15 ) An increasing trend in density along a trajectory requires that if then ( with strict inequality for at least some of the ages ) . We now look at the effect of shedding a single vesicle . This analysis is sufficient due to the independent increment property of the proposed model . If is the time of shedding of the vesicle , we compare to for ( i . e . , ) . Rearranging Eq . ( 15 ) we get if ( 16 ) where is the [Hb] in the vesicle . Eq . ( 16 ) gives an upper bound on the vesicle Hb concentration . The bound is calculated from the initial cellular Hb mass and membrane surface area , as well as the vesicle's radius . Taking ( pg ) , ( µm2 ) and ( nm ) , the bound on the concentration is 600 For reference , [Hb] in human RBCs is never outside the range 20–50 g/dl , and it is physically impossible to achieve such high [Hb] ( based on the values in Table 5 ) . The immediate conclusion is that we always have increasing dry density along a cell trajectory . This result could be demonstrated experimentally if it were feasible to monitor a single RBC in circulation for at least several hours . We thus find that cell dry density must increase monotonically with each vesicle shed . This analysis of dry density along a path shows that the difference in the two models ( as seen in Figure 3 a-b ) is quantitative and not qualitative , since both models have an underlying increasing trend . The second model ( see Figure 3b ) has a milder increase per vesicle and fewer vesicles along the trajectory , which make the trend less apparent against the background population variation . In Figure 6 we show the raw SMR data: dry mass versus dry density . The young RBC ( reticulocyte ) data was not collected for all samples and is required for the simulations . In Figure 7 we show those samples with simultaneous measurements of Hb mass and dry density . Here the reported vesiculation rate is calculated per sample according to Eq . ( 7 ) , adjusted to the Hb mass model ( Eq . ( 8 ) ) for model 1 , assuming for model 2 . The area model ( Eq . ( 9 ) ) can generate an estimate of , but because we have no individual measurements of cell area , we use 3 single-cell measurements of volume and surface area and match the initial volumes as described above . Theses assumptions probably reduce the variability both within and between samples . The estimates we obtain are in the range . Note that surface area parameters used in Eq . ( 9 ) and in Figure 2 are based on Table 4 which is independent of the surface area used in the simulations here , in the context of predicting dry mass-density profiles , that are based on data from [29] . Among the samples in Figure 7 , S2 ( used in Figure 3 ) , S4 , and S5 are more typical of healthy human adults based on the average Hb mass lost . S1 and S3 show a smaller than typical Hb mass loss ( Advia measurement ) . For S3 we see that while the estimated vesiculation rate is only 2 . 5 times more when estimated using the Hb mass model versus the surface area model , the dry density spans the same range and thus the difference between models 1 and 2 persists . In sample S1 we see that the vesiculation rate is similar when estimated by either the Hb mass or surface area model . In this case the Hb mass lost is 0 . 33 pg , which is 13 . 75% of the mean loss in our 21 healthy human adults . Further investigation of this anomaly is beyond the scope of the current study . It is possible that new biophysical measurements like dry-mass and dry-density have diagnostic potential in the context of some forms of anemia which are associated with a reticulocyte population located much closer to the general population ( in contrast to Figure 1 ) . It is likely that these pathologic conditions show different RBC maturation patterns . In this work several data sources have been integrated via our mathematical modeling . We collected the parameter names , methods , and references in Table 7 . Here we briefly describe those methods . The RBC volume and hemoglobin content are measured via the Siemens Advia 2120 automated clinical hematology analyzer [30] . This instrument is essentially a flow cytometer that uses an isovolumetric-sphering reagent prior to the light scattering to render the measurement invariant to cell presentation . Using a pair of small and large angle light scattering intensities and Mie scattering theory , the cell volume and hemoglobin concentration for each cell are calculated . The surface area is measured using a microfluidic device , either by fixing the cells in a constriction of known geometry [3] or by controlling the flow and utilizing symmetry-based calculations form a 2-dimensional image [29] . The cellular dry mass and dry density are measured using the SMR , a microfabricated mass sensor which implements Archimedes principle in a microfludic device for individual cells . Measuring the buoyant mass , or mass in fluid , of a cell sequentially in two fluids of known density allows inference of the cell's mass , volume , and density . When the two fluids are H2O-based and D2O-based , the cell exchanges its water content with D2O and thus the measurements yield only the dry mass and dry density of the RBC , since only the dry content , and not the aqueous content , contributes to the buoyant mass in either fluids [16] . Vesicle radii were measured either from atomic force microscopy ( AFM ) images , assuming spherical symmetry , [25] or recently by μ NMR [20] . The μ NMR method involves labeling micro-vesicles with target-specific ( CD235a antibody ) magnetic nanoparticles and quantifying their concentration using a miniaturized nuclear magnetic resonance system . For both RBC lifetime estimation [26] and vesiculation clearance rate [11] , [26] , [31] , [32] , the experimental design includes the isolation of RBCs/vesicles from a blood sample , labeling of the isolates , reinfusion of the labeled sample , and measurement of the fraction of labeled cells/vesicles repeatedly over time . The study protocol was approved by the local institutional review board ( IRB ) at Massachusetts General Hospital , in accordance with the principles expressed in the Declaration of Helsinki . We used blood samples that had been collected solely for non-research purposes ( such as medical treatment or diagnosis ) . | Red blood cell concentration ( RBC ) , mean volume ( MCV ) , and hemoglobin content ( MCH ) are routinely measured in the complete blood count , a fundamental clinical test essential for the screening , diagnosis , and management of most diseases . Variation in MCV and MCH is associated with many important clinical conditions , but we do not understand the mechanisms controlling these red blood cell physical characteristics . Vesicle shedding is thought to be most important , but we show here that a dominant role for vesicle shedding violates empirical geometric and biophysical constraints . An additional unknown process must be primarily responsible . We show that this important unknown process must be coupled to changes in RBC surface area , and we quantify the magnitude of its effects . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
]
| [
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"population",
"dynamics",
"blood",
"counts",
"integrative",
"physiology",
"blood",
"volume",
"mathematics",
"population",
"modeling",
"population",
"biology",
"stochastic",
"processes",
"theoretical",
"biology",
"biophysics",
"hematology",
"probability",
"theory",
"systems",
"biology",
"hematocrit",
"blood",
"anatomy",
"physiology",
"biology",
"and",
"life",
"sciences",
"hematopoietic",
"system",
"physical",
"sciences",
"computational",
"biology",
"markov",
"processes"
]
| 2014 | In Vivo Volume and Hemoglobin Dynamics of Human Red Blood Cells |
The ( asymptotic ) degree distributions of the best-known “scale-free” network models are all similar and are independent of the seed graph used; hence , it has been tempting to assume that networks generated by these models are generally similar . In this paper , we observe that several key topological features of such networks depend heavily on the specific model and the seed graph used . Furthermore , we show that starting with the “right” seed graph ( typically a dense subgraph of the protein–protein interaction network analyzed ) , the duplication model captures many topological features of publicly available protein–protein interaction networks very well .
In the past few years , protein–protein interaction ( PPI ) networks of several organisms have been derived and made publicly available . Some of these networks have interesting topological properties ( e . g . , the degree distribution of the yeast PPI network is heavy-tailed; that is , there are a few nodes with many connections ) . It has been argued that the degree distribution of these networks are in the form of a power law [1 , 2] ( some recent works challenge this by attributing the power law–like behavior to sampling issues , experimental errors , or statistical mistakes [3–7] ) . Since well-known random graph models also have power-law degree distributions [8–10] , it has been tempting to investigate whether these models agree with other topological features of the PPI networks . There are two well-known models that provide power-law degree distributions [11–13] . The preferential attachment model [9 , 14] was introduced to emulate the growth of naturally occurring networks such as the web graph; unfortunately , it is not biologically well-motivated for modeling PPI networks . The duplication model , on the other hand [15–17] , is inspired by Ohno's hypothesis on genome growth [18] by duplication . Both models are iterative in the sense that they start with a seed graph and grow the network in a sequence of steps . The degree distribution is commonly used to test whether two given networks are similar or not . However , networks with identical degree distributions can have very different topologies ( e . g . , consider an infinite 2-D grid versus a collection of cliques of five nodes; in both cases , all nodes have a degree of four ) . Furthermore , it was observed in [3] that given two networks with substantially different initial degree distributions , a partial ( random ) sample from those networks might give subnetworks with very similar degree distributions . Thus , the degree distribution cannot be used as a sole measure of topological similarity . In the recent literature , two additional measures have been used to compare PPI networks with random network models . The first such measure is based on the k-hop reachability . The 1-hop reachability of a node is simply its degree ( i . e . , the number of its neighbors ) . The k-hop reachability of a node is the number of distinct nodes it can reach via a path of ≤k edges . The k-hop reachability of all nodes whose degree is λ is the average k-hop reachability of these nodes . Thus , the k-hop reachability ( for k = 2 , 3 , . . . ) of nodes as a function of their degree can be used to compare network topologies . An earlier comparison of the k-hop reachability of the yeast network with networks generated by certain duplication models concluded that the two network topologies are quite different [19] . The second similarity measure is based on the graphlet distribution . Graphlets are small subgraphs such as triangles , stars , or cliques . In [4] it was noted that certain “scale-free” networks are quite different from the yeast PPI network with respect to the graphlet distribution . This observation , in combination with that on the k-hop degree distribution , seems to suggest that the known PPI networks may not be scale-free , and that existing scale-free network models may not capture the topological properties of the PPI networks . There are other topological measures that have been commonly used in comparing social networks , etc . , but not PPI networks . Two well-known examples are the betweenness distribution and the closeness distribution [20] . Betweenness of a node v is the number of shortest paths between any pair of nodes u and w that pass through v , normalized by the total number of such paths . Closeness of v is the inverse of the total distance of v to all other nodes u . Thus , one can use betweenness and the closeness distributions , which respectively depict the number of nodes within a certain range of betweenness and closeness values that can be used to compare network topologies .
As mentioned above , scale-free network generation models such as the preferential attachment model and the duplication model can have very similar degree distributions under appropriate choice of parameters . ( See Materials and Methods for exact definitions for the two network generation models . ) Moreover , the degree distribution of these models converge to a power-law degree distribution whose shape is determined solely by the edge deletion and edge insertion probabilities , and not by the initial “seed” graph [11] . Hence , it has been tempting to assume that networks generated by these models are similar in general; moreover , the effect of the seed graph in shaping the topologies of these networks has largely been ignored in recent literature . We start with the observation that two networks with very similar degree distributions may have very different topologies . For example , a network generated by the preferential attachment and another generated by the duplication model may have very different k-hop reachability , graphlet , betweenness , and closeness distributions while having almost identical degree distributions . Figure 1 depicts the degree distribution , k-hop reachability , and graphlet frequency of the duplication model and the preferential attachment model with 4 , 902 nodes ( as per the yeast PPI network [21] ) . Both models start with identical seed graphs; we set r = 0 . 12 , p = 0 . 365 ( the two key parameters of the duplication model ) , and c = 7 ( the single key parameter of the preferential attachment model ) so that the average degree of nodes in both models is seven ( again as per the yeast PPI network [21] ) . Figure 1 compares the k-hop reachability achieved by the two models for k > 1 . As can be seen , the k-hop reachability is quite different , especially for k = 3 , 4 . Figure 1 also shows how the graphlet distributions differ , especially for dense graphlets ( e . g . , graphlets 17–29 and 85–145 ) . In terms of betweenness and closeness , there are some differences as well . We now show that the seed graph has a role in characterizing the topology of the duplication model . Figure 2 depicts how various topological features of the duplication model with fixed parameters ( p = 0 . 365 and r = 0 . 12 ) vary as the seed graph changes . The first seed graph ( red ) is obtained by highly connecting two cliques of ten and seven nodes , respectively , by several random edges . To reduce the average degree , some additional nodes were generated and randomly connected to one of the cliques . The second seed graph ( blue ) is obtained by enriching a ring of 17 nodes by random connections so as to make the average degree match that of the first seed graph . The third seed graph ( green ) is formed by sparsely connecting two cliques of ten and seven nodes , respectively , with some added nodes randomly connected to one of the cliques . All three networks were grown until all had 4 , 902 nodes as per the yeast PPI network [21] . ( We depict the “average behavior” of five independent runs of each of the models . ) It can be observed that although all of them have very similar degree distributions , their graphlet distributions may be quite different , especially for dense graphlets . Figure 2 also compares the k-hop reachability , closeness , and betweenness distributions . As can be seen , the k-hop reachability and the closeness distribution can vary considerably . Note that both the graphlet and the closeness distributions are in logarithmic scale , and seemingly small variations in the figure may imply several factors of magnitude of a difference between the two distributions . The key question we aim to address in this paper is the following . If the seed selection has such an impact in shaping the topology of the network generated by the duplication model , is it possible to select the “right” seed graph so that all interesting topological features of the PPI networks in question can be captured ? Also , is there a systematic way to determine a subgraph of a PPI network that can provide a good seed graph ? We answer the above questions positively by demonstrating that the duplication model applied on the right seed graph can result in a network that accurately captures all key features of the PPI networks we considered . The PPI networks we consider in this study include ( the largest connected component of ) the complete Database of Interacting Proteins ( DIP ) yeast PPI network [21] with 4 , 902 proteins and 17 , 200 edges ( as of July 2006 ) as well as the smaller but more accurate core yeast network from the DIP [22] . We also tested the lesser-developed DIP worm network [21] . ( See Materials and Methods for a detailed description of these networks . ) As will be demonstrated , we were able to closely approximate all the interesting topological features of these networks via the duplication model using specific seed graphs that largely exist as a subgraph in the corresponding PPI network . A crucial observation toward obtaining the right seed graph is that the duplication model is unlikely to generate “large” cliques ( a set of nodes which are fully connected ) . Notice that the only way to produce a clique of size h through the duplication model is starting with a clique of size h − 1 , duplicating one of its nodes , and making sure that none of the new node's edges that are connected to the clique are deleted . The probability of this happening is negligible for large values of h . The size of the maximum clique in the yeast PPI network is ten nodes . In our experiments with the duplication model , even if we started with a seed graph that included a clique of nine nodes ( but not ten ) , the chances that we ended up with a clique of ten nodes ( in <5 , 000 steps ) turned out to be negligible . Thus , the seed graph has to include a clique of ten nodes . We enriched the seed graph by adding to the clique of ten nodes another ( independent ) clique of seven nodes that is present in the yeast PPI network . We also included the edges between the two cliques and some additional nodes so that the normalized degree distribution of the yeast PPI network would be similar to that of the seed graph . The total number of nodes in the resulting seed graph was 50 . As mentioned before , there are two key parameters associated with the duplication model: p , the edge maintenance probability; and r , the edge insertion probability . These two parameters alone determine the ( asymptotic ) degree distribution and the average degree of the generated network . We chose p = 0 . 365 and r = 0 . 12 so that the degree distribution of the duplication model matches that of the yeast PPI network ( see Methods and Materials for the exact mathematical expressions for p and r ) . Also , for the preferential attachment model , we choose the value c = 7 so that the average degree of the graph created using preferential attachment would be equal to that of the yeast PPI network . We used the duplication model and preferential attachment model described above to generate a network with 4 , 902 nodes . The resulting networks are compared with the yeast PPI network in terms of the k-hop reachability , the graphlet , betweenness , and closeness distributions in Figure 3 . Under all these measures , the yeast PPI network is very similar to the network produced by the duplication model ( and not similar to the network produced by the preferential attachment model ) . In fact , the duplication model approximates both the k-hop degree distribution and the graphlet distribution of the yeast network much better than the random graph models described earlier in the literature ( [4] and [19] ) —which were specifically devised to capture the respective features of the yeast PPI network . Another evidence of the power of the duplication model in capturing the topological features of available PPI networks is through comparing the duplication model with the main component of the core subset of the yeast network . The core subset contains the pairs of interacting proteins identified in the yeast that were validated according to the criteria described in [22] . It involves 2 , 345 nodes and 5 , 609 edges . The values of r and p were set to r = 0 . 12 , p = 0 . 322 as prescribed by the average degree formula a = 2r / ( 1− PS − 2p ) and the fact that PS is a function of r and p ( see the next section for explanation ) . The seed graph we used was very similar to that used for the complete yeast network . Also , for the preferential attachment model , we set a value c = 4 . 8 so that the network generated using the model has the same average degree as the CORE yeast PPI network . The results are shown in Figure 4 . Although the yeast PPI network is the most reliable PPI network available , it is still far from completion . Following up on [3] , we also considered the effect of sampling errors on the duplication model with respect to all the topological features used . In order to emulate the effect of sampling and thus the ( potential ) presence of false negatives in the yeast PPI network , we used the duplication model to generate larger networks than the available ones and applied the sampling strategy proposed in [3] to “shrink” them to the size of the available networks . The sampling strategy of [3] involves two parameters: the bait sampling probability ( the probability that a node is kept in the network during sampling ) and the edge sampling probability ( the probability that an edge of a bait is kept in the network ) . We demonstrate the effect of sampling as per [3] on the emulation of both the full yeast and the CORE yeast PPI networks below . We used a bait sampling probability and an edge sampling probability of 0 . 7 each ( resulting in 70% “bait coverage” and again 70% “edge coverage” ) for our emulation of the full yeast PPI network . A comparison of the features of the resulting network with that of the full yeast PPI network is given in Figure 5 . We then used a bait sampling probability and edge sampling probability of 0 . 5 each for emulating the core yeast PPI network ( resulting in 50% “bait coverage” and 50% “edge coverage” ) . A comparison of the core yeast PPI network against the resulting network is given in Figure 6 . As can be seen , the topological features of both the full yeast PPI network and the core yeast PPI can still be closely captured by the networks obtained via the duplication model , which have been subject to sampling errors . The seed graphs used in both tests involving sampling are identical to those used in the tests that do not involve sampling . Uniform sampling reduces the size of the maximum clique in the resulting networks significantly , as can be seen at the tail end of the graphlet distributions . In reality , the sampling errors are not uniform . Very dense subnetworks such as cliques are better covered by both the full yeast network and the core yeast network of the DIP . It is interesting to note that although the core yeast network has only 5 , 609 edges in comparison to the full yeast network's 17 , 200 edges , the maximum clique size in the former is nine nodes , whereas it is ten nodes in the latter .
Perhaps the best-known PPI network database is DIP [21] , which includes the Saccharomyces cerevisiae ( yeast ) PPI network ( the best-developed PPI network available , with 4 , 902 proteins and 17 , 200 interactions ) . DIP also includes a more accurate but much smaller core yeast network ( 2 , 345 proteins and 5 , 609 interactions ) [22] . Our results are mainly on these two networks . Although there are other PPI networks available through DIP [23] ( e . g . , those of the fruit fly , human , and mouse ) as well as through BIND [24] , IntAct [25] , and MINT [26] databases , they are not sufficiently well-developed to perform a conclusive analysis . For comparison purposes , we also provide results on the DIP Caenorhabditis elegans ( worm ) PPI network ( which includes 2 , 387 proteins and 3 , 825 interactions ) as Text S1 . The two network models we study here , namely the preferential attachment model and the duplication model , both start with a small seed graph and create an additional node in each iteration as described below . For notational convenience , let G ( t ) = ( V ( t ) , E ( t ) ) be the network at the end of time step t , where V ( t ) is the set of nodes and E ( t ) is the set of edges/connections . Let vt be the node generated in time step t . Given a node vt , we denote its degree at the end of time step t by dt ( vt ) . The preferential attachment model ( as analyzed in [9 , 11 , 14 , 27] ) generates a network as follows . In iteration t a new node vt is generated and is connected to every other node vt in the network independently with probability . Here , c is the average degree of a node in G . The duplication model ( as analyzed in [15–17] ) , in contrast , generates a network as follows . In iteration t , an existing node vt of G ( t − 1 ) is picked uniformly at random and “duplicated” ( i . e . , an exact copy of vt as vt is generated ) . The edge set of vt is then updated: first , each existing edge of vt is deleted independently with probability ( 1 − p ) ; then each node vt not connected to vt is connected to vt independently with probability r / |V ( t ) | . Here , p and r are user-defined parameters . Note that it is possible to maintain a constant average degree ( a ) throughout the generation of the network by setting r = ( 1/2 − p ) · a . As mentioned earlier , the degree distributions of both the preferential attachment model and the duplication model asymptotically approach a power law [9 , 11 , 12 , 14] . More specifically , the frequency of nodes with degree d is proportional to d−b , where b is a constant typically between 2 and 3 . The value of b is solely determined ( asymptotically ) by the values of p and r in the duplication model or the value of c in the preferential attachment model . Both the preferential attachment and the duplication model produce many singletons [13] ( i . e . , nodes that are not connected to any other node ) . ( For example , in the duplication model where r = 0 , p = 1/2 , the proportion of singletons asymptotically approaches 1 . ) In contrast , the number of singletons in known PPI networks is very small ( this is not surprising , as genes with “no functionality” are not maintained by evolution ) . To avoid the generation of singletons , it is possible to use a slightly modified duplication model that deletes each singleton node as soon as it is generated . This modified duplication model has also been shown to achieve a power-law degree distribution [13] . However , it is not known which values of p and r ensure that the expected average degree can be set to a desired value and is kept fixed through all iterations . In this paper , we derive conditions on p and r that are necessary for having a constant expected degree . We later use these conditions so that the modified duplication model can approximate the degree distribution of the yeast PPI network as tightly as possible . Perhaps the ultimate way to test whether two networks are topologically similar or not is through the use of graph isomorphism as described below . Unfortunately , graph isomorphism and approximate graph isomorphism are computationally hard problems . Thus , it is very common to use some of the topological features of networks as a basis of checking their similarity . In this paper , we focus on five such features: the degree distribution , the k-hop reachability , the graphlet frequency , the betweenness distribution , and the closeness distribution . Graph isomorphism . Two networks , G and G′ , are called isomorphic if there exists a bijective mapping F from each node of G to a distinct node in G′ , such that two nodes v and w are connected in G if and only if F ( v ) and F ( w ) are connected . G and G′ are called approximately isomorphic if by removing a “small” number of nodes and edges from G and G′ , they could be made isomorphic . Ideally , a random graph model that aims to emulate the growth of a PPI network should produce a network that is approximately isomorphic to the PPI network under investigation . Unfortunately , the problem of approximate isomorphism is NP-complete ( through a trivial reduction from subgraph isomorphism—a known NP-complete problem ) ; thus , this measure cannot be used to practically test similarity of two networks . k-hop reachability . Let V ( i ) denote the set of nodes in G whose degree is i . Given a node v , denote by d ( v , k ) its k-hop degree ( i . e . , the number of distinct nodes it can reach in ≤k-hops ) . Now we define f ( i , k ) , the k-hop reachability of V ( i ) , as Note that f ( i , k ) is the “average” number of distinct nodes a node with degree i can reach in k-hops ( e . g . , f ( i , 1 ) = i by definition ) . Graphlet frequency . The graphlet frequency was introduced in [4] to compare the topological structure of networks . A graphlet is a small connected induced subgraph of a large graph ( e . g . , a triangle or a clique ) . The graphlet count of a given graphlet g with r nodes in a given graph G = ( V , E ) is defined as the number of distinct subsets of V ( with r nodes ) whose induced subgraphs in G are isomorphic to g . In this paper , we consider all 141 possible graphlets/subgraph topologies with three , four , five , and six nodes . In addition , we consider cliques of sizes seven , eight , nine , and ten . We enumerate these graphlets as shown in the final figure in Text S2 . Betweenness distribution . The betweenness of a fixed node of a network measures the extent to which a particular point lies “between” point pairs in the network G = ( V , E ) . The formal definition of betweenness is as follows . Let sx , y be the number of the shortest path from x ∈ V to y ∈ V for all pairs x , y ∈ V . ( Note that sx , y = sy , x in undirected graphs . ) Let sx , y ( v ) be the number of shortest paths from x ∈ V to y ∈ V which go through node v . The betweenness Bet ( v ) of node v is now defined as Closeness . For all x , y ∈ V , we define dx , y as the length of the shortest path between x and y . The closeness of a node v ∈ V is defined as Thus , closeness of a node v is simply the inverse of the average distance of v to all other nodes in G . The network comparison methods in use: The yeast PPI network versus the Erdos–Renyi random graph model . The network features described above can be used to test whether a given random graph model can emulate an available PPI network . Here , we consider the standard Erdos–Renyi random graph model [28] in comparison to the yeast PPI network . As shown in Figure 7 , each of the features we consider point to significant differences between yeast PPI ( red ) and ( five independent runs of ) the Erdos–Renyi ( green ) model . In this section , we show how to determine the deletion probability 1 − p with respect to the insertion probability r so that the expected average degree of the network can be set to any given value . For this , we make the assumption that the degree frequency distribution and the average degree of nodes are fixed asymptotically once the values of p and r are determined . Let G ( t ) = ( V ( t ) , E ( t ) ) be the network generated by the modified duplication model and let n ( t ) = |V ( t ) | and e ( t ) = |E ( t ) | . Also , let nk ( t ) be the number of nodes in time step t with degree k and a ( t ) be the average degree of nodes in G ( t ) . Finally , let Pk ( t ) = nk ( t ) / n ( t ) , the frequency of nodes with degree k at time step t . We assume that Pk ( t ) is asymptotically stable ( i . e . , Pk ( t ) = Pk ( t + 1 ) for all 1 ≤ k ≤ t for sufficiently large values of t . In other words , we assume that Pk ( t ) = dk for some fixed dk . By definition: Now we can calculate the average degree a ( t + 1 ) under the condition that degree frequency distribution is stable and a ( t ) = a , a constant . Let Prs ( t ) be the probability that vt+1 ends up as a singleton . Since this probability does not depend on t asymptotically , we can set Prs ( t ) = Prs . Now we can calculate the expected number of nodes and the expected number of edges in step t + 1 . Comparing the above equation with the first equation for Exp[e ( t + 1 ) ] , we get Solving Equation 10 results in a = 2r / ( 1 − Prs − 2p ) , where Prs is a function of p , r , and dk only . The discussion above demonstrates that the two key parameters p and r of the ( modified ) duplication model are determined by the degree distribution and the average degree of the PPI network we would like to emulate . | The interactions among proteins in an organism can be represented as a protein–protein interaction ( PPI ) network , where each protein is represented with a node , and each interaction is represented with an edge between two nodes . As PPI networks of several model organisms become available , their topological features attract considerable attention . It is believed that the available PPI networks are ( 1 ) “small-world” networks , and ( 2 ) their degree distribution is in the form of a “power law . ” In other words , ( 1 ) it is possible to reach from a protein to any other protein in only a small ( approximately six ) number of hops , and ( 2 ) although most proteins have only a few interactions ( one or two ) , there are a few proteins with many more interactions ( 200 or more ) and that act as “hubs . ” It has thus been tempting to develop simple mathematical network generators with topological features similar to those of the available PPI networks . One such model , the “duplication model , ” is based on Ohno's model of genome growth . It starts with a small “seed network” and grows by “duplicating” one of the existing nodes at a time , with an identical set of interactions; a randomly selected subset of these interactions is then deleted , and a few new interactions are added at random . It has been mathematically proven that the duplication model provides a small-world network and also has a power-law degree distribution . What we show in this paper is that by choosing the “right” seed network , many other topological features of the available PPI networks can be captured by the duplication model . The right seed network in this case turns out to include two sizable “cliques” ( subnetworks where all node pairs are connected ) with many interactions in between . In this paper , we also consider the preferential attachment model , which again grows by adding to a seed network one node at a time and connecting the new node to every other node with probability proportional to the existing degree of the second node . Because the preferential attachment model also provides a small-world network and has a power-law degree distribution , it has been considered equivalent to the duplication model . We show that the two models are vastly different in terms of other topological features we consider , and the preferential attachment model cannot capture some key features of the available PPI networks . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
]
| [
"saccharomyces",
"computational",
"biology"
]
| 2007 | Not All Scale-Free Networks Are Born Equal: The Role of the Seed Graph in PPI Network Evolution |
Immunity-related GTPases ( IRG ) play an important role in defense against intracellular pathogens . One member of this gene family in humans , IRGM , has been recently implicated as a risk factor for Crohn's disease . We analyzed the detailed structure of this gene family among primates and showed that most of the IRG gene cluster was deleted early in primate evolution , after the divergence of the anthropoids from prosimians ( about 50 million years ago ) . Comparative sequence analysis of New World and Old World monkey species shows that the single-copy IRGM gene became pseudogenized as a result of an Alu retrotransposition event in the anthropoid common ancestor that disrupted the open reading frame ( ORF ) . We find that the ORF was reestablished as a part of a polymorphic stop codon in the common ancestor of humans and great apes . Expression analysis suggests that this change occurred in conjunction with the insertion of an endogenous retrovirus , which altered the transcription initiation , splicing , and expression profile of IRGM . These data argue that the gene became pseudogenized and was then resurrected through a series of complex structural events and suggest remarkable functional plasticity where alleles experience diverse evolutionary pressures over time . Such dynamism in structure and evolution may be critical for a gene family locked in an arms race with an ever-changing repertoire of intracellular parasites .
Immunity Related GTPases ( IRG ) , a family of genes induced by interferons , are one of the strongest resistance systems to intracellular pathogens [1]–[4] . The IRGM gene has been shown to have a role in the autophagy-targeted destruction of Mycobacterium bovis BCG [5] . Recently , whole genome association studies have shown that specific IRGM haplotypes associate with increased risk for Crohn's disease [6] , [7] . The IRG gene family exists as multiple copies ( 3–21 ) in most mammalian species but has been reduced to two copies , IRGC and a truncated gene IRGM , in humans [8] . Analysis of mammalian genomes ( dog , rat and mouse ) has shown that all IRG genes except IRGC are organized in tandem gene clusters mapping to mouse chromosomes 11 and 18 ( both syntenic to human chromosome 5 ) [8] . A comparison of the mouse and human genomes identified 21 genes in mouse but only a single syntenic truncated IRGM copy and IRGC in human [8] . We investigated the copy number and sequence organization of the IRG gene family in multiple nonhuman primate species in order to reconstruct the evolutionary history of this locus .
Sequence analysis of two different prosimian species ( Microcebus murinus and Lemur catta ) confirmed the mammalian archetypical organization with three IRGM paralogs in each species ( Figure 1 ) . FISH analysis showed that genes in these species are organized as part of a tandem gene family similar to the organization observed within the mouse genome ( Figure 2 ) . In contrast , FISH and sequence analysis of various monkey and great ape species ( see Text S1 ) confirmed a single copy in each of these species . Based on the estimated divergence of strepsirrhine and platyrrhine primate lineages , we conclude the IRGM gene cluster contracted to a single truncated copy 40–50 million years ago within the anthropoid lineage of evolution . We next compared the structure of the IRGM gene in various primate species . One of the three mouse lemur IRGM genes ( IRGM9 ) preserves a complete ORF based on the mouse model and shows the greatest homology to mouse Irgm1 . The ORF encodes a putative 47 kD protein including a classical N-terminal region as well as classical motifs at the end of the carboxyl-terminus associated with most functional murine IRGM loci [8] , [9] ( see Text S1 ) . The second mouse lemur gene , IRGM8 , is likely a pseudogene because of a mutation generating a stop codon within the G domain and a frameshift mutation at the C terminus . The third mouse lemur gene , IRGM7 , is atypical because it has substitutions in the G domain that disrupt the G1 motif that interacts with the nucleotide phosphates and is highly conserved in P-loop GTPases [10] ( Figure S1 and Text S1 ) . In contrast to mouse and prosimian species , all anthropoid primate lineages show the presence of an AluSc repeat immediately after the splicing acceptor that disrupts the ORF of the sole remaining IRGM gene ( Figure 1 and S2 ) . Sequencing of the IRGM locus in four New World monkey species revealed the presence of the same two stop codons disrupting the ORF of IRGM in all species . We similarly identified a common frameshift mutation resulting in premature stop codons within the IRGM locus in eleven diverse Old World monkey species suggesting that IRGM had become pseudogenized before the radiation of these species . Sequencing of the gene in multiple individuals in the same species ( five unrelated Rhesus macaque and baboon ) suggested that the frameshift mutations were fixed ( Figure S3 and Text S1 ) . In total , these data argue that the IRGM locus has been nonfunctional since the divergence of the New World and Old World monkey lineages ( 35–40 million years ago ) likely as a result of an Alu repeat integration event that disrupted the ORF of the gene in the anthropoid ancestor ( Figure 1 ) . In contrast to New World and Old World monkeys , sequencing of the IRGM locus in humans and African great ape species reveals a restored , albeit truncated , ORF of ∼20 kD in length . This is consistent with an antiserum raised against peptides from the human IRGM protein that detected a specific signal at ∼20 kD by Western blot [11] . In contrast to humans and the African great apes , analysis of the orangutan genome assembly predicted a nonfunctional protein ( C to T transition at nucleotide position 150 with respect to the start codon resulting in a premature shared stop codon in the ORF ( Figure 1 and Text S1 ) . This is the same substitution identified among all Old World monkey genomes suggesting that ancestral ape species carried a pseudogene . We resequenced the IRGM gene in twelve different orangutans and five different gibbon species . Six of the twelve individuals from orangutan and one of the five species from gibbon are heterozygous for the C to T substitution . In addition , we noted that all ape IRGM copies also shared a new translation initiation codon with a preferred Kozak sequence immediately after the Alu integration . These data indicate that the gene can exist as either a pseudogene or as a complete 20 kD ORF among these Asian ape lineages as a result of either balancing selection or recurrent mutational events . It will be necessary to examine a larger number of individuals within each species to establish the evolutionary history of this locus among the Asian apes . We noticed an important structural difference in the gene organization for species that regained putative IRGM function when compared to those primates with a pseudogenized version . In the common ancestor of humans and great apes , an ERV9 retroviral element integrated within the 5′ end of the IRGM gene ( Figure 1 ) . We reasoned that this structural difference may have conferred expression differences and analyzed the RT-PCR expression profile of IRGM in human , macaque and marmoset . Full-length cDNA sequencing and 5′ RACE revealed that the human transcription start signal mapped specifically within the ERV9 repeat element ( Figure 1 and Figure S4 ) resulting in the addition of a novel 5′UTR exon and an alternative splice form . Although there are five distinct , alternative splice forms of human IRGM , all human copies share this first intron . In humans , we observe constitutive levels of expression of IRGM in all tissues examined , with the highest expression of IRGM in the testis ( Figure 3A ) [8] . Although IRGM does not encode a functional protein in marmoset and macaque , we find evidence of low levels of expression , albeit in a more restricted manner ( Figure 3B ) . Macaque and marmoset , for example , show no expression in the kidney with marmoset IRGM expression restricted to testis and lung . Furthermore , we find no evidence in macaque of splicing of the first intron based on the human IRGM gene model ( Figure 3C ) but rather evidence that the first intron remains as a continuous unspliced transcript . We also failed to confirm 3′ downstream splicing events of macaque IRGM suggesting that even if stop codons were reverted , a full-length cDNA ( comparable to human ) could no longer be produced . These data strongly suggest that ERV9 integration significantly reshaped the expression and splicing pattern of IRGM in the common ancestor of humans and apes ( Figure 3 ) . We note that structural changes of the human IRGM locus continues to occur within the human lineage with a 20 . 1 kb LTR-rich deletion polymorphism , recently identified and sequenced , located 2 . 82 kb upstream of the ERV9 promoter region [12] . Our preliminary data suggest that this deletion polymorphism alters the relative proportion of alternative splicing of IRGM transcripts ( Figures S5 and S6 ) . We tested for natural selection on IRGM coding sequence using maximum likelihood models to estimate evolutionary rates for individual branches in the phylogeny as well as specific codon changes [13] , [14] . Based on the structural differences in IRGM organization , we first divided our species into three groups: Group 1 consists of species that carry a single copy of IRGM with the ERV9 element ( human ( Hs ) , chimpanzee ( Ptr ) , gorilla ( Ggo ) and orangutan ( Ppy ) ) ; Group 2 consists of species that carry a single copy of IRGM but lack the ERV9 element ( Macaque ( Rh ) , baboon ( Pha ) and marmoset ( Cja ) ) ; while Group 3 was formed by species ( dog and mouse lemur ) that had multiple copies in a tandem orientation ( Figure 4 ) . Phylogenetic branch estimates of dN/dS revealed striking differences between Group 2 ( ω = 0 . 9254 ) and Group 3 ( ω = 0 . 3866 ) with an intermediate value for Group 1 ( ω = 0 . 6073 ) . Group 3 was found to be under constrained evolution ( ω = 0 . 3866 ) and it was significantly different ( P = 6 . 09E−12 ) from a model of neutral evolution . In contrast , Group 1 and 2 gene evolutions were indistinguishable from a model of neutral evolution ( see Text S1 ) .
There are two possible interpretations of our results . First , the IRGM gene is not functional in humans having lost its role in intracellular parasite resistance ∼40 million years ago when the gene family experienced a contraction from a set of three tandem genes to a sole , unique member whose ORF was disrupted by an AluSc repeat in the anthropoid primate ancestor . In light of the detailed functional studies [11] and the recent associations of this gene with Crohn's disease [6] , [7] , we feel that this interpretation is unlikely . For example , McCarroll and colleagues recently demonstrated that a 20 . 1 kb deletion upstream of IRGM associates with Crohn's disease as well as the most strongly associated SNP and that the deletion haplotype showed a distinct pattern of IRGM gene expression consistent with its putative role in autophagy and Crohn's disease . An alternate scenario is that the IRGM gene became nonfunctional ∼40 million years ago ( leading to pseudogene copies in Old World and New World monkeys ) but was resurrected ∼20 million years ago in the common ancestor humans and apes ( Figure 5 ) . In addition to the genetic and functional data , several lines of evidence support this seemingly unusual scenario . First , we find evidence of a restored ORF in humans and African great apes . Second , this change coincided with the integration of the ERV9 element that serves as the functional promoter for the human IRGM gene . Such retroposon-induced alterations of gene expression are not without precedent in mammalian species [15] , [16] . Third , we find that ape/human codon evolution is consistent with a model of nucleotide constraint resulting in depressed dN/dS ratios in the hominid branch ( Figure 4 ) when compared to the Old World and New World species . It is intriguing that the orangutan and gibbon populations possess both a functional and nonfunctional copy of IRGM , which would open the possibility to long-term balancing selection or recurrent mutations ( see Text S1 ) . The inactivating stop codon is shared with all Old World monkey species suggesting an ancestral event . Moreover , we and others [17] find that the structure of the locus is continuing to evolve in humans altering the expression profiles of IRGM transcripts in different tissues . These structural changes are thought to underlie the strong association with Crohn's disease , perhaps , by modulating the efficiency of the autophagic response [17] . Our data suggest remarkable functional plasticity where alleles experience diverse evolutionary pressures over time . Such dynamism in structure and evolution may be critical for a gene family locked in an arms race with an ever-changing repertoire of intracellular parasites .
We retrieved whole genome shotgun sequence of the IRGM locus for chimpanzee ( Pan troglodytes ) , gorilla ( Gorilla Gorilla ) , orangutan ( Pongo pygmaeus ) , rhesus macaque ( Macaca mulatta ) , marmoset ( Callithrix jacchus ) , baboon ( Papio hamadryas ) , and Gray Mouse Lemur ( Microcebus murinus ) from NCBI Trace Archive ( http://www . ncbi . nlm . nih . gov/Traces/trace . cgi ? ) and constructed local sequence assemblies using PHRAP ( http://www . phrap . org ) . We sequenced and confirmed the IRGM genome organization based on DNA samples from four different New World monkey species and from eleven different Old World monkey species . We also resequenced the IRGM gene in unrelated macaques ( n = 5 ) , baboons ( n = 5 ) , orangutans ( n = 12 ) and gibbons ( n = 7 ) . For Microcebus murinus with multiple copies of IRGM , we first isolated large-insert BAC clones , subcloned and sequenced PCR amplicons corresponding to the different copies . Metaphase spreads were obtained from lymphoblast or fibroblast cell lines from human ( Homo sapiens ) , rhesus macaque ( Macaca mulatta ) , marmoset ( Calithrix jacchus ) and lemur ( Lemur catta ) . FISH was performed using either human IRGM probe WIBR2-3607H18 or lemur IRGM BAC DNA LB2-61D22 , LB2-77B23 and LB-61A22 , directly labeled by nick translation with Cy3-dUTP ( Perkin-Elmer ) . Lemur BAC probes were obtained by library hybridization screening of a L . catta genomic library ( CHORI Resources: LBNL-2 Lemur BAC Library [http://gsd . jgi-psf . org/cheng/LB2] ) . Full-length human IRGM transcript was obtained by 5′RACE PCR followed by subcloning ( PGEM-T easy ) and sequencing ( EU742619 ) . RT-PCR experiments were performed using cDNA synthesized ( Advantage RT-PCR , Clontech ) from mRNA extracted ( Oligotex isolation kit , Qiagen ) from total RNA ( RNA Easy , Qiagen ) . Total RNA was obtained from tissues isolated from chimpanzee , rhesus macaque , marmoset and human . IRGM splice variants were detected by a quantitative PCR assay using the LightCycler SYBR Green System ( Roche ) with primers IRGM ( b ) - ( c ) - ( d ) and IRGM all primers ( Text S1 ) . Transcript levels were normalized to the amount of the GAPDH and UBE1 transcript , which also served as positive controls for RT-PCR experiments . We generated multiple sequence alignments using Clustal-W[18] , [19] and constructed neighbor-joining phylogenetic trees ( MEGA 3 . 1 ) [20] . Tests of selection ( ω = dN/dS ) were performed by maximum likelihood using PAML [13] applying the Sites Model [14] to calculate the percentage of codons under positive , neutral evolution or purifying selection and the Branch model [21] to estimate evolutionary pressures at different times during evolution . The Likelihood Ratio Test ( LRT ) was used to assess the significance of different values of ω for different groups . | The IRG gene family plays an important role in defense against intracellular bacteria , and genome-wide association studies have implicated structural variants of the single-copy human IRGM locus as a risk factor for Crohn's disease . We reconstruct the evolutionary history of this region among primates and show that the ancestral tandem gene family contracted to a single pseudogene within the ancestral lineage of apes and monkeys . Phylogenetic analyses support a model where the gene has been “dead” for at least 25 million years of human primate evolution but whose ORF became restored in all human and great ape lineages . We suggest that the rebirth or restoration of the gene coincided with the insertion of an endogenous retrovirus , which now serves as the functional promoter driving human gene expression . We suggest that either the gene is not functional in humans or this represents one of the first documented examples of gene death and rebirth . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
]
| [
"genetics",
"and",
"genomics/comparative",
"genomics",
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics",
"evolutionary",
"biology/human",
"evolution",
"genetics",
"and",
"genomics/functional",
"genomics",
"immunology/innate",
"immunity",
"molecular",
"biology/bioinformatics",
"immunology/genetics",
"of",
"the",
"immune",
"system",
"evolutionary",
"biology",
"genetics",
"and",
"genomics/bioinformatics",
"infectious",
"diseases/gastrointestinal",
"infections"
]
| 2009 | Death and Resurrection of the Human IRGM Gene |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.